# The Palgrave Handbook of Family Policy

*Edited by*  Rense Nieuwenhuis · Wim Van Lancker

The Palgrave Handbook of Family Policy

Rense Nieuwenhuis · Wim Van Lancker Editors

# The Palgrave Handbook of Family Policy

*Editors* Rense Nieuwenhuis Swedish Institute for Social Research (SOFI) Stockholm University Stockholm, Sweden

Wim Van Lancker Centre for Sociological Research University of Leuven Leuven, Belgium

ISBN 978-3-030-54617-5 ISBN 978-3-030-54618-2 (eBook) https://doi.org/10.1007/978-3-030-54618-2

© The Editor(s) (if applicable) and The Author(s) 2020. This book is an open access publication. **Open Access** This book is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s)

and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this book are included in the book's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the book's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Cover image: © DrAfter123/Getty Images

This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

# **Acknowledgments**

The distinguishing feature of this Handbook—the multilevel structure brought together scholars from research communities that have operated separately from each other to some degree. These include scholars attending conferences from Community, Work & Family (CWF) and the Work and Family Researchers Network (WFRN) on the one hand, and the European Network for Social Policy Analysis (ESPAnet) on the other. Bringing these worlds together and expanding their overlap proved to be intellectually stimulating and a joy to work on.

We thank the organizers of the 2019 Community, Work & Family conference in Malta—and in particular Dr. Anna Borg—for facilitating us in hosting a symposium. Discussing the chapters with many of the authors in the same room greatly improved the quality and integration of the chapters in this handbook.

Our gratitude goes out to all the authors who are the foundation of this Handbook. Their enthusiasm for this project and their stimulating contributions were highly motivating. With this Handbook, and its diverse chapters, we hope to contribute to the next generation of family policy research.

Stockholm/Ghent May 2020

Rense Nieuwenhuis Wim Van Lancker

# **Contents**



#### **Part III National**




## **Part VI The Next Decade of Research**


# **Notes on Contributors**

**Willem Adema** Senior Economist in the Social Policy Division at the Organisation for Economic Co-operation and Development (OECD), Paris.

**E. Anne Bardoel** Professor of Management at the Department of Management & Marketing, Swinburne University of Technology, Melbourne.

**Alzbeta Bartova** Postdoctoral researcher at the Department of Public Administration and Sociology of Erasmus University Rotterdam.

**Katia Begall** Assistant Professor in Sociology at Radboud Social Cultural Research, Radboud Universiteit Nijmegen.

**Heejung Chung** Reader in Sociology and Social Policy at the School of Social Policy, Sociology and Social Research, University of Kent.

**Chris Clarke** Junior Economist, Organisation for Economic Co-operation and Development (OECD), Paris.

**Rosa Daiger von Gleichen, Ph.D.** candidate in Social Policy, Department of Social Policy and Intervention, University of Oxford.

**Mary Daly** Professor of Sociology and Social Policy, Department of Social Policy and Intervention, University of Oxford.

**Maja Djundeva, Ph.D.** Researcher at the Netherlands Institute for Social Research (SCP).

**Laure Doctrinal** Ph.D. student in sociology at the Swedish Institute for Social Research (Stockholm University).

**Pearl A. Dykstra** Professor of Empirical Sociology, Erasmus University Rotterdam.

**Tom Emery** Deputy Director of the Generations and Gender Programme, Netherlands Interdisciplinary Demographic Institute.

**Cassandra Engeman** Research Fellow at the Swedish Institute for Social Research (SOFI), Stockholm University.

**Marie Evertsson** Professor of Sociology in Social Policy at the Swedish Institute for Social Research (SOFI), Stockholm University.

**Fernando Filgueira** Professor and Director of the Methods and Data Analysis Unit. Facultad de Ciencias Sociales, Universidad de la República.

**Mia Hakovirta** Academy Research Fellow in Social Policy, Department of Social Research, University of Turku.

**Jennifer L. Hook** Associate Professor of Sociology, Department of Sociology, University of Southern California.

**Eva Jaspers** Assistant Professor of Sociology, University of Utrecht.

**Jana Javornik** Associate Professor in Work and Employment Relations at the University of Leeds.

**Jane Jenson** Professor of Political Science, Political Science Department, Université de Montréal.

**Renske Keizer** Professor in Family Sociology, Erasmus University Rotterdam.

**Meiying Li** Ph.D. student, Department of Sociology, University of Southern California.

**Henning Lohmann** Professor of Sociology, University of Hamburg.

**Laurie C. Maldonado** Assistant Professor at Molloy College.

**Ylva Moberg** Postdoctoral fellow at the Swedish Institute for Social Research (SOFI) at Stockholm University.

**Rense Nieuwenhuis** Associate Professor in sociology at the Swedish Institute for Social Research (SOFI) at Stockholm University.

**Zachary Parolin** Postdoctoral researcher at the Center on Poverty and Social Policy, Columbia University.

**Shahra Razavi** Chief of Research and Data Section, UN Women.

**Cecilia Rossel** Associate Professor at the Department of Social and Political Sciences, Universidad Católica del Uruguay.

**Pia S. Schober** Professor of Sociology, Department of Sociology, University of Tübingen.

**Sebastian Sirén** Ph.D. student in sociology at the Swedish Institute for Social Research (Stockholm University).

**Christine Skinner** Professor of Social Policy, Department of Social Policy and Social Work, University of York.

**Olivier Thévenon** Social Policy Analyst, Organisation for Economic Cooperation and Development (OECD), Paris.

**Tanja van der Lippe** Professor of Sociology of Households and Employment Relations at the Department of Sociology and Research School (ICS) of Utrecht University.

**Wim Van Lancker** Assistant Professor in social work and social policy at the Centre for Sociological Research (CESO) at KU Leuven in Belgium.

**Michel Vandenbroeck** Associate Professor in Family Pedagogy, department of social work and social pedagogy, Ghent University.

**Linda A. White** RBC Chair in Economic and Public Policy and a Professor of Political Science and Public Policy at the Munk School of Global Affairs and Public Policy, University of Toronto.

**Mara A. Yerkes** Associate Professor of Interdisciplinary Social Science at the Department of Interdisciplinary Social Science, Utrecht University.

**Hannah Zagel** Postdoctoral researcher in the Department of Social Sciences at Humboldt-University Berlin.

# **List of Figures**


**xv**



#### **xviii List of Figures**


# **List of Tables**



# **Part I**

**Introduction**

# **1**

# **Introduction: A Multilevel Perspective on Family Policy**

**Rense Nieuwenhuis and Wim Van Lancker**

What do the United Nations, the nation-state, a big city, and your local supermarket have in common? They all craft policies for families. At each of these levels of governance, family policies are formulated, voted for, implemented, and carried out—or not. And it is this whole set of multilevel policies that ultimately affect families' and individuals' choices, opportunities, constraints, and capability in terms of work, care, and well-being.

Of course, it is not simply a matter of trickle-down politics, with the highest level deciding and the other level following suit. There is constant interaction, exchange of norms and ideas, and policy feedback and learning between levels. The story of family policies and their outcomes can be read in many different ways, from bottom-up to top-down, from horizontal to vertical. Still, it is not the supermarket that designs family policies to be carried out at higher levels nor does the supermarket's company regulations affect many people outside of their shop. In contrast, Directives by the European Commission do affect all European citizens, at least in principle, and limit the scope of decision making for national politicians. For that reason, we begin our story at the top tier of governance, and throughout the book we

R. Nieuwenhuis (B)

Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden

e-mail: rense.nieuwenhuis@sofi.su.se

W. Van Lancker

Centre for Sociological Research, University of Leuven, Leuven, Belgium e-mail: wim.vanlancker@kuleuven.be

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_1

descend over national and subnational policies to the local and company-level policies.

It is often overlooked that family policies are formulated and implemented at very different levels, that range from *supranational* (and international) organizations such as the United Nations (UN), the Organisation for Economic Co-operation and Development (OECD), and the European Union (EU), *national* or federal policies and legislation, the *subnational* municipalities, states, or regions in which people live, and finally the *organizations* in which people work. The historical dominance of the nation-state means that the heavy lifting in terms of policymaking is still carried out at that level. As a result, the academic literature on family policies at the national level is developed most extensively. Still, in recent years separate strands of the literature emerged in which each policy level has been studied separately, ignoring possible synergies and discrepancies between these different levels at which family policies are formulated and implemented.

Family policies have had a long-standing and profound impact on how families live their lives, how they are formed, and even what is considered to be a 'family.' The area of family policy constitutes a broad range of legislation and policies, that set out to achieve goals that include poverty alleviation, compensation for the economic costs of children, fostering employment, improve gender equality, support early childhood development, and raise birth rates (Thévenon, 2011). Although often not considered in analyses of 'family policy,' policies that relate to reproductive rights extend to policies on contraceptives, and induced abortion (Levels, Need, Nieuwenhuis, Sluiter, & Ultee, 2012). Family policies were understood early on as an important nexus between family and the state (Gauthier, 1996). Not surprisingly, the literature on family policy outcomes blossomed, producing important concepts and insights on how individuals interact with the organizational and institutional context shaped by family policies, inspired by a variety of disciplines (Van der Lippe & Van Dijk, 2001). While the literature on family policy used to be Western-centric, we now witness a true globalization of both the development of family policy and academic research on the issue (Robila, 2012).

Today, several societal developments point toward an increased importance of examining family policy at different levels. Organizations, often employers, offer attractive work(–family) arrangements in the competition for talented workers, and compensate to some extent for the absence of national-level policies (Den Dulk, Peters, & Poutsma, 2012). At the same time, they are increasingly forced by national legislation to allow workers to request flexible working hours. Federal states have faced pressures to decentralize family policies or added new layers of governance, giving room to ideological differences to shape regional policy differences (Béland & Lecours, 2005), which is clearly illustrated in the United States where states or even large cities such as New York have implemented family leave arrangements (Boushey, 2016). Urbanization and associated population decline in rural areas is associated with regional differences in the availability of public services. National policymakers are facing persisting gender and economic inequality, family diversity, and other so-called new social risks, budget deficits, and a seemingly constant concern about economic growth. Increasingly, supranational organizations issue recommendations or directives to which other policymakers have to relate. Examples include recommendations by the OECD, nations committing to the Beijing Platform for Action by the UN, and the 'Directive on work-life balance for parents and carers' adopted by the European Commission.

Addressing the disconnect between the academic literatures on different levels of governance with respect to family policies and their outcomes is one of the key aims of this Handbook.

### **Purpose and Contribution of This Handbook**

The purpose and main contribution of this handbook is to provide a multilevel view on family policy outcomes, thus, as described in the previous section, combining insights on family outcomes at levels ranging from the potential impact of supranational organizations, via national policies, subnational/regional, and finally organizations or employers. Although the focus is on outcomes, at each of these levels some of the chapters also address (factors behind) the development of family policies.

The handbook seeks to make three additional contributions. The handbook will both take stock of existing theories and empirical research, as well as develop an agenda for the next decade of research on family policy outcomes. To that end, at each of the four levels discussed above, chapters are included that examine key concepts, summarize lessons learned, and focus on the frontiers of the research agenda.

Next, the handbook will be sensitive to heterogeneous policy impacts. As the literature on the outcomes of family policies burgeoned, it became increasingly clear that individuals and organizations react to family policies in heterogeneous ways. Family policy sits at the intersection between parenthood, family and employment, and taking due account of heterogeneity of parents in terms of preferences, family structure, class, opportunities and constraints in relation to decisions about care and employment, helps understand the outcomes of these policies more accurately (Mandel, 2012).

The final contribution of this handbook—inherent to its multilevel structure—is to bring together contributions from various academic communities that have operated with some degree of separation from each other. Such wide-ranging overview is not yet available and will combine academic literature that is well-developed (such as the organization level and the national level) to literatures that are rapidly gaining more prominence (such as the subnational/regional and supranational levels). Bringing together various academic communities in a comprehensive handbook fosters learning interdisciplinary lessons, cross-cutting across disciplines that include (but are not limited to) sociology, political science, business administration, demography, and economics.

# **Outline of the Book**

As part of the introduction to this Handbook, Mary Daly (Chapter 2) sets out the framework of what constitutes 'family policies,' how they are changing, and how family policy should be theorized and studied going forward. The chapter provides the conceptual foundation on which the other chapters in this volume build, and provides the starting point for many of the empirical, conceptual, and theoretical discussions that will be tackled more in detail in particular chapters throughout. Daly's overall aim is to set out the constituent elements of family policy and explore the significance of family policy, especially in a context of socioeconomic, demographic, and political change. Focusing on long-standing welfare states, the first part of the chapter considers the functions of family policy, outlining both the classic aims and objectives and those found in more recent family policy. The different instruments of family policy are identified, including cash and tax payments, services, and leaves. These three mainstays of family policies are the focus of many of the subsequent chapters in this volume. In all of this, variation is emphasized and examined. In its second part, the chapter links changes and debates to the complex issues animated by family policy, such as questions about personal and national values, intergenerational balance and equity, gender equality and the differential politicization of aspects of individual and family life. Daly concludes that the agency of family policy change operates at different levels of policymaking. It is to these levels that we now turn.

### **Supranational**

The attention for the role played by supranational organizations and governing bodies such as the European Commission, the Organisation for Economic Co-operation and Development (OECD), the World Bank (WB), the International Labour Organisation (ILO), and the United Nations (UN) in the making of family policies has long been somewhat underwhelming. Yet, these organizations have a long history of influencing national policies, setting minimum standards, and adopting binding decrees or directives. For instance, the European Commission is increasingly involved in social outcomes and policies, including family policies, of European Union member states. This is quite clear in the development from the Lisbon agreements, over the social investment package to most recently the launching of the 'European Pillar of Social Rights.' As part of the latter, the European Commission issued a 'Work-Life Balance directive' in April 2017, with proposals that pertain to paternity leave, parental leave, carers' leave, and flexible working. This was adopted by the European Parliament in May 2019. Even more global, as early as 1919, the International Labour Organisation (ILO) influenced policymaking at lower tiers of governance through what Kahn and Kamerman (1978) termed 'implicit family policies,' policy domains and decisions that have consequences for families with children. In this case, a concern for equal treatment of men and women in the workplace and occupational health led to the adoption of a convention that obliged countries to put flesh to the bone of their family policies across the globe.

To fully capture the importance of the supranational level of governance in understanding the outcomes of family policies, this section starts with a chapter by Jane Jenson (Chapter 3) in which she charts the intellectual history of how international organizations (IOs) began to expand their areas of competence and intervention in the realm of family policy by proposing new policy directions in the name of sustaining employment, fighting poverty, social investment, or social development. Many of these touched directly and even more indirectly—on policy fields that have long been part of the domain of family policy. Of course, exerting influence on policymaking at national, subnational, and company levels is not a one-way street, but it is the result of a complex exchange of ideas, influence, and power. The chapter explores these exchanges both horizontally across IOs and vertically across policy tiers (bottom-up as well as top-down). In the end, it seems that the three mainstays of family policy identified in this handbook were regularly underexposed because the policy frames shifted toward children or workers, stripping individuals from the family they are living in.

#### **8 R. Nieuwenhuis and W. Van Lancker**

Moreover, it is not because IOs of supranational bodies exert influence on norms and ideas or launch binding directives, that these are always fully transformed into actual policymaking at the national or lower policy levels. In her chapter, Linda White (Chapter 4) argues that some countries or national governments may regard the authority of IOs with suspicion or may regard IO endorsement as a reason *not* to adopt a policy domestically. Thus, even if policy ideas or norms become dominant at the international level, there can be varying levels of receptivity to those ideas within domestic policy processes. The chapter examines international norms on gender equality and human rights in general and family policies in particular, and shows that even if IOs may not directly affect decision making, they can facilitate the spread of ideas of best practices among member countries through their reports and conferences. Yet, White shows for liberal welfare states (Australia, Canada, New Zealand, UK, USA) that international norms are diffusing to domestic policy processes rather unevenly, particularly in countries in which decision making on family policies is decentralized. As such, IOs are still less authoritative than the national level when it comes to on-the-ground policymaking.

Interestingly, though, IOs are usually not monoliths but often comprise different agencies, secretariats, offices, and administrations within the overarching organization. In her chapter, Shahra Razavi (Chapter 5) provides a comprehensive overview of how three agencies within the United Nations (UN) look at family policy through different lenses, shaped by their different mandates and objectives, institutional context, intellectual history, and administrative capacities. These perspectives can be conflicting or not in sync with one another. With its focus on labor standards and decent work, the International Labour Organisation (ILO) had an important impact on families through norm-setting mechanisms and procedures, for instance through its conventions on maternity protection. At the same time, its focus on women as workers overlooks gender stereotypes and the role of men in family life. In a similar vein does UN Women impact family life, centering on key feminist concerns such as domestic violence and reproductive rights. UNICEF, then, sees families through the lens of children's rights, which tends to confine women to maternal roles, responsible for breastfeeding and ensuring children's nutritional needs. This unavoidably leaves out the needs of working parents, in particular mothers who do the bulk of unpaid care. At the same time, advocacy work on children's well-being has a tangible impact on women's rights and gender equality.

As becomes clear throughout the chapter, family policy sensu stricto is usually not front and center in these agencies' ideas and norms, but their actions do shape family life on a global scale. Although there is pathdependency in how different UN agencies approach family policy, there is also learning and changing of position. This is exemplified by a recent turn to family-friendly policies, connecting children's rights to women's rights and feminist concerns. It is still an open question whether these changes at the ideational level will fully translate into the programmatic level.

In any case, to fully understand how the supranational policy level and IOs shape family policies around the globe, one not only has to take due account of differences *between* IOs, their objectives and institutional makeup, but also within IOs. Just as there is no 'one UN' approach, as highlighted by Razavi, there is no 'one IO' approach or strategy to affect family policy to date.

## **National**

At the national level, few countries had implemented a coherent set of family policies up until the 1980s. From that point on, however, state support for families expanded strongly across OECD countries. As a corollary, the study of national family policies became a rapidly expanding academic field, producing important insights on how individuals interact with the institutional context shaped by family policies drawing on qualitative as well as quantitative methods, inspired by a variety of disciplines (Van der Lippe & Van Dijk, 2002). Alfred Kahn and Sheila Kamerman (1978) provided one of the first attempts at developing a comparative framework to study family policies. Since then, several studies showed how spending on national family policies increased strongly over time and how more employmentrelated forms of family policy such as childcare services and leave schemes developed alongside traditional cash components such as marriage subsidies and family benefits. Importantly, a key insight from the literature is that family policies do not emerge in a political vacuum but that they are path dependent and tied to dominant norms on gender and motherhood, and that they interact with and emerge in the context of broader welfare state arrangements (Gauthier, 2002; Montanari, 2000; Thévenon, 2011).

By drawing on specific conceptual and analytical frameworks, scholarship on family policies played an important role in *understanding* family policy outcomes and its variation across countries. In particular, the study of national family policies is strongly connected to the inclusion of gender in welfare state research. A famous example was the introduction of the concept of 'defamilization' (Lister, 1994; McLaughlin & Glendinning, 1994) to counterbalance the central role played by 'decommodification' in Esping-Andersen's (1990) influential welfare state typology—and to bring in a focus on gender and family and care relations. Decommodification is the degree to which individuals are able to uphold an acceptable standard of living independent of employment. In other words, if one does not work, how well is one protected by the welfare state? Yet, when the male breadwinner model is dominant and women's roles in providing unpaid care are subordinated to men's paid employment, such an approach of classifying welfare states and understanding its outcomes tends to be 'gender blind.' In response, Lister (1994) suggested the concept of familism (and the corresponding defamilization function of the welfare state) to better reflect how institutional arrangements cater to people engaging in care for children or relatives. Defamilization, then, is the degree to which individuals can uphold an acceptable standard of living independent of care relations within the *family*. The concept of defamilization subsequently was used to understand the outcomes of welfare states, *a forteriori* family policies, and how it affects individual's behavior in terms of care and employment given the gendered relations in terms of work and care within the family (Korpi, 2000; Lewis, 1992; Orloff, 2009). Policies enabling defamilization, such as paid parental leave and public provision of childcare, are instrumental in supporting women's labor force participation (Thévenon & Luci, 2012). As such, the study of family policy is directly linked to the study of gender in relation to the state, the labor market, and the family.

Two chapters in this Handbook chart key concepts and theoretical approaches used in the analysis of national-level family policies. Hannah Zagel and Henning Lohmann (Chapter 6) provide a systematic and thorough discussion of the theoretical concepts used in comparative family policy, gender, and welfare state research. Besides providing an overview of the state of play of the theoretical ideas in this research area, they propose six criteria for evaluating and properly understanding family policy concepts. Such a multifaceted approach toward the conceptualization of family policies helps understand how and why family policies shape the lives of families. Zagel and Lohmann argue that a systematic discussion of the value and meaning of concepts used in the area of family policy research is crucial, since there are many similarities in the underlying ideas of concepts but also important differences which have mainly gone unnoticed. One prime example is the concept of defamilization which we have used ourselves in the previous paragraph. Several versions of the concept are used in the literature, usually without paying due attention to the history of the concept or the subtle differences in the interpretations of different varieties of the concept. So, gaining thorough understanding of the outcomes of family policies starts with gaining better understanding of the concepts underlying such research.

In the next chapter, Jana Javornik and Mara Yerkes (Chapter 7) take the multifaceted approach to understanding family policies even further. They argue that translating concepts into measurable indicators of family policies remains a conceptual and analytical challenge. Drawing on examples of British and Swedish policies on childcare, they discuss the difficulties involved in conceptualizing family policies in a cross-country comparative perspective and propose using the Capabilities Approach (CA) as a way forward. They demonstrate that the CA helps capture the complexity of analyzing family policies and their outcomes in a meaningful way, in particular by shifting focus to what extent family policies actually allow families, and women, in particular, to make meaningful choices within varying contexts. Applying the CA helps to evaluate family policies in a broader perspective than is usually done.

It is important to recognize that family policies are developed and implemented in the pursuit of attaining certain objectives. Those objectives are usually tied to a particular time, place, and context. Childcare, for instance, is a mainstay of modern family policies, receiving ample political, administrative, and academic attention. In Chapter 8, Michel Vandenbroeck dissects where this attention originated and provides a brief history of the development of childcare policies, or more broadly Early Childhood Education and Care (ECEC) policies, in a diverse set of countries. The chapter shows how childcare policy evolved from its nineteenth-century objective of providing custodial care for poor, working-class children to a twenty-first century objective of improving cognitive and noncognitive abilities of all children (Kamerman, 2006). To this end, improving quality and accessibility for all families became important policy goals. At the same time, as is also demonstrated elsewhere in this book, policies shape different opportunities for families at different intersections of inequality: disadvantaged children who would benefit most from high-quality childcare are usually the ones left out (Van Lancker & Ghysels, 2016). Vandenbroeck argues that policies successful in equalizing access to high-quality childcare are policies that consider childcare as a public good rather than a marketized commodity, and polices that are built on the notion of proportionate universalism in which universal policies are combined with targeted actions toward disadvantaged families. In the majority of countries, however, this will require substantial investments in both quantity and quality.

The empirical literature on the outcomes of national family policy is well-developed since Kamerman and Kahn's pioneering work. In Chapter 9, Willem Adema, Chris Clarke, and Olivier Thévenon take on the challenge of summarizing such a vast field and provide an overview of research findings. They present an overview of the changes in family and work behavior and highlight the diversification of family models. Importantly, similar to Vandenbroeck, they point at growing inequalities associated with these changes. The chapter also looks forward by identifying important challenges ahead, such as the growing number of singles and single-parent families, the diversification of family forms, and growing inequalities between families. Recent developments in family policy are aimed largely at helping parents to balance work and family commitments and reducing gender inequality, including the provision of ECEC services, fathers-only paid parental leave and support for flexible working at different levels of governance, all three themes which are developed further in the subsequent chapters.

By now, it is clear that the vast majority of scholarly work in the field focuses on OECD and EU countries. Yet, it is important to emphasize the importance of studying family policy developments in other parts of the world as well, which enables to test the aforementioned concepts, classifications, and methods outside of the traditional Western-centrist 'worlds of welfare.' That is the purpose of the chapter by Fernando Filgueira and Cecilia Rossel (Chapter 10). They discuss country experiences across the globe in developing systems of family allowances, leave schemes, and childcare services. A comparison with the European and OECD countries shows that the majority of these countries have a long road ahead in growing a seed of family policies into a strong foundation of the welfare state. In many developing economies, family policies were established relatively recently and usually carry little weight in the total fiscal effort of these countries. And while some countries have seemingly generous leave policies in place, offering for instance full wage replacement rates, this is merely a paper reality. Most workers don't actually qualify for these policies. At the same time, however, the breadth and depth of family policies are growing in the majority of countries across the globe. In that respect, it is illuminating to see that in many of these countries the policy blueprint resembles the one in European welfare states albeit often with a particular path-dependent twist. This demonstrates that the conceptual and analytical framework applied in the literature to understand the outcomes of family policy bears broader relevance.

Let us now return to the issue of inequality. While earlier attempts at quantifying the outcomes of family policies and the effects of these policies on families' work and care behavior focused on the 'average family,' more recent studies acknowledge the heterogeneity of policy impacts and the existence of unintended consequences of 'deliberate policy action' (Merton, 1936).

For instance, the use of formal childcare was found to be socially unequal, biased toward higher educated mothers, leading to the observation that the 'Matthew effect' prevailed in family policies (Pavolini & Van Lancker, 2018). Trade-offs were found (Pettit & Hook, 2009), with for instance provisions for part-time work and childcare services increasing the number of employed women while at the same time increasing gender inequality in wages. The provision of leave was found to be an indispensable measure to facilitate care and employment in motherhood while at the same time very long periods of leave were found to be a mechanism of exclusion of women from the labor market (Nieuwenhuis, Need, & Van Der Kolk, 2017). Moreover, research showed that family policies have the potential to particularly benefit specific groups, such as children (Bradshaw & Finch, 2002), the lower classes and the lower educated (Korpi, Ferrarini, & Englund, 2013), large families and single parents (Nieuwenhuis & Maldonado, 2018), in turn contributing to lower poverty rates and higher levels of well-being (Engster & Stensöta 2011; Van Lancker & Van Mechelen, 2015).

Several chapters explicitly take this heterogeneity in family policy outcomes into account. Jennifer Hook and Meiying Li (Chapter 11) examine trade-offs, interactions, and unintended consequences of the impact of family policies, parental leave, and childcare services in particular, on women's labor market outcomes. They present a concise overview of what they refer to as the 'uneasy consensus' that emerges from the literature on the effects of publicly funded childcare and parental leave on employment outcomes: both measures are considered to be supportive for women's employment. At the same time, this uneasy consensus falls apart once the heterogeneity of policy effects by social class is considered. They argue that a clear gap in the current state of the literature is our limited understanding of heterogeneity in policy effects as well as how inequality more broadly affects the relationship between family policy and women's employment. Childcare and parental leave policies are sometimes only to the benefit of middle- and upper-class women, while gains in terms of inclusion on the labor market can be offset by increasing inequalities between men and women in the labor market. To evaluate whether family policies attain their intended objectives, it is of utmost importance to be aware of trade-offs and heterogeneities. But studies come to contrasting conclusions, and there is still no consensus on the actual impact of educational level, or contextual factors such as labor market institutions, or cultural norms, on the outcomes of family policies. Hook and Li thus argue that our current conclusions in this area are based on, and they put it mildly, less than ideal policy data. How data limitations impede further progress in the field of family policy research is a key issue that is discussed throughout the book, in particular in Chapter 24.

Christine Skinner and Mia Hakovirta (Chapter 12), explore the current policy designs and principles pertaining to child support. There is a long policy history in most Western countries, but child support policies and separated families make unhappy bedfellows. The institutional and operational challenges are considerable as policies try to keep pace with changes in intimate family relationships, changes in household structures and changes in gendered patterns of employment. For example, to provide accurate determinations of child support, policymakers must track the procreation activities, living arrangements, and employment patterns of parents. It therefore matters whether parents go on to establish new partnerships and have more children. It also matters for children entitled to child support, who they live with and how the patterns of parental shared care are agreed upon. Given that the lifetime of a child support case could last for the duration of an individual's childhood (possibly as long as 18–21 years), the propensity for complexity is inherent within the policy design and principles. And it shows in the results. Skinner and Hakovirta demonstrate on the basis of a new expertbased dataset that a change toward greater gender equality in the division of labor in families is not reflected very much in child support systems over the last decade. They too point to heterogeneity in policy effects and trade-offs, in that the growing recognition of gender equality in child support systems might potentially jeopardize the poverty-reducing effects of those systems.

Further exploring the issue of changes in household structures and how policies cope, Laurie C. Maldonado and Rense Nieuwenhuis (Chapter 13) examine both the intended and unintended consequences of family policies for single-parent families. They analyze how national family policies support the specific group of single parents, addressing in particular how family policy interacts with their resources and employment. In doing so, they examine how redistributive and financial support policies not only operate differently than work–family reconciliation policies but also have a different impact for single parents in comparison with couples.

While most national family policies are geared toward families with children, the issue of care is obviously not limited to parents caring for children alone. As such, policies facilitating employment or care, and policies protecting against financial hardship while providing care, bear broader relevance for families in which children care for their parents. Pearl Dykstra and Maja Djundeva (Chapter 14) consider cross-national differences in European national policies regarding caring responsibilities for older family members. They argue that it is unfortunate that care policies are separated—conceptually as well as in practice—in 'young' and 'old' sections, because such separation overlooks interdependencies across generations and only provides a fragmented notion of what families are about. As such, they broaden the notion of family policies and consider questions such as the public provision of services for frail older adults, pension credits, and cash for care schemes. The chapter differentiates between policies that *free* family members from caring responsibilities and policies that *enable* them to care for older generations. The chapter shows that long-term care systems across Europe showed '*limited convergence*' (Ranci & Pavolini, 2013, p. 312): while universalistic systems retrenched their provisions, most of the residual care regimes expanded theirs. Throughout Europe, there has been a trend toward *refamilialization* of care, that is, shifting responsibility for long-term care from the state to individuals and their families. The chapter furthermore brings intergenerational relationships into the framework of heterogeneous policy impacts.

No matter how encompassing the literature on national family policies has become, many issues and challenges are still not well understood. Exciting new pathways are currently being explored, with for instance researchers focusing on understudied groups. Alzbeta Bartova and Renske Keizer (Chapter 15) focus on fatherhood, in search of support for the dual-carer family model. Starting from the observation that there is much heterogeneity in the care roles fathers adopt, their chapter explores family policies that have a potential to promote the dual-carer family model across 13 European countries. They find that although all countries offer fathers an individual nontransferable entitlement to child-related leave, in general, these policies do not sufficiently challenge the gendered distribution of paid and unpaid work. Since highly educated fathers and fathers with higher income are more likely to devote their time to childcare compared to their counterparts with lower educational attainment and income, Bartova and Keizer examine to what extent these policies help fathers with a different socioeconomic background to take up more childcare responsibilities. This, in turn and in combination with the relatively new EU Directive on Work–Life Balance for parents and carers, will presumably lead to more gender equality.

Marie Evertsson, Eva Jaspers, and Ylva Moberg (Chapter 16) introduce the concept of parentalization, defined as the ability to become a parent and be recognized as such, both legally and via social policies. In two parts, they examine who can become a parent and how, and who can share in the care of the child. Empirically, this chapter uses register data to show trends in the number of children with same-sex female and same-sex male parents, and how family leave policies apply to same-sex couples in the four largest Nordic countries and the Netherlands. These European countries were in the forefront of legally recognizing same-sex marriage and parenthood. This is a clear research frontier, since same-sex couples face greater difficulties being legally recognized as parents and to share the leave between them. This is important, not least as parental leave rights are key to long-term shared parenting and care, enabling a more equal division of leave for those who so desire.

# **Subnational**

The chapters in this section focus on family policies that are formulated or implemented at the subnational level, for instance the level of the state, province, or municipality. Examining regional variation at these levels seems increasingly important in the light of ongoing decentralization. More generally, family policy provisions at the national level may not always be indicative of what citizens in different parts of the country can expect. For instance, although the United States is often described as one of the very few countries in the world that has no federal legislation on paid parental leave (Heyman & Earle, 2009), states and cities are implementing family policies at a rapid pace (Boushey, 2016).

Indeed, in many cross-national policy analyses, the United States is conceptually understood as the ideal-typical liberal regime that offers marketbased solutions to social problems, including work–life balance issues. Fittingly, the country lags behind most others in the world with respect to family policy generosity. However, there is considerable subnational variation that cross-national comparative research often overlooks. Therefore, in Chapter 17, Cassandra Engeman addresses how family leave policies developed across US states over time. Drawing on state legislative documents, the chapter identifies the dimensions on which state policies typically vary. The chapter shows how state policies generally shifted from maternity leave to gender-neutral family leave to paid leave schemes, characterizing some states as more innovative than others. Comparing leave policy development in the United States to other countries, the chapter highlights often overlooked, innovative aspects of US family policy, most notably, the emphasis on individual entitlements to leave rights and the distinction between leave for health and caregiving needs that are gender-neutral in purpose. The chapter thus presents subnational variation in family policy as a rich area for future research and demonstrates that cross-national comparative research on family policy outcomes should consider subnational variation when including federal systems, such as the United States, in the analysis. Taking stock of evidence of increasing subnational divergence throughout the past two decades, Zachary Parolin and Rosa Daiger von Gleichen (Chapter 18) investigate the extent to which a family's ability to achieve financial recourse from the welfare state and/or market varies across the 50 United States. The findings demonstrate that states' family policy packages vary widely with respect to the generosity and accessibility of social assistance, health insurance, net incomes from minimum wage employment, publicly supported childcare and pre-K, and paid family leaves. Moreover, the chapter investigates the extent to which this cross-state variation in family policies helps to explain differences in family employment and poverty outcomes across the United States. The findings indicate that states became more similar in the amount of benefits received from the federally funded Supplemental Nutrition Assistance Program (SNAP), Supplemental Social Security Income (SSI), and the Earned Income Tax Credit (EITC), but less similar in terms of benefits received from Temporary Assistance for Needy Families (TANF). Moreover, divergence in policies has not always translated into divergence in social outcomes.

Drawing on the case of Germany, Pia Schober (Chapter 19) argues that considering different levels of regional and local variations within a county offers great potential for generating new insights on mechanisms of how family policies affect families' practices and choices. The chapter first describes the institutional context of childcare provision in Germany and the existing variation in regional provision and take-up of services. It then reviews different theoretical perspectives on potential drivers of policy variations at the municipality and federal state level and connects these to existing empirical evidence. The second part of the chapter proposes a framework for investigating socially stratified parental work-care choices at subnational levels by connecting a macro–micro rational choice perspective with the capability approach and the accommodation model of childcare choices. Particular attention is paid to variations across families with varying levels of education, income, migration background, and partnership status. After reviewing existing evidence on the effects of regional variations in childcare provision on social inequalities in take-up, maternal employment and work– family balance, and on some of the mechanisms, the chapter concludes by pointing to research gaps and new frontiers of regional family policy analysis. It outlines the current challenges and new demands for data collection and linkages necessary to realize the full potential of regional family policy analysis.

The Netherlands provides extensive and flexible childcare provision to support young parents. Even though the Netherlands is not a federal state, unlike the United States and Germany covered in previous chapters, the national-level policy on childcare isn't evenly distributed, and provision is better for some than for others. Childcare, like education and healthcare, is centered around physical infrastructure and this ensures that provision and access are structured by traditional geographies of inequality. In Chapter 20, Tom Emery examines detailed data on childcare in the Netherlands from a range of data sources to explore social gradients in access to childcare at the very local level. Survey respondents are geocoded and are linked to the *Landelijk Register Kinderopvang en Peuterspeelzalen* (Register for Childcare and Toddler Groups). In linking this data, three often neglected dimensions of childcare provision are captured. First, at the most basic level the geographic proximity of childcare services is captured (i.e., distance to nearest provider). Second, the availability of choice is measured (i.e., number of providers within 2 kms). Third, the diversity of options is measured (i.e., the ratio of different types of providers). The richness of the geographical registry data and the survey data allow for a detailed assessment of childcare provision at the regional and subregional level and the extent to which it is associated with wider socioeconomic cleavages. The findings illustrate the strong association between the geographical availability of childcare and the ability of women to work longer hours. However, more informal forms of childcare support such as childminders or family networks are not directly associated with a return to employment. Specifically, these results enable the identification of specific subregions where geographical challenges in the provision of childcare persist and an overview of subnational initiatives to address these childcare deserts. Precise estimates of geospatial contexts translate directly into more precise policy interventions and a shorter line between research and policymakers.

# **Organizations**

A vast literature exists on how organizations support their workers. Organizational norms and practices such as opportunities for part-time work, flexible working hours, overtime regulation, employer-provided childcare facilities, and telework options are now known to affect the well-being of workers (Lewis, Anderson, Lyonette, Payne, & Wood, 2017). More recently, crossovers started to emerge between the literature on what organizations do and the literature on what states bring about with respect to work and family, and their interaction (Abendroth, Van der Lippe, & Maas, 2012). This has led, for instance, to the important insight that organizations compensate for the absence of national-level policies to combine work and family, but only to a limited extent (Den Dulk et al., 2012).

The motivation to study the workplace context is that existing research shows there is often a gap between policy and practice, that is, workers refrain from taking advantage of existing national and/or workplace policies. The fact that workers do not always utilize policies, even when they need to, suggests that there are constraints which influence their sense of entitlements and claims to existing policies. Hence, to fully understand how work–family policies play out at the workplace we also need to address how arrangements and policies are mediated, translated, and implemented within organizations. It is at the workplace and work-organizational level that formal work–family policies are converted into entitlements and claims, where requests are granted or denied. This can have profound consequences for social inequality between workers within the same company, or between workers in different companies. Heejung Chung (Chapter 21) develops the argument that not only do organizations define the 'final availability' workers actually have toward various family policy arrangements, organizations may also provide various additional arrangements through occupational policies which are not set out in the national-level agreements that are crucial in addressing reconciliation needs of workers. This chapter empirically examines what types of arrangements are provided at the organizational level to address work–family demands of workers. It further provides a synthesis of the literature on why organizations provide such policies as well as to whom it is provided, which is linked to the possible outcomes of organization-level family-friendly policies.

With the importance of organizational family policy and practice clearly demonstrated, a question of central importance is who benefits from these policies. This pertains not only to which organizations provide such policies, but also which workers within organizations can access them. This is examined by Katia Begall and Tanja van der Lippe (Chapter 22), using uniquely rich, cross-national data with a multilevel design. Their chapter focuses exclusively on the question to what extent access to and use of organizational policies related to work–life balance, flexibility, training, and health is stratified by education. The argument is that work–family research has paid little attention to differences by education above and beyond the differences which can be expected based on occupational and job differences. Whether lowskilled (measured by education) workers are able to make use of these policies and in how far they can benefit as much from them as higher-skilled workers in organizations are therefore central questions in this chapter. Advocating a multilevel perspective, the chapter analyses how the availability, use, and consequences of work–family policies differ by education incorporating a wide range of indicators on the individual and organization level.

Despite organizations increasingly operating in multiple countries, there has been limited discussion among researchers regarding the roles and responsibilities of human resources (HR) managers in multinational enterprises and to work–life management in the global context. Anne Bardoel (Chapter 23) discusses and analyzes multinational enterprises (MNEs), where managing work–life issues presents a number of challenges for HR departments because of the complexity of implementing policies that require sensitivity to local issues such as cultural traditions and legislation. A tension-centered approach to analyzing these complexities in MNEs with a particular focus on work-life management and strategy development provides insights into constraints and challenges to organizations operating globally. There is evidence that tensions often exist in MNEs between corporate/global HR, local HR, and operational line managers involved in the implementation of work-life policies and practices. The promise of the tension-centered approach is that it provides insight into the ways these tensions are resolved in practice, and that it can point toward strategies to improve relevant practices.

## **The Next Decade of Research**

This Handbook is testimony to the multitude of methods, research traditions, and data sources that have been deployed to study family policies and their outcomes. While early attempts to cross-national family policy research had to use fragmented data, recent studies have a whole range of cross-country comparable survey data on income, employment, and living conditions at their disposal, alongside publicly accessible databases with policy indicators in the fields of childcare, child benefits, leave policies, and fiscal policies. An increasing number of studies rely on more technically advanced methods, such as microsimulation studies, quasi-experiments, or studies based on model family types to examine the particular design of family policies. Still, scientific progress in terms of understanding the outcomes of family policies at different levels is often hampered by a lack of high-quality data. In their chapter, Sirén, Doctrinal, Van Lancker, and Nieuwenhuis (Chapter 24) focus on the salient issue of childcare. Arguably it is the policy area which gets the most attention in today's discussions on family policy, while the data and indicators available to study it are least developed. In their chapter, they discuss the current availability of comparative data sources and their limitations, and spell out what kind of data investment is needed to push forward the childcare research agenda in particular, bearing broader relevance for data issues in family policy research in general. In particular, they argue for the need to engage in a research agenda that integrates family policies, including social care services, as essential components of social citizenship.

In Chapter 25, Nieuwenhuis develops a research agenda for examining family policy outcomes with respect to vertical economic inequality between households, arguing that family policies have wrongly been neglected as a determinant of vertical economic inequality. Three questions are central to this research agenda: who uses family policy, to what income effect, and with whom do they live? Family policies have been linked to women's employment and earnings, and to lower vertical income inequality. Yet, the literature also makes abundantly clear that family policies come with trade-offs along the lines of gender and class, as well as Matthew effects. These mechanisms need to be better understood to integrate family policy in analyses of—and recommendation against—high and rising inequality. The challenge ahead is to understand what (combination of ) family policies may be inclusive to a wide range of families across the full width of the income distribution.

In the final chapter 26, Van Lancker and Nieuwenhuis highlight five major societal challenges for the future outlook and outcomes of family policies, and reflect on what the handbook teaches us about how to effectively address these challenges. The challenges pertain to the (1) levels of policy implementation, and in particular globalization and decentralization, (2) austerity and marketization, (3) economic inequality, (4) changing family relations, and (5) welfare states adapting to women's empowered roles. The chapter concludes with a reflection on what we learned, and are yet to learn, regarding the role played by family policies in cushioning economic, social, and health shocks of various kinds. How well current theories and empirical findings help us understand to what extent family policies support families during extraordinary times remains a challenge for the next decade of research.

## **References**


Boushey, H. (2016). *Finding time*. Cambridge, MA: Harvard University Press.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **2**

# **Conceptualizing and Analyzing Family Policy and How It Is Changing**

**Mary Daly**

This contribution provides a conceptually based analysis of family policy in Europe, identifying the defining constituent elements as well as the main changes underway. The three questions that underlie the chapter are: What constitutes family policy? How is it changing? And how should we conceive of and study family policy going forward? To address these, we need a clear conception of family on the one hand and the relevant policies on the other. A core aim of the chapter is to set out an analytic framework which identifies the interrelationships, functions, policy constituents, and trends in family policy. Toward this end, the chapter is organized into two main parts. The first part focuses on the conceptualization of family policy. It proceeds by first drawing from the existing literature and second reviewing briefly both the historical evolution and particular models that have prevailed historically in Europe. The second part of the chapter identifies major contemporary trends, focusing on three main areas of family policy: income supports for families with children, early childhood education and care (ECEC), and parentingrelated leaves from employment. As well as setting out the detail, this section also considers the significance of these changes in terms of underlying large shifts and changing conceptualizations of family policy. In the final section I set out some considerations for evaluating and theorizing family policy going forward. The chapter is informed throughout by a comparative sensibility,

University of Oxford, Oxford, UK e-mail: mary.daly@spi.ox.ac.uk

M. Daly (B)

mainly from a perspective of the European Union (EU) countries. An overarching line of interrogation and intellectual challenge for the field of family policy analysis as a whole is to move beyond a quite particularistic and narrow orientation which characterizes the field.

# **Conceptualizing the Field of Family Policy**

## **Insights from Existing Literature**

In terms of focus and conceptualization, research and scholarship to date have concentrated mainly on identifying the different dimensions and areas of family policy, strongly favoring analysis of institutional features and underlying exigencies, especially in terms of the "problems" policy is intended to ameliorate and hence its functions (Ferrarini, 2006; Hantrais, 2004; Kaufmann, 2002; Wennemo, 1994). This literature has shown that the policy field is potentially quite broad and that a varied set of measures has been developed for the purposes of supporting families and regulating family life. These include cash transfers, tax credits and tax allowances, employment leaves, child-related education and care services, family services, employment leave arrangements, and legal measures to encode rights and responsibilities (Fox Harding, 1996; Millar & Warman, 1996; Saraceno, 2011). The comparative literature—which is a marked characteristic family of policy research—also makes clear that there is great variation in the primacy, role, and constituent elements of family policy cross-nationally (Daly, 2010; Gauthier, 1996; Saraceno & Keck, 2010).

All of this has a deceptive simplicity to it. In fact, though, there are many complexities to be worked out, especially in a policy field that is developing quite rapidly with policymakers appearing to be more and more prioritizing the family as a focus of policy development and intervention.

The definition and scope of family policy is not clear-cut and there is no consensus about either. Gauthier (1999, p. 32) terms family policy a "wide umbrella of policies." As such, some serious decisions are required about what to include and exclude. These decisions cohere around four main sets of questions. The first is where to draw the boundary around family policy in the policy universe overall: which areas and measures should be included or excluded? The most widespread consensus in scholarship is to define family policy as policies associated with families with children. But what about other domains that touch on matters of private and family life? Care for older people is a prime example (this point is further developed in Chapter 14 by Dykstra & Djundeva in this volume). Why might this be seen as family policy? One possible reason is that it falls within the role of and set of responsibilities assigned to family membership and the family as a social institution. However, while true historically in some European countries, this is less and less the case now, especially legally—only in three European Union (EU) member states do families have a legally enshrined responsibility to care or provide for their elderly relatives (Hungary, Latvia, & Lithuania) (Spasova et al., 2018). Hence, on this count anyway, it seems appropriate to adopt a narrower definition of family policy—as centered around the well-being, functioning, and responsibilities of families with children.<sup>1</sup> Employment is another example. One could argue for an inclusive approach here along the lines that family and other policies are concerned with employment or affect employment-related behavior by virtue of the incentives and disincentives that are built into them, especially regarding the behaviors of parents and spouses. However, that said, family policy is not employment policy and its primary purpose arguably lies elsewhere (to financially secure the family as institution, for example, or to monitor and support child-rearing). There is an insight here about what is directly targeted by policy as against looser interconnections, which I will follow up below. The point to note now is that it calls on the analyst to be mindful that there are broader and narrower perspectives on what it is that family policy may aim to do and that family policy is always part of a wider social policy constellation.

A second and related issue pertains to the level(s) of analysis and in particular what attention to give to the vertical (as against horizontal) dimensions. The tradition in the field is for analysis situated at a single dimension or level, usually the nation state level. This closely reflects the policy world, with family policymaking centralized at nation-state level, especially in terms of cash transfers. It is a view biased by viewing family policy as cash transfers though; when one brings family services into the analysis other levels come into play. As this handbook testifies, family policies are formulated and implemented at other levels also. In addition to the nation-state level, such other levels include supranational organizations such as the United Nations (UN), the Organization for Economic Cooperation and Development (OECD) and the EU, the subnational level of municipalities, states, or regions in which people live, and the organizations in which people work (companies or firms). Each of these has an analytical purchase. However, in the European context, it should be pointed out that the EU has no direct competence in family policy—that is a member state jurisdiction. This notwithstanding, the EU has

<sup>1</sup>It should be noted, though, that some of the care literature includes both care for adults and that for children within the same framework (e.g., Saraceno & Keck, 2010).

taken action in matters that are closely related to family functioning, work– family balance, for example. The significance of this is that the transnational level should be part of any theoretical framework on family policy. So too should the possibility of decentralization within the nation-state. Elements of family policy are frequently administered at a subnational level, especially family-related services (such as ECEC). Companies or firms too are potential family policy actors, most often by virtue of their provision of services (again with ECEC as the prime example but also—and especially in the past—income supplementation to male workers with family responsibilities).

Thirdly, review of the literature also raises a question about whether to adopt what is officially defined or conceived as family policy, or to go outside or beyond states' self-representations of what constitutes family policy. The latter allows the analyst to take a more theoretically informed and even critical approach. Kamerman and Kahn (1978, p. 3) bring this issue to the fore when they differentiate between explicit and implicit family policy. To qualify as explicit, policies and programs are put in place to achieve explicit goals regarding the family and the situation of families; implicit family policy is where governmental actions and policies may not be specifically or primarily addressed to the family but which have indirect effects on the situation of families and the well-being of the individuals who comprise them. In the former, family has to be a specific policy focus to merit the label "family policy." There is a deliberateness about it with intentionality as a key factor driving policy. An explicit family policy could not exist without a strong sense of the family as a unit or institution of importance in society. This in turn rests upon acceptance of the state as a legitimate actor in regard to the family. The core meaning of "implicit," on the other hand, is that the family as such is not targeted but is envisaged to be affected by policies nonetheless. In the implicit scenario, family policy is not a recognizable policy entity, resembling more a "perspective" than a "field" in Kamerman and Kahn's terms. It seems to me that—given that family policy is a recognizable field within and across the European and other countries now (unlike when Kamerman and Kahn were writing 40 years ago)—an explicit understanding of family policy provides a rich field of analysis.

However, it is not as easy as that for we need to update what an explicit policy approach might mean nowadays. In this, we cannot regard policy as fixed or follow blindly what states or other entities consider as "family policy." The risk with the latter is that we miss relevant aspects of policy and also that the broader impact and focus will slip from view. Hence, as well as a sensitivity to explicit or implicit family policy, our understanding needs to be

centrally informed by a conception of family<sup>2</sup> (what it is and what it does). The existing literature suggests that a complex understanding would differentiate between family as a set of individuals and as a collective structure (Mätzke & Ostner, 2010). In the former, emphasis is placed on family as a set of roles and relationships; the latter emphasizes the functions performed by the family as a structure or mode of organization. The former is more novel than the latter as a line of analysis in family policy. The need to disaggregate family to focus on the individuals who comprise it comes especially from the feminist literature which has sought to move beyond the conception of family as a unit(y) of common interests among members to highlight internal processes, often associated with power imbalances between family members, that have an impact (Williams, 2004). Disaggregation is important also from a generational perspective, especially from a child-centered vantage point (Daly, 2020a). While there are limits to the extent to which this chapter can take account empirically of the degree to which family policy instruments affect individual family members, for the purposes of a theoretical framework it is important to note that it can and does seek to influence individual family relationships. The following section will demonstrate some such effects.

#### **Insights from Contextualizing Family Policy**

Family policy in Europe has a rich history (Bahle, 2008; Therborn, 2004). There is much to learn from it in terms of not just insights about practice but also from a more theoretical perspective. I undertake a brief historical overview—focusing on the policy modalities as well as variations—in order to develop insights for a theoretical perspective.

It is generally agreed that state responsibility for families developed later than other areas of social policy, especially in comparison to social policies oriented to income redistribution and securing the adult life course (Gauthier, 1996). But we need to temper this interpretation somewhat by recognizing the broader origins of family policy and the philosophical underpinnings involved. The deep roots of family policy lie especially in the institutional and legal context and how the institutions governing family life—like marriage, parenthood, and childhood—have been and are legally constituted (see Hank & Steinbach, 2019, for a recent overview). These are very long-standing. Thinking about them brings two insights to the fore. First, the legal institutions form the backdrop or underpinning to family policy and highlight an

<sup>2</sup>Note that when I say family I mean "families" in recognition of the diversity of families. For this reason also I try to avoid the use of the term "the family."

early intentionality on the part of the state associated with the regulation of the family. Second, they underline a long-standing concern with the family as a structure or mode of organization.

The interest in family as a form of organization gave rise to a number of family policy. One was cash or financial supports through taxation or vouchers which have been very dominant in the field of family policy. Among the first family policies were those offering financial assistance to families mainly in the form of income supplements for children (most widely known as "family benefits" or "child benefits"). These dated from the 1940s in many countries (and earlier in some) (Gauthier, 1996). They varied in terms of whether they compensated for all children and the degree to which they differentiated the level of support on the basis of the child's age or number of children in the family (universal or selective). Eliminating poverty and hardship among families was a widespread motivation for the introduction of child-related financial transfers. This trained the spotlight initially on the most needy sectors of society and so the first such transfers, introduced between the 1870s and 1920s, were directed at special categories of families, typically necessitous mothers, widows, and orphans. There was also a second route to the growth of financial transfers to families—employers adding subventions to wages for the fathers among their workforce. This too was selective in that only some employers engaged in the practice and it took hold only in some countries (principally Austria, Belgium, France, Germany, Italy, Netherlands, and Spain). It underlines the point made earlier about the firm as a potentially important level of family policy implementation. Over time though, national states took on the practice of supporting workers and others with the costs of raising children, either by paying civil servants with children additional wages or allowances or by introducing child benefits and/or tax allowances more universally for families with children. The interest in the family as a particular form of organization was to be seen especially in the channeling of support to fathers, which served to underpin the male breadwinner model of employed father and stay-at-home mother—the industrious father and the caring mother at home symbolizing the appropriate moral order. With the male breadwinner family as the preferred form, early social and family policies linked closely to the idea of a family wage, and provided subsidies for both marriage and the "dependants" of the breadwinner (Crouch, 1999). This, together with the widespread belief that young children should be cared for at home, made for a strong gender division of responsibilities and roles. This helps to explain cross-national and other variations in the degree to which income supports were favored over services. Countries which limited their family policy to income supports tended to support a traditional family model whereas those that offered both income and services were less doctrinaire about the best type of family model (true especially of the Nordic countries and France).

Family policy has also been concerned about protecting family-related actors and activities. Protection for mothers has been key here and also has deep roots in family policy. One of the primary policy modalities here has been income and employment protection for new mothers. Maternity benefits are the familiar policy here—combining income subsidies with employment leave for mothers while they are out of employment for childbirth. This is one of the oldest social security benefits in Europe and elsewhere, predating child income support or family income support in many locations. This is a domain of policy that has grown and expanded considerably beyond its roots (as we will see in the next section of the chapter especially). The last two to three decades have seen a notable trend to extend a range of child- and family-related employment leaves to include first parental and later paternity leaves.

It will be obvious that this goes beyond the family as a structure or mode of organization—targeting the roles and behaviors of family members. This is a counterpole (although not necessarily an oppositional one) to the focus on family as structure or institution. In some countries—such as those in the Nordic region where family is not a strong mobilizing concept—familyrelated policy sought to or was utilized to support employment and equal opportunities on the part of both female and male parents. In order to achieve this families were given access to high-quality childcare and other services as well as ensuring income sufficiency. This model did not operate with a strong or uniform concept of family as a collective or even separate institution but was more focused on individual well-being, opportunity, and equality (Ellingsaeter & Leira, 2006). Women's role and identity as workers was written into the institutions of state and market. Furthermore, while family membership might be a source of emotional stability and identity, family as a privileged social unit was much less supported as compared with other parts of Europe. Service provision for families was widespread and access tended to be anchored in social rights.

The foregoing highlights that the underlying model of family was very important historically in influencing the content and orientation of family policy. Looking at developments in context, further, draws attention on the one hand to social policy's interest in the family as a structure—as an organizational unit and a social institution—and on the other hand to family as comprising relationships, roles, and sets of responsibilities among people bound together by ties of kinship. The underlying point here is that family policy modalities potentially affect both. Income supports, for example, if given to the father perpetuate a traditional division of resources whereas if given to the mother allow for some financial independence and autonomous recognition of her role—hence policy (whether it intends to or not) affects the relationships between parents and respective roles and activities in regard to family life. All of this leads me to suggest that family policies serve two main functions: supporting/resourcing individuals/the collective unit and regulating family-related behavior and relationships. Hence, it is vital not to adopt a perspective that limits particular policy instruments to particular functions or orientations (a reason why we need to have a more critical perspective on explicit family policies). I will keep these differentiations and the many complexities to the fore as we proceed.

# **Recent Changes**<sup>3</sup>

Family policy in most countries has mainly concerned itself with supporting families with income and giving help with child-rearing. Raising children incurs both direct and indirect costs, the former arising from the additional costs involved and the latter mainly from income foregone from employment because of child-bearing and parenting. Welfare state and other policy actors have long compensated for the former but they have been much slower to take on the latter, although moves toward gender equality and work–family balance do contribute to reducing indirect costs. What is happening in each of the three main spheres of family policy today?

# **Child Income Supports**

As mentioned, historically across Europe child benefits or family allowances were the main pillar of family policy. Designed to assist families with some of the costs of raising children, they date mainly from the period around World War II. Before 1960 child income support was mainly provided through employment-based child supplements to wages (usually paid to fathers), especially in continental European countries, i.e., Austria, Belgium, France, Italy, Netherlands, and Switzerland. Once the system of public support started to be established more extensively, child-related income supports represented a truly innovative form of social right since they tended to have no conditions attached to receipt (Montanari, 2000, p. 309). Usually paid until the

<sup>3</sup>This section draws from Daly (2020a) and Daly and Ferragina (2018).

child reaches school-leaving age, some variant of these allowances exists in almost every country of the EU today, although their levels vary considerably and they are not universally available to all families. Means-testing of these allowances tends to be characteristic of the Mediterranean and post-socialist countries although the UK under Conservative leadership also introduced means-testing of the benefit by family income in 2010 (thereby undoing a considerable history of universalism in these benefits). When countries target this type of support, they tend to do so usually on the basis of income, although some countries also target by the age of the child and the number of children (often wishing to support larger families and hence encourage higher fertility). Child benefits are usually funded from general taxation revenues and in this and other ways are an expression of solidarity with people who are raising children (and therefore oriented to horizontal equity).

Looking at the last 20 years or so, there are a number of significant changes to report. While the long-term trend—that is over the 50 years from the 1970s—is for greater generosity in child income support (Daly & Ferragina, 2018), the most recent period (since the recession that set in in 2008) has seen reducing generosity in a number of countries (European Commission, 2017). In the EU, 12 member states (mainly Eastern European and Mediterranean countries) made major cutbacks and, between 2008 and 2012, spending on child and family income support fell in 21 out of 28 member states. The cutbacks are instituting significant reforms. Two trends are of particular note.

The first is greater use of targeting. This follows a strongly categorical logic and spells a change in regard to which families with children are prioritized for state support. There are strong moves in some countries (e.g., Greece, Poland, Portugal, Romania) to target support toward larger families or those on low incomes. A focus on financial need is spearheading this move away from universalism and a form of egalitarianism that includes all families. The policy thrust is a move away from supporting families regardless of size, that is away from more generic family support (Eurofound, 2015; European Commission, 2017). A second trend is toward fiscalization of financial support to families (Ferrarini, Nelson, & Höög, 2012). OECD data suggests that the average value of financial support to families through the tax system now rivals that given to families through the benefit system.<sup>4</sup> While this approach—effected through tax credits and tax allowances for example—was and still is especially favored by the liberal-oriented countries, it is increasingly found also in other parts of Europe (e.g., Austria, Belgium, Germany, and Italy). As a move away from cash support it portends a change

<sup>4</sup>See the OECD family database at: http://www.oecd.org/els/soc/PF1\_1\_Public\_spending\_on\_family\_ benefits.pdf.

in both the form and conditions under which families with children are supported financially and is at its core an attempt to more closely link child and family income support with parental earnings and employment. The behavior of individual family members becomes more important as do the links between family policy and employment. Through these moves, family as an institution is still supported but it is family as an economic unit of gainfully employed individuals that is targeted.

## **Early Childhood Education and Care (ECEC)**

This type of provision has become much more important over time, associated with a mix of push factors. These include the growth of children's rights, countries' elevated concern with the educational performance and general behavior and achievements of their young populations (especially linked to the perspective of social investment, e.g. Hemerijck, 2015) and also policy's interest in greater gender equality. The provision of daycare and education for children is the classic policy response here. However, looking across countries and periods of time, this is not a uniform provision in that some countries viewed it as care for children whereas others saw it more in terms of education and early learning (see also Vandenbroeck in this volume). Taking a historical look at Europe as a whole, minimal ECEC provision existed until after World War II at which point the national and local authorities started to assume a responsibility to provide ECEC (Moss, 2006; Scheiwe & Willekens, 2009). From the 1960s to 1970s on, and especially in the last two decades, what was a stream has become a major arterial flow. Here we see the full range of policy actors or levels—transnational, national, local/municipal, and employer—at play.

In terms of changes and transitions, three general trends are noteworthy. First, there has been a general move in the direction of the educational model. Moss (2006) suggests that this took place in two waves: the 1970s and the last decade or so. The Nordic countries were to the fore in the first wave. While ECEC in Scandinavia has its roots in the care or welfare paradigm, these countries led the way from the 1970s on in moving toward universal ECEC provision that is strongly pedagogical in orientation. An educational model epitomized in ECEC as the term now used to describe the field—is becoming much more the norm as countries move away from the notion of services as caring for children and toward educating them and developing their abilities at the earliest opportunity. Second, one of Europe's strongest trends in recent years has been a growth in out-of-home ECEC for very young children (up to age three) (European Commission, 2013). Here cross-national agencies have been very important. In Europe this has been led by the strong stance taken by the EU in the Barcelona targets set in 2002 which aimed for 33% of the zero–two-year-old cohort and 90% of those aged between three and six years in ECEC by 2010 (ibid.). The OECD has also been a leader, its message generally dovetailing with that of the EU (Mahon, 2010). A third, related, trend is a growth of guarantees to ECEC for children. While such guarantees are often rhetorical or abstract, some seven EU member countries now guarantee a legal right to ECEC for each child under two years, often immediately after the end of parental childcare leave (Eurydice, 2019). These are Denmark, Estonia, Finland, Germany, Latvia, Slovenia, and Sweden. In most of these countries, the entitlement usually implies a full-time place. Other countries start the guarantee later. In Belgium, France, Luxembourg, Hungary, Malta, Spain, and the UK, a place in publicly subsidized ECEC is guaranteed from the age of three or a little earlier. So, over half the EU member states grant children a right to ECEC. Fourthly, it is important to note that the trend toward ECEC expansion for young children is not unequivocal as in some places it is accompanied by a choice-oriented introduction of home care allowances (e.g., Finland, France, Norway, Sweden, Germany) (Lohmann, Peter, Rostgaard, & Spiess, 2009).

Read through the perspective developed earlier we see a move toward a greater focus on the resources available to individuals (especially children) and the shifting of some child-rearing outside the home (although the responsibility remains with parents and the family more generally). The fact that some countries are drawing back from this and have two approaches in play that is offering services but also giving incentives for children to be cared for at home—underlines some ambivalence and a tussle between making child development a public good as against supporting and buttressing the family as a unit and location for child-rearing.

#### **Parental and Maternal/Paternal Leaves**

Historically, maternity has been the main focus of provisions oriented to parents. This continues to be an important plank of provision but the innovation nowadays is generally elsewhere: in other types of family-related leave. From the 1990s on parental leave has become more important and more widespread, generalized as a norm of modern welfare state provision and endorsed especially by the EU and the OECD. Essentially, there has occurred a two-fold shift in focus: from mothers to parents, and from mothers to fathers.

Increasingly now, countries offer parental leave which is available to both mother and father. The matter of the division of the leave between the parents and whether to prescribe certain portions of it for the mother or the father, part of a general attempt to increase fathers' uptake of the leave, is an important line of innovation and development in these leaves. The core trend has been for parents being left to themselves to decide how they will utilize and divide the parental leave between them. There is a countertrend though in that sometimes, a certain period is set aside for the parent who is not on leave; an individual, non-transferable right—the so-called "father quotas" was pioneered in Norway to encourage fathers to take child-related leave and has taken firm hold in the other Nordic countries. In addition to maternity and parental leave, most European countries now offer father-specific paternity leave which is job-protected leave provided to the employed father, normally to be taken near the time of childbirth in order for him to spend time with his newborn child and offer support to the mother and any other children in the home. Most widely, it is of short duration—a number of weeks—and paid at a flat-rate. The motivations for its introduction vary but it is at root an attempt to improve gender equality between parents and also to enable the new father to have a bonding period with his newborn.

The situation now is that almost all countries in the EU have three types of such leave available: maternity, parental, and paternity (European Commission, 2017). To summarize a complex set of developments, EU countries have seen moves to consolidate maternity leave, expand parental leave, move to the couple as the unit of entitlement rather than the individual parent(s), and increase the availability of paid leave to fathers (either through extending paternity leave and/or making a portion of the parental leave attractive to fathers) (Blum, Koslowski, & Moss, 2017). The implications for parents are obvious, those for the construction of childhood and the treatment of children less so. Here, we see a focus on family-related behavior rather than structure through the greater engagement by the state in how children are reared and the respective roles and investments of the two parents. We could read the current policy consensus as assuming that child well-being and a "good childhood" are best secured by: (a) having both parents present in the first month or two, (b) being cared for at home by the mother with some input from the father for about another year, and (c) having access to an increasing volume of out of home ECEC from the age of 1 on.

## **Conclusion**

Family policy, at its most developed in Europe among world regions, refers to state policies oriented to the welfare of families with children and the support and regulation of the family as an institution and a way of life. While societies have their own specific histories on how they organize their family policy (and indeed how active and recognized it is as a policy field), the main policy modalities that exist across countries are in the form of cash benefits or tax allowances for families with children (most widely to help them with the costs of rearing children), services oriented to early childhood care and education (committed to a variety of goals, including children's development and education and also support of parental employment) and a series of leaves from employment for child-related reasons (also intended to be multifunctional but especially oriented to the organization of employment and family life of parents).

As this whole chapter has shown, family policy is a dynamic area of policy within and across countries. During the last 20 years or so, not only have governments become more active on family policy but new areas of policy have been added and developed and existing provisions fundamentally reformed. There is a move toward a more diversified set of family supports wherein cash benefits sit alongside tax allowances, services, leaves, and employment-based measures. Again with cross-national variations, it is generally the case that the field is broadening and deepening and services are becoming a larger part of the redistribution toward families. This reflects especially a greater concern with children's early education, development, and well-being. As they have been developed through family policy, concerns around children's development and resources have seen both greater access for (especially young) children to early education services and also targeting of income support on the most deprived families. We might conclude from this that relationships and resource distribution among generations are increasingly foregrounded (Daly, 2020a).

Policy concern with the family as an institution or structure remains strong and is even growing in some countries, although the degree of support for a traditional family form is generally declining and it is now the characteristics of the family as an economic unit (rather than a particular physical structure such as based on marriage) that dominate. Hence, policy continues to reproduce the family, but it does so in somewhat different ways as compared with the past and arguably also more directly. Consider as an example the sense of a greater degree of regulation of child-rearing and the employment-related behaviors of adult family members (Papadopoulos & Roumpakis, 2019). The notion that family policy may now have a more instrumentalist, economic cast—as against resourcing the family as having moral authority (which was true especially in some of the continental European countries in the past)—is also worth considering.

A further trend is a general move to balance or "reconcile" work and family life (Lewis, 2009). This has had two main expressions in policy. On the one hand, employed parents are given increased incentives and support to take time off work to care for their children when they are deemed to need it. Both parents can be targeted in this regard an attempt to affect the share of unpaid work done by each but there is increased interest in encouraging fathers to take (short periods of ) time off work to care for their very young children. On the other hand, there is a general move toward activating people to be employed so as to raise the level of gainful activity (and associated sense of responsibility). It will be obvious that there is a certain tension if not contradiction—between the two types of 'reconciliation'. Some have seen a 'feminization' of the male life course (Esping-Andersen, 2016) but to me the father-oriented measures (outside of Scandinavia especially) are too weak to be other than a symbolic policy (Daly, 2020b).

I opened this chapter by considering how to conceptualize family policy and I want to take up that matter again here, in line with the chapter's overall aim of developing a framework for the analysis of family policy in this volume and elsewhere. There are a number of first principles of relevance it seems to me. The first is that the two-fold framework of family as structure/unit and as a collection of individuals has wide application for the analysis of family policy. It is also clear, though, that these rather crude categories need to be more finely calibrated so as to pick up nuance in both policy and its reform. Generally, I tend to agree with Papadopoulos and Roumpakis (2019, p. 249) that we need to move beyond narrow conceptualizations of the family and engage with the family as a collective socio-economic agent. Second, while the policy details are complex, essentially the analysis confirms that it is meaningful to regard family policy as exercising two main functions: resourcing families and regulating them (with both understood as matters of degree). Third, it seems essential to contextualize family policy in a broader societal context, by understanding family as a social institution in which various actors are invested. While we have not examined the agency or implementation of family policy in detail here, the analysis tends to confirm the editors' contention that the agency in family policy development potentially operates at four levels: cross-national, national, subnational, and at the level of the firm or company.

## **References**


perspectives In K. Scheiwe & H. Willekens (Eds.), *Child care and preschool development in Europe* (pp. 1–22). London: Palgrave Macmillan.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Part II**

**Supra-national**

# **3**

# **Beyond the National: How the EU, OECD, and World Bank Do Family Policy**

#### **Jane Jenson**

In recent decades, numerous international organizations adopted positions that use components of a policy frame familiar from family policy at the national level. They sought to advance one or more of three classic goals of that domain: stabilizing demography, ensuring income security, and supporting parents' labor force participation (Saraceno, 2018, p. 443). The policy instruments promoted included income transfers, services, and leaves (Daly, in this volume). They were justified variously in the name of equality, social development, or social investment. This chapter examines three such organizations. The Organisation for Economic Co-operation and Development (OECD) and the World Bank are standard international organizations (IO), being intergovernmental. They are composed of independent member countries and they have no institutions to impose policy choices, although each has a range of policy tools that can compel, induce or encourage their members to act in particular ways. The European Union (EU) is not a standard IO. While it does have substantial intergovernmental dimensions, it is also in significant ways a supranational organization. It is a grouping to 27 (28)1 member states that have transferred a measure of sovereignty to European-level institutions, thereby providing them some legislative,

J. Jenson (B)

© The Author(s) 2020

<sup>1</sup>With the withdrawal of the United Kingdom from the European Union, the number of members drops to 27.

Université de Montréal, Montréal, QC, Canada e-mail: jane.jenson@umontreal.ca

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_3

executive, and regulatory capacity. This transfer touches, moreover, on policy domains important to families and family policy.

Family policy has had moveable boundaries. Goals have changed over time and instruments sidelined. Concerns about demography, whether family planning or pronatalist initiatives, have come and gone as an objective. More permanently sidelined is the principle of redistributing economic resources horizontally from families without children to those raising children. That goal underpinned post-1945 family policy for decades, and generated many countries' choice of universal family allowances to supplement family income. The objective of interfamily horizontal redistribution is currently much less present, often replaced by the goal of vertical redistribution from rich to poor families and means-tested benefits intended to reduce disadvantaged families' risk of poverty. Finally, there is an on-going shift from income maintenance for male breadwinners to measures promoting labor force participation of mothers and shared family care (Bonoli, 2013).

This chapter sets the boundaries of analysis by concentrating on these three goals—stabilizing demography, family income security, and parental labor force participation. It focuses in particular on the expressed policy goal(s), the targets, and policy instruments highlighted by the organizations and the policy frame used to justify each. In the past decades, each organization has modified its position on all three as well as weighting each differently. The chapter proceeds as follows. First, it provides a brief overview of the three organizations, setting out the policy capacity of each. The next three sections examine the position of each with respect to the three goals. A brief conclusion offers some general discussion of differing trajectories but a shared process leading to *non*-*familialization*.

# **Purpose, Competence, and Actions of Three International Organizations**

The OECD, the World Bank, and the EU have achieved the authority and independence to identify and frame policy problems, using their claims to expertise to prompt their preferred responses. This chapter adopts the analytic stance that ideas and actions of IOs result from the behavior of policy advocates and entrepreneurs within them as much as or even more than from outside pressure. While there is no space in this chapter to trace the influence of internal advocates and entrepreneurs, the premise implies that IOs never simply act at the behest of their principals (for discussions of this literature, see Ellinas & Suleiman, 2011; Mahon & McBride, 2008; Weaver, 2007).

One of a number of international organizations created in 1944 to ensure postwar reconstruction in Europe and Asia, the World Bank's official name is International Bank for Reconstruction and Development. Over the next decades, however, its mission of reconstruction fell away, and it now seeks to be the world's premier development institution, its credo being "our dream is a world free of poverty." There are two vectors to the Bank's work. One is operational, involving the elaboration of country-specific projects. The Bank provides financial resources via loans for project or policy directions and technical expertise to governmental and other agencies, usually with conditions about policy design and implementation. Decisions about such loans require large numbers of Bank staff in both Washington and the field (King, 2007, p. 161). The second vector developed later because "… no systematic research program existed until the early 1970s" (King, 2007, p. 162). The disciplinary training of this research staff has expanded slightly, including both some noneconomists and a wider range of policy expertise, including education and health economists, among others (King, 2007). The Bank's work nonetheless remains heavily influenced by economists and their analytic frames (Rao & Woolcock, 2007).

The OECD's Convention, signed in Paris in December 1960, commits it to promoting policies meant to achieve the "highest sustainable economic growth and employment and a rising standard of living in Member countries, … thus to contribute to the development of the world economy …."<sup>2</sup> This international agreement gave the OECD "… no regulatory capacity, no independent source of funds, no money to lend, and no instruments within its control" (Wolfe, 2008, p. 28). Nevertheless, Article 3 of the Convention did provide a pillar on which the OECD has built its intellectual leadership; it requires member countries to furnish the information and data necessary for the staff to undertake analytic work. Building on these data and the possibilities for analysis and comparison they provide, the OECD has flourished as a knowledge organization with a large research capacity dominated by economists (Mahon, 2019, p. 3). It has refined techniques for the surveillance of member countries' economic performance and outcomes and provides its members with standardized comparative data, "… aimed at shaping the ideas and preferences…" (Leimgruber & Schmelzer, 2017, p. 24). It is also a forum for exchanging ideas and policy prescriptions across the international community of experts that circle around the staff, their meetings, and their consultations. Without the possibilities of but also the constraints from

<sup>2</sup>The text of the convention can be read online: https://www.oecd.org/general/conventionontheorgani sationforeconomicco-operationanddevelopment.htm (Last accessed on March 27, 2020).

operational projects that restrain the World Bank, the OECD can act with suppleness to refocus its analytic lens quickly (Jenson, 2017a).

The European Union has greater authority over its member states than any standard IO; indeed, the EU is in large part a supranational organization. It has legislative power and a court that can impose conformity to Union law. Bureaucratic structures tasked with action encase its research functions. Moreover, the large bureaucracy is composed of citizens of all member states with a wide range of training and education that tends to follow the traditional expertise of each policy domain as well as varying by national traditions of postsecondary education (Ellinas & Suleiman, 2011). Since the 1957 Treaty of Rome, the mix of supranational and intergovernmental competences has altered and the influence of the EU in additional policy areas has expanded considerably. Social policy, and thus family policy, is one such area. This expansion occurred over a number of decades, as European institutions pushed the definitional boundaries of their treatygranted competences (Anderson, 2015, pp. 4–5). Within the general area of social policy, the EU deploys several governance practices: "Social regulation means setting legal standards and norms in social policy at the EU level. Hard coordination is EU-facilitated policy coordination that is legally binding and that involves high surveillance and enforcement of member state policy. In contrast, soft coordination is EU-facilitated policy coordination that, while including surveillance, involves weak enforcement" (De la Porte, 2018, p. 478). As we will see below, while some of the components of family policy use the tools of social regulation (for some policies governing parental employment), most involve soft coordination instruments for shaping labor markets and welfare states. Diffusion of ideas within policy communities is also an aim, with these communities defined to include social partners (employers and unions) as well as civil society organizations (De la Porte, 2018, pp. 479, 482–483). Despite much activity, the EU does not "make" much social or family policy, however. It influences national decisions, to be sure, via its diffusion of learning, by its fiscal decisions, and by its limited regulatory authority, but the power to choose, establish and maintain the bulk of social policy remains primarily a national prerogative. The EU's trajectory in the domain of family policy is, therefore, complex and opaque.

# **The EU: "An Actor Without a Role"**<sup>3</sup>

In the treaty-based division of powers in the European Union, the principle of subsidiarity continues to apply to most of what might be family policy, meaning that it remains a national or subnational competence (Anderson, 2015, p. 4). Moreover, member states carefully guard their policy autonomy in the field. Nonetheless, the EU has a long history of acting on objectives of demography, income security, and parental employment, often undertaking these actions without a family policy frame, however.

Substantial consequences for both income security and parental employment, for example, follow from legislative and judicial decisions around the fundamental right to freedom of movement within the Union, one of the four freedoms at its heart. European institutions have developed family law and policy that constrain member states' treatment of migrant families, including the rights of all family members, and the social benefits to which they are entitled. Such interventions arose because promotion of workers' mobility after the Treaty of Rome immediately raised matters touching on the rights of other family members, both adult and children, to reside with the worker in another member state and their access there to social protection benefits and services. Regulations and court decisions extended the right of residence in another member state to accompanying family members, including a spouse, children, and other dependent relatives. Subsequent years brought regulations granting migrants access to the same social and tax advantages as nationals of the member state (for a list see Anderson, 2015, p. 90). Attention also turned to child support and custody in cases of divorce involving several member states (Stalford, 2002). All of these issues, informed by more or less traditional understandings of family, were decided within the framework of the right to freedom of movement upon which the parents' employment and the family's income security rested.

The EU began to sidle further into the field of family policy with the Social Action Programme in 1974, using the door provided by its competences over labor markets. In the name of achieving "full and better employment in the Community," among several objectives listed were actions to achieve equality between women and men and "to ensure that the family responsibilities of all concerned may be reconciled with their job aspirations" (Ross, 2001, p. 180). These actions' target was women, and particularly their equal pay and protection from discrimination because the equal treatment of women and men rested on solid legal grounds, from both the Treaty of Rome (article 119)

<sup>3</sup>Title adopted from Ross (2001).

and key Directives in the mid-1970s (Hantrais, 1995, p. 80; Jacquot, 2015, pp. 31–35). Over the next years, the policy linkage tightened between equal opportunities and access to childcare services and leaves in order to allow mothers to take up and retain employment.

The Equal Opportunities Unit housed in Directorate-General V (dealing with employment and social policies, now Directorate-General for Employment, Social Affairs and Inclusion DG EMPL) pushed for childcare and leaves, setting up an expert network in 1986; in 1991 its name became Network on Childcare and Other Measures to Reconcile Work and Family Responsibilities. The Equal Opportunities Unit and expert networks around it framed better childcare and improved leaves as policy tools for promoting gender equality in family relations. Greater sharing of family responsibilities was the mechanism to work social change (Ross, 2001, pp. 180–187; Stratigaki, 2004). Given the treaty-based requirement to stay resolutely anchored to the employment field, the measures for reconciliation of work and family remained firmly focused on parental (usually seen as mothers') employment.

In the 1980s, however, the EU began to consider developing an explicit family policy, the entry point being demographic challenges, conceptualized as threats to the functioning of labor markets but also to families' income security. In a 1983 Resolution the European Parliament called for family policy to become "an integral part of all Community policies" (quoted in Hantrais, 1995, p. 80). Policy research and discussion began within various agencies, driven by the widespread demographic concerns of member states.4 The 1988 European Council established high-level consultations with officials from member states; the next year the Commission released its *Communication on family policies* (for more of this story see Ross, 2001, pp. 188ff.). The document presented a broad and ambitious analysis, "because as children are becoming more rare, the demographic future of Europe rests with the family" (EU Commission, 1989, p. 12). It began by painting the "disturbing demographic situation in Europe" and ended with a portrait of the "fundamental role of the family as the basic unit of society" and important for intergenerational solidarity. Between these two visions was analysis of family transformations reshaping demography and labor markets, particularly rising rates of women's labor force participation and marriage patterns. The communication justified EU-level advocacy of childcare services—"a key component of family policy." The Communication's purpose was to make a case for Europe deploying a family policy frame for new and "feasible" action

<sup>4</sup>Demography troubled the EU in the 1980s but member states—and demographers—had pointed to falling birth and fertility rates from the mid-1960s and the end of the baby boom (Jenson & Sineau, 2001, pp. 21–22). For contemporaneous data see Hantrais (1997, pp. 346–347).

(EU Commission, 1989, p. 3). The Social Affairs Council that received the report toned it down somewhat, however, pushing the analysis of childcare back onto "reconciliation" of work and family (and thus the EU's mandate for employment) rather than framing it as a tool for changing gender relations in the family as the Commission had done. It agreed the EU might take action with respect to freedom of movement and equal opportunities, but otherwise the EU would engage in information gathering and diffusion. The European Observatory on National Family Policies began data gathering in 1989 but, as its name indicated, EU actors were well aware that profound ideological differences existed among member states and that they were not about to cede the policy field to the Union.

Nonetheless, European actors moved decisively onto the social policy field in conjunction with the 1992 Treaty of Maastricht (Daly, 2008). While doing so, they continued to identify improving the demographic situation as integral to the EU's purpose because of the threats from both aging (and early retirement practices) and falling fertility.5 The Commission published several major reports on demography over these years (EU Commission, 1994, 2006). Well into the new century, European institutions highlighted demographic challenges (falling fertility, aging populations, more diverse family structures) as policy drivers. Policy treatment remained primarily confined, however, to areas of European competence over the operation of labor markets and equal opportunities. The EU pushed policies directly related to childbearing (leaves and childcare) as well as those demographers assume have indirect effects on fertility, such as gender equality and working time (Daly, 2020; Nieuwenhuis, Need, & van der Kolk, 2012, Chapter 6).

Through the years, the European institutions prescribed support for parental employment. The key theme initiated in the 1980s and worked on through the decades remained supporting parental employment in ways that would permit "reconciliation," particularly for women (for the documents and their summaries see Hantrais, 1997, p. 340; Jenson, 2008; Ross, 2001, pp. 14–15). The EU used its treaty-provided tools assiduously. Instruments of social regulation were deployed in a number of directives (pregnant workers and working time' 1990; parental leaves, 1995) as well to prevent gender-based discrimination. Soft coordination applied to the key service identified as enabling reconciliation of work and family—childcare services (De la Porte, 2018, p. 479). The Lisbon Agenda in 2000 and subsequent

<sup>5</sup>Article 7 of the protocol on social policy annexed to the treaty instructed the Commission to report annually "… on progress in achieving the objectives of Article 1, including the demographic situation in the Community."

declarations set targets for women's employment rate and for childcare services.

Nonetheless, as many have observed, the EU often simultaneously deploys more than one paradigm, often promoting the same policy instrument to attain multiple objectives (Knijn & Smit, 2009 for example). These alternative perspectives exist in flexible compromise. When the 2000 Lisbon Agenda opened the third "growth spurt" for European social policy, a paradigm that gained prominence diagnosed social exclusion as the main challenge (Daly, 2008, p. 3). The Lisbon Agenda announced the need to "modernize" social policy to make benefits sustainable and promote social inclusion. The policy sources and actors driving this approach were not the same as those that had led the thrust for quality childcare and parental leaves, a push that had come from the Equal Opportunities Unit and the expert, often feminist, networks around them. The Lisbon Agenda's take on social exclusion had different roots: "… the multi-faceted understanding of social exclusion developed through the EU-funded poverty programs and Observatory on Policies to Combat Social Exclusion during the 1980s and 1990s, a turn to activation on the part of the EU (as expressed especially through the European Employment Strategy), and the strong promotion of mobilization by the 'poverty sector'…" (Daly, 2008, p. 6). Adoption of a social inclusion lens was the signal that a social investment strategy was taking shape, and attention to equality both between women and men and across classes was on the sidelines (Cantillon & Van Lancker, 2013; Jenson, 2009; Saraceno, 2015).

One of the foundational documents of the social investment perspective, *A New Welfare Architecture for Europe,* prepared for the Belgian presidency in the second half of 2000, sought to concretize principles of the Lisbon Agenda and its lens of inclusion while also incorporating the ambient demographic concerns. In this perspective, if social policy did not succeed in "harmonizing motherhood with employment," three negative consequences were likely: higher child poverty; "labour shortages or, alternatively, a shortage of births"; and "wasted human capital" when well-educated women without access to nonfamilial childcare limited their labor force participation (Esping-Andersen, et al., 2002, pp. 9–10). As the social investment perspective developed through the 2000s, not families but children became the target. In the Commission and Council's annual *Joint Reports on Social Inclusion*, over time "poverty in general has been almost excised," while child poverty emerged as a "strong issue" (Daly, 2008, p. 10). Indeed, combatting child poverty became a highlight of the European Union's 2008 *Renewed Social Agenda*, under both the heading "Children and Youth – Tomorrow's Europe" and "Combating Poverty and Social Exclusion" (Jenson, 2009, p. 459). When the *Social Investment Package* was finally issued in 2013 it called for breaking the "intergenerational transmission of disadvantage" and actions to "prepare" not "repair." With its life-cycle approach, children were a central focus and childcare was in the first instance for them and only secondarily for parents (Jenson, 2017b, p. 277). "Access to early childhood education and care (ECEC) has positive effects throughout life, for instance in terms of preventing early school leaving, improving employment outcomes, and facilitating social mobility. ECEC is key to addressing challenges faced by disadvantaged children by providing early intervention. Furthermore, it is essential in removing barriers to the labour market participation of parents" (EU Commission, 2013, p. 21). In this major document, families received little attention except to note that earning adequate income was a challenge for families and they needed family-friendly workplaces and employment policies.6

By 2018 when the EU laid out a *Pillar of Social Rights* in 20 principles, attention to families again was hard to find. The only mention that Europeans might live in families came with respect to income security, in the form of the need for "adequate minimum wages" (principle #6). Women and men had the right to equal treatment and equal pay while "all groups" must have equal opportunities (principles #2 and #3). Then in what was a fundamental turn away from representations of childcare in earlier decades, it became an intervention *only* for the young: "children have the right to affordable early childhood education and care of good quality" and "the right to protection from poverty," while children "from disadvantaged backgrounds have the right to specific measures to enhance equal opportunities" (EU Commission, 2018). A major policy instrument for family policies for years would no longer serve its double purpose of supporting children's development *and* the employment of their parents. This discussion of non-parental care was also silent on any goal of changing family relations.

Yet at the same time and in parallel, the Commission announced a directive on work–life balance for parents and carers that returned to familiar and solid ground with respect to parental care (Daly, 2020, Chapter 7). It picked up principles #2 and #3 of the Pillar, and promised better implementation of the principle of equality between men and women with regard to labor market opportunities and treatment at work (EU, 2019). The directive extended rights for paternity leave, parental leave, and carer's leave. Thus, alongside the Pillar that focused on individuals and children was a parallel

<sup>6&</sup>quot;Families" and "family" were rarely mentioned in the Social Investment Package, while "child" and "children" received numerous mentions (EU Commission, 2013). The quote here is the only mention of parents.

move finally to address the needs of adults caring for children and others, improvements intended both to encourage sharing of care between women and men and to maintain rates of labor force participation. The directive was firmly anchored by the EU's traditional frame and competence for economic growth: "… this is not only about strengthening the rights of individuals. The new rules are a model for how to align social and economic priorities."7

# **The OECD: The Work–Family Nexus**

In 2017 the Secretary-General of the OECD gave a speech in Seoul titled "Better family policies can help combine work and family commitments: Lessons from OECD countries" and another in 2018 in Montreal, "Familyfriendly policies a key driver of economic growth."8 The OECD's mission is large—"to promote policies that will improve the economic and social well-being of people around the world"—with economic growth being the primary goal. Yet it has none of the EU's institutional capacity to legislate or to regulate. Indeed, the OECD *does not have policies*; it has positions and analytic frames that allow it to promote some policies and instruments. Both the Secretary-General's recent speeches singled out good practices in some member countries, a familiar strategy in the OECD's provision of intellectual leadership in processes of social learning (Wolfe, 2008, pp. 25, 41). This embrace of family policy instruments to achieve the IO's economic growth mandate had been a long time coming. It is only in the new millennium that family-friendly or family policies received significant analytic attention, with the aim of increasing employment rates and reducing poverty, especially among lone-parent families. Given this context, this section tracks when and how the OECD interpreted the three classic components of family policy as germane to its mission.

Fully a decade and half after its creation, the OECD established the Directorate for Employment, Labour and Social Affairs (DELSA) in 1974. It convened a first meeting of social affairs ministers only in 1988. DELSA is an overarching directorate that does statistical and analytical work on social protection within sections covering employment, social policies, migration, and health. In the 1980s, when the EU was creating space for social policy

<sup>7&</sup>quot;Work-life balance: Commission welcomes provisional agreement" https://ec.europa.eu/social/main. jsp?langId=en&catId=1311&furtherNews=yes&newsId=9285 (Last accessed 27 Mar 2020).

<sup>8</sup>For the first, see http://www.oecd.org/els/family/better-family-policies-can-help-combine-work-andfamily-commitments.htm (Last accessed 20 Apr 2020). The second is at https://www.oecd.org/new sroom/family-friendly-policies-a-key-driver-of-economic-growth.htm (Last accessed 20 Apr 2020).

action and even explicitly contemplating *doing* family policy, DELSA was still attached to its "welfare as burden" stance that shaped its analyses (and others' appreciation of it as fundamentally neoliberal). By the early 1990s, however, OECD analysts began to vaunt social expenditures' role in maintaining social cohesion and even serving as investments (Deacon & Kaasch, 2008, pp. 227–229; Jenson, 2017a, p. 212).

The report *New Orientations for Social Policy* (OECD, 1994) participated in the widespread reboot of thinking of social policy from income maintenance to instruments to facilitate labor market participation (Bonoli, 2013). Social affairs ministers deliberated on this new approach in 1992, hearing that the socioeconomic context was one of slow growth, labor-market problems, economic insecurity, and tight budgets for social programs. The search for effective and efficient social policies would have three challenges different from those of the post-1945 years: population aging, less familiar family structures, and continuing high rates of women's employment. Moreover, the difficult economic conditions of high unemployment and poverty meant "young families – which are often over-represented among those with low incomes – are finding it difficult to fulfil all their aspirations for their children." In other words, the demographic situation was both a risk for and a consequence of the economy. The major challenge the OECD saw was effects on pension sustainability and care for the vulnerable elderly. These concerns drove most of the search for new orientations. Nonetheless, the gesture toward "aspirations" was a reminder that falling fertility was troubling, and the ministers approved "a review of the wide range of policies in support of families and children."<sup>9</sup>

Four years later the high-level conference *Beyond 2000: The New Social Policy Agenda* brought ministers together with invited experts. The choice of experts guaranteed that the ministers would hear about the components of family policy, from fertility through parental employment and adequate income. Gøsta Esping-Andersen, as he would four years later for the Belgian Presidency of the EU, argued that "family and labour market policies can no longer be based around the notion of the male breadwinner; such a framework fails to take account of the needs of two-earner households, and can therefore even inhibit family formation and functioning" (OECD, 1997, p. 9). Chiara Saraceno (ibid.) analyzed three shifts in family behavior with consequences for social policies—women's rising labor force participation rate; higher rates of unstable marriages; and declining fertility rates. All three put pressure on traditional family practices of care.

<sup>9</sup>See the report of the meeting of the Employment, Labour and Social affairs Committee at ministerial level, Paris, 8 and 9 December 1992. SG/PRESS(92)94.

The emphasis on care continued with the next meeting whose theme was a "new social policy agenda for a caring world." The draft document sketched a troubling situation on its first line: "The demographic challenge to social policy arises from declining fertility and greater longevity. The key point is that there will be fewer people of working age for every person who is retired" (OECD, 1998, p. 2). Pension and health costs were still driving concerns but so was the need for employment-oriented social policies. A lifecourse approach to social policy was deemed appropriate. Drawing directly on Saraceno's earlier analysis, a short section on family formation declared the male-breadwinner model of family relations "outmoded" and enumerated restructured fertility and marital relations, including the increase in loneparent (or what the OECD terms single-parent) families (OECD, 1999, pp. 14–18). With the life-course frame came attention to early childhood interventions of many kinds—from childcare to visits from social workers, targeted to the most disadvantaged, as in the United States from whence the data came (OECD, 1999, pp. 84–86). Data analysis also identified different cross-national patterns in mothers' labor force participation and the effects of social policies on lone parents' employment. All this appeared beneath the heading "reconciling paid work and unpaid work," a formulation in line with the notion that care (unpaid work) was a key challenge. The section focused on leaves and childcare (OECD, 1999, pp. 88–90). Later discussion of policy analysis for "adapting to family developments" hemmed and hawed about a number of instruments and interventions that might be used to keep social protection costs down (OECD, 1999, pp. 135–136). Childcare services were a possible policy instrument, but the OECD's major concern was the "problem" of lone parents: their poverty and the "inconsistent policy signals" some governments gave by subsidizing them to care and not requiring them to seek employment. In addition, family planning and sex education to avoid teen pregnancies were positively mentioned (OECD, 1999, p. 136). Consideration of care for the frail elderly appeared in sections on health policy, however, thereby sidelining unpaid care by family members. The OECD was still having trouble making the analytic leap from aging societies and dependency ratios to a structured approach to family policy that would consistently support employment and ensure adequate income.

By the start of the new decade, the OECD's stance on family matters was taking clearer shape. DELSA was in the midst of the review of the wide range of policies in support of families, called for a decade earlier. One product was the five *Babies and Bosses* volumes providing a review of the reconciliation of work and family life and the implications for social and labor market trends. In 2007 the final report synthesized the approach for the OECD's main business: "If parents cannot achieve their desired work/family life balance, not only is their welfare lower but economic development is also curtailed through reduced labour supply by parents. A reduction of birth rates has obvious implications for future labour supply as well as for the financial sustainability of social protection systems. As parenting is also crucial to child development, and thus the shape of future societies, policy makers have many reasons to want to help parents find a better work/family balance" (OECD, 2007, p. 217).

Throughout the *Babies and Bosses* series DELSA remained firmly focused on labor force participation while combining all three of the components of standard family policy, although the preferred vocabulary was "familyfriendly" policy (for example OECD, 2007, p. 16). Instruments promoted were publicly supported (albeit not necessarily provided) childcare, parental leaves of a length that would not impede a return to employment and with a design to encourage sharing, various tax measures and some transfers, and flexible workplace measures. The synthesis report criticized policies that allowed, even encouraged, lone parents to remain out of the labor force, retaining the idea of "moderate benefit sanctions" for not seeking work, an idea that had been part of the OECD's toolbox since the 1990s (Mahon, 2008, p. 265; OECD, 2007, p. 20).

Parallel to DELSA's work on the *Babies and Bosses* series was the Education directorate's *Starting Strong* analyses, sparked by the 1996 meeting of education ministers whose theme was "Making Lifelong Learning a Reality" and whose final communiqué prioritized improving access to and quality in ECEC. The series was an important element of the OECD's move to a childcentered social investment approach (Jenson, 2017a, p. 213). Located in the Education directorate, the resulting analyses in five reports focused on the supply, forms, and training for ECEC. Parental employment was simply a driver of demand while the next generation's reduced poverty and boosted social mobility could result from successful programming (Bennett, 2006; OECD, 2017). The focus on child protection and child well-being continued in *Doing Better for Children*, which called for careful and planned investments across the "child life cycle." The document deployed a "well-being" lens, reviewed the evaluation data for a wide range of policy interventions and recommended concentrating spending early, even in utero, and on at-risk children in order to change the intergenerational transmission of disadvantage (OECD, 2009, pp. 178–183).

DELSA's work on supporting families and promoting family-friendly policies resumed in *Doing Better for Families*, returning to demographic concerns and family income (OECD, 2011a). Recommendations for interventions targeted making work pay and supports for work–family balance as well as early interventions especially for young children at risk of developmental delays and socio-emotional difficulties. This analysis also underpinned work on indicators and data development, including the creation of huge databases on family policies and outcomes (Adema, del Carmen Huerta, Panzera, Thevenon, & Pearson, 2009). Then, two other analyses reported simultaneously. The International Futures Programme that usually tackles issues like space and risks from climate and other disasters, focused on families in the next decades (OECD, 2011b). The report examined large drivers of change and the challenges generated, and pointed to the need for preparation, explicitly framing issues as ones of family policy, in part via the composition of its steering group. Its "taking the long view" section identified a likely scenario of continuing constraints on public finances that would generate pressure for replacement of universal services by targeted ones and a reassignment of responsibilities from the public to the private or community sectors and among individuals, the family and the state (OECD, 2011b, pp. 35–36). A second initiative that continued the emphasis on contributions of effective family policies to economic growth was the organization-wide "gender initiative" launched in 2010. Among policies promoted in the areas of education, employment, and entrepreneurship were familiar "family-friendly" ones such as childcare and parental leaves, including incentives for parental sharing of care in the name of equity (Adema, 2014, pp. 124–125).

All of these efforts typifying the OECD's mandate to "promote policies" lay behind the Secretary-General's two speeches cited above. They continue to inform the work of this IO as it develops a new overall strategy for inclusive growth in response to rising inequalities worldwide (Jenson, 2017a, p. 215). Of the three international organizations examined here, the OECD remains most firmly anchored to a consideration of the three elements of family policy. But this may change as it shifts its analytic lens in the next years.

# **The World Bank: Population, Women, Children… and Families?**

The World Bank took longer than the other two IOs to pursue sustained analytic attention to social policy design, setting up its Social Development Network in 1997 and publishing the first full-Bank *Sector Strategy Paper* (SSP) on social development in 2005 (Vetterlein, 2007, pp. 516–517). In earlier years, only a few projects received loans to advance social development goals for health, education, or nutrition. In the 1980s, social lending narrowed, with loans to countries being conditional on policies favoring economic restructuring over social spending, under influences similar to those that had produced the OECD's "welfare as burden" stance. When this strategy's human costs appeared, additional tools, including "social funds" (sometimes termed "social investment funds") targeted projects concentrated on creating earning opportunities, social and economic infrastructure, and some social services. Later decades brought other strategies. In none was family policy explicitly evoked, although over the decades several initiatives touched significantly on the three components of the domain, particularly stabilizing the demographic situation. Indeed, in the mid-1990s the Bank claimed it was "among the largest sources of financing for population and reproductive health programs" (World Bank, 1995, p. 4).

Major Bank interventions to reduce fertility started early. "A positive population policy seemed to be the overwhelmingly sensible course of action and was widely adopted starting in the mid-1960s" (Robinson & Ross, 2007, p. 424). Population growth had been troubling economists close to the Bank since the 1950s; leading figures, including Gunnar Myrdal and the team of Ansley Coale and Edgar Hoover, informed the IO's research and policy (Robinson & Ross, 2007, p. 3). The institution moved decisively, however, only when Robert McNamera became president in 1967. A first loan to support family planning policies and interventions (1970, in Jamaica) received approval (King, 2007). By the 1980 *World Development Report*, the Bank's established position was that human development depended on reducing fertility. This was not "an end in itself "—a caveat repeated several times in the WDR—but an action that followed from as well as promoted a range of other positive outcomes (World Bank, 1980, p. 97; see also Razavi in this volume). Policy instruments for family planning were multipurpose, meant to promote health, reduce poverty, and support economic growth and social development.

This perspective on population control was not without its detractors. Feminists and other advocates of women's reproductive rights and health pressed for strategies to end women's subordination rather than simply providing, incentivizing, or imposing contraception (some methods of which were dangerous) and even sterilization.<sup>10</sup> The Bank in the 1990s did finally accept a reproductive-rights frame, but nonetheless continued to treat financial incentives as a useful instrument to structure clients' family-planning choices (Chowdhury et al., 2013, Chapter 2, p. 58; for a critique see

<sup>10</sup>Chowdhury et al. (2013, Chapter 2) discuss in some detail the ethical concerns of practices prior to the 1994 Cairo conference, which was a turning point in international thinking (see also Smyth, 1996).

Smyth, 1996). Environmentalists with a "limits to growth" view of the earth's carrying capacity advocated family planning as a policy instrument for managing population size. The Bank adopted this position, in opposition to the US government whose domestic politics were rife with controversy over abortion and contraception and which claimed that promoting economic growth was sufficient (Ringel, 1993; Whitworth, 1994, Chapter 4). Third, some religious forces were simply pronatalist. None of these positions linked demography to classic family policy frames, however, with the quite limited exception of feminist positions that also argued for women's employment rights. They used a gender equality frame rather than a parental employment one, however. The Bank continued to treat family planning primarily within a health frame, thereby generally leaving it detached from other elements of family policy such as supporting parental employment or ensuring adequate family income.

Nonetheless, elements of other family policy objectives did appear, for example, in the major publication intended to correct popular views of the Bank as enthralled by the Washington Consensus—*Investing in People. The World Bank in Action* (World Bank, 1995). Laying out the Bank's approach to human development and overseen by the Directors of the Education and Social Policy Department and the Population, Health and Nutrition Department the document began this way: "Investing in people means helping people invest in themselves and their children. It means empowering households, especially poor households, to increase the quantity and quality of investments in children. For people to break the cycle of poverty and improve their lives, they must have access to adequate social services…" (World Bank, 1995, p. 3). The emphasis on reducing fertility with family planning instruments remained central. But attention also went to strategies to "increase the demand for smaller families" that policy communities believed would follow from health improvements and lower infant mortality, better access to education (particularly for girls), and reduced poverty and higher incomes, all described as broad social policy goals relating to family welfare (World Bank, 1995, p. 10). Nevertheless, analytic interest in the welfare of families or even household units was much less evident than were concerns about health, about partnerships for community development, and even about gender equity.

As the Bank adopted the social investment perspective in the 1990s and 2000s, it moved along two main policy tracks, with each exhibiting no more than partial sensitivity to classic family policy matters. On a first track, the policy work complemented the usual foci on family planning and schooling with a push for early child development (ECD) programs. The outcome sought was improved "human capital," with educational interventions back into preschool being the preferred policy instruments to combat the effects of childhood poverty and health problems, including malnutrition. The Bank's policy entrepreneurs drew on research that reported efficiencies from investments in improving school readiness (via preschool and health) and by supporting disadvantaged groups as well as allowing siblings (older sisters for the most part) to remain in school (Jenson, 2017a, p. 210). This was a childfocused analysis. Like the OECD's *Starting Strong* series, the Bank's ECD work drew on expertise in child development and children's health, rather than on family policy's concern with ensuring parental employment and adequate income. It focused more on girls than adult women and more on the children's social mobility that ECEC might foster than on ensuring adequate family income. It was a pro-poor argument but one that was, as so much of the social investment perspective, more future-oriented than concerned with families in the present.

The second policy track for the social investment perspective clearly targeted family income. Beginning in the late 1990s, a new policy instrument complemented efforts to achieve the pay-offs of investments in ECD and to battle poverty, including its intergenerational transmission. These were conditional cash transfers (CCTs). They transfer money to mothers who fulfill a certain number of conditions with respect to prenatal health and use of health and education services by their children. Countries instituted CCTs with the encouragement and support of the Bank. Indeed, they were a key policy instrument about which the Human Development Network of the Bank deployed vast amounts of social knowledge, expertise, and large loans. By 2015 all Latin American countries had at least one CCT and between 2008 and 2010 the number of cash transfer programs in Africa went from 21 to 37 (Jenson, 2017a, p. 210; see also Chapter 10 by Filgueira and Rossel in this volume). CCTs emerged as a policy innovation in the 1980s in several Latin American countries but in 1997 Mexico created the CCT that we now associate with the social investment perspective. The World Bank quickly joined the CCT bandwagon, offering technical expertise and funding for the extension of these instruments across Latin America (Jenson & Nagels, 2018). While Bank technocrats continued to praise the Mexican CCT, primarily because it had been designed to provide reliable evaluation data, the version that was standardized and diffused owed more to the Brazilian *Bolsa Família* that imposed lighter conditions and used a less neoliberal discourse (Jenson, 2017a, p. 211).

Targeted to the very poor, CCTs were designed to increase the disposable income available to families to invest in their children. Mistrustful of men and assuming mothers would "invest" more in their children, designers of CCTs paid the transfer to mothers (Jenson & Nagels, 2018, pp. 332–333). This done, however, there was no control over spending and older children and fathers could benefit from the higher family income (for a discussion see Fiszbein & Schady, 2009, p. 183). Objectives included reducing children's participation in work, which clearly interfered with human capital acquisition. Bank economists were also fearful that a transfer imposing no employment conditions would encourage parents to reduce their own labor force participation; no such findings were corroborated (Fiszbein & Schady, 2009, pp. 117–119). Finally, the economists were also afraid families would increase their fertility, in order to maintain access to a CCT, but again numerous evaluation studies detected no such pattern (Fiszbein & Schady, 2009, p. 121).

Recent Bank analysis has not brought all three components of family policy together, even when work is the analytic target. The 2019 WDR report, for example, examined changing patterns of work, described the Bank's human capital indicators project in detail, reiterated the need for ECD interventions, and approvingly noted CCTs do not reduce employment (World Bank, 2019, pp. 56–57, 73–76, 108). The analysis, nonetheless, while describing many workers, students, and young children living in poor families, does not target the family for interventions. The WDR carefully explored the links between various kinds of social protection and workers' needs, called for more state responsibility for provision and explored possible policy mixes, including a universal basic income. But as with recent EU policy pronouncements, recipients were framed as individuals without families or even children. The section on working women simply remarks that ECD that offers childcare *might* affect mothers' decisions about employment, while another short paragraph mentions the effects of unpaid domestic labor and the advantages of parental leaves (World Bank, 2019, pp. 75, 97). Overall, however, gender inequalities are ascribed to legal and other discriminations rather than family circumstances. While the Bank has long addressed demographic concerns and poverty, it has rarely put them together with employment to frame them as family policy.

# **Conclusion**

How might we understand the differences in these three IOs' treatment of the goals of family policy listed at the start of this chapter? The OECD remained committed longer than the other two IOs to family-friendly policies

promoting mothers' employment and income supplementation in response to demographic concerns. Its *All on Board. Making Growth Happen* (OECD, 2014) continued to highlight policies for reconciliation and income support as well as the challenges of demographic change (OECD, 2014). In other words, even the shift to an inclusive growth focus on inequalities did not marginalize the family interests of this IO. Nonetheless, the innovative New Approaches to Economic Challenges (NEAC) project, run out of the Secretary-General's office from 2012 to 2015, may sideline the family frame. It was an interorganizational (with the Ford Foundation) and crossorganizational way of working. Fully 14 directorates or centers and numerous committees were acknowledged project participants (OECD, 2015, p. 2). The result was a panoramic overview with significantly thinner consideration of families. The final synthesis reviewed at the ministerial level folded all family-friendly policies (childcare, leaves, working time) into a single paragraph focused on gender (not families) while identifying ECEC as a support for higher PISA scores (OECD, 2015, pp. 42–43).

This example reveals again the effect of bureaucratic location and the knowledge base that comes with it, because each brings a way of seeing, a policy frame and preferred instruments. Over almost two decades the OECD's DELSA jousted with alternative knowledge located in the Directorate for Education. *Babies and Bosses* "saw" working parents, couples making fertility decisions and poor families, where the *Starting Strong* experts "saw" young (even unborn) children needing early interventions in order to achieve transitions to school and school success. Disadvantaged family circumstances threatened that success and ECEC (alongside other interventions) might counter it. When both the education and the social directorates were thrown in with 12 units framing "new approaches" they had to contend with other economists' often more familiar frames for economic growth, and were forced back to stripped-down individualized versions of their policies.

The OECD was not alone in experiencing such processes of *nonfamilialization*. In the EU, too, frame competition occurred. The emerging social investment perspective turned the lens from families, women, and children toward poor children, bringing into the mix other frames and instruments. In addition, the EU's organizational decision to move responsibility for gender from a social DG to one concerned with human rights resulted in greater reliance on a discrimination frame and away from even a reconciliation one (Jacquot, 2015, Chapter 5). Non-familialization of the Pillar of Social Rights was one result. Nonetheless, the EU is a large bureaucratic organization and the search for bridges to consensus across policy frames is the standard way of working, as the 2019 directive on work–life balance indicated well. Thus, gender and family policy advocates concerned with women's employment, family income and demography may be able to retain a toehold in a political context in which paradigm plurality is the norm.

The World Bank's way of working is different yet again. Its large bureaucracy dominated by economists manages to retain separate spheres, cocooning child development apart from departments dealing with monetary transfers, and those concerned with family planning. The lack of cross-reference to each other's frames is notable. This is now particularly remarkable with respect to the objective of stabilizing demography. In the mid-1990s family planning was fully entwined with health, education, and nutrition, reflecting that moment's limited but still integrated bureaucratic structure for the social sector. With the expansion of the Human Development Network over the next two decades, the cross-network ties significantly loosened. It was possible to provide a detailed analysis of CCTs that never mentioned family planning and addressed high fertility only as a possible (but ultimately not demonstrated) negative consequence of cash transfers to mothers. Concerns about unequal gender relations in the family were confined to the effects on girls' schooling (Fiszbein & Schady, 2009). When the 2019 WDR, for its part, analyzed "working women," only 10 lines went to what the OECD calls "family-friendly" policies.

Thus, analytic frames matter within any large organization. The story of these three IOs is both one of change and one in which the popularity of child-centered social investment has frequently trumped all three of the components of family policy that this chapter has followed, thereby fostering non-familialization.

# **References**


Anderson, K. M. (2015). *Social policy in the European Union*. London: Palgrave.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **4**

# **Do International Organizations Influence Domestic Policy Outcomes in OECD Countries?**

**Linda A. White**

This chapter examines the role that transnational and supranational actors and organizations, and particularly international organizations (IOs) play in promoting domestic policy change in areas related to family policy in OECD countries. Existing research demonstrates a mixed record in terms of transnational and international non-state actors' influence on domestic policy-making generally, particularly in the Global North (Avdeyeva, 2010; Kollman, 2009; Paxton, Hughes, & Green, 2006; Zwingel, 2012). This chapter argues that international norms in the areas of gender equality in general and family policy in particular are diffusing to domestic policy processes rather unevenly across OECD countries (see also Linos, 2011; White, 2011, 2014, 2017b). For countries that are part of the European Union (EU), this supranational organization has been an important driver of social policy related to the family and balancing work and family life. It can set common principles and standards in the form of directives on member countries, and the European Court of Justice (ECJ) can rule on member countries' compliance with EU laws (see also Jenson in this volume). Most recently, the European Council and European Parliament enacted a Work-Life Balance Directive (PE-CONS 20/19) in 2019 based on a 2017 European Commission proposal that set even higher minimum standards in member countries regarding paternity leave, parental leave, carers' leave, and flexible work than the 2010 Council Directive on parental leave (Council

L. A. White (B)

University of Toronto, Toronto, ON, Canada e-mail: linda.white@utoronto.ca

© The Author(s) 2020

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_4

Directive 2010/18/EU). For example, the Directive requires member states to provide paternity leave for at least ten working days after a child is born paid at a level equivalent to maternity leave provisions. It also requires member governments to expand parental leave from at least four months unpaid to paid and expand the number of months that are non-transferable between parents from one to two months and to permit it to be taken in flexible forms, such as on a part-time basis. Member governments are also required to permit flexible work arrangements for parents and caregivers for children up to age eight. Finally, the Directive requires either paid or unpaid leave of up to five working days per year for someone to provide personal care or support to a relative, or to a person who lives in the same household as the worker, and who is in need of significant care or support for a serious medical reason.

Yet, country governments within the EU and certainly outside of the EU still exercise considerable autonomy in taking up ideas of IOs. A significant constraint on IO influence is the willingness of country national governments to respond to international norms and standard setting and to regard IOs as sources of legitimate policy advice. Even if policy ideas become normative at the international level, there can be varying levels of receptivity to those ideas domestically. At the same time, while IOs may not directly affect domestic decision making, they can, through their reports and conferences, facilitate the spread of ideas of best practices among member countries. IOs can thus help to create norms and transmit those norms to domestic policy actors who then advocate for their adoption. Generally, then, IOs are not as much authoritative actors as epistemic actors who contribute to policy ideas, but with policy implementation remaining the responsibility of domestic political actors and coalitions.

The chapter begins by reviewing the extant literature on supranational actors generally and IOs particularly as sources of authority and of domestic policy influence. It then focuses more specifically on the track record with regard to gender equality in general and family policy and programs in particular.

# **International and Transnational Influences on Domestic Policy Making: A Review of the Literature**

The literature on transnational policy diffusion points to myriad sources of domestic policy ideas, including international organizations (IOs) (Barnett & Finnemore, 2004), transnational advocacy networks (Keck & Sikkink, 1998), epistemic communities (Haas, 1992), networks of state actors (Slaughter, 2004), and government-to-government transfer of policy ideas (Marsh & Sharman, 2009). These myriad actors and institutions are involved in international regime setting: that is, in establishing "social institutions consisting of agreed upon principles, norms, rules, procedures and programs that govern the interactions of actors in specific issue areas" (Zwingel, 2012, p. 116).

International organizations such as the European Union (EU), International Labour Organization (ILO), Organisation for Economic Cooperation and Development (OECD), and the United Nations (UN), particularly the UN Educational, Scientific, and Cultural Organization (UNESCO), the Office of the High Commissioner on Human Rights (OHCHR), and the United Nations Children's Fund (UNICEF), along with transnational actors and activists have been key players in transmitting and supporting ideas around democracy, human rights, and labor standards. Those ideas have become embedded in a dense array of laws and conventions and supported by formal laws and institutions (Barnett & Finnemore, 2004; Goodman & Jinks, 2013; Greenhill, 2010; Klotz, 1995; Payne, 2001; Simmons, Dobbin, & Garrett, 2008).

IOs and other actors and institutions have also been key in promoting gender equality in general, including voting rights, gender quotas, and sexual orientation and transgender rights as normative (Avdeyeva, 2010; Berkovitch, 1999; Fernández & Lutter, 2013; Hughes, Krook, & Paxton, 2015; Kollman, 2007; Krook & True, 2012; Ramirez, Soysal, & Shanahan, 1997; Swiss, & Fallon, 2017).1 The question is, have gender equality norms to support women, children, and families and to help balance work and family life become embedded in domestic policies and institutions? And has the status of women, children, and families improved directly because of international norm setting? After, all, as Zwingel notes with regard to gender equality, "As many violations of women's rights are rooted in sociocultural traditions and perpetrators are often nonstate actors, such violations were long seen as a problem to be solved by incremental change from within a given society, not by international interference or even pressure" (2012, p. 116). So, does the work of supranational actors in general, and IOs in particular make a difference? Why would we expect that they would?

Simmons et al. (2008, p. 10) examine the processes by which ideas diffuse and become embedded institutionally. They argue that transnational diffusion of policy ideas occurs as the result of four mechanisms: competition, coercion, socialization/learning, and emulation. Governments may

<sup>1</sup>Norms are defined as collective expectations of "right" or "appropriate" action characteristic of particular actors (Jurkovich, 2019, p. 2).

adopt similar policy ideas as an instrumentally rational means to *compete* with other countries economically. Governments may also experience *coercion* from leading or powerful countries which can encourage or even impose policy harmonization. It is thus important to observe countries' position in the world order. Governments may be *persuaded* to adopt a particular policy because they come to believe it is the "right" thing to do, for example, in order to be seen as "modern" (Meyer, Boli, Thomas, & Ramirez, 1997). Alternatively, states may "actively assess the *content* of a particular message – a norm, practice, or belief – and 'change their minds'" in a process of learning (Goodman & Jinks, 2013, p. 22). As Greenhill (2010, p. 128) argues, policy leaders such as IOs or leading states can use economic and other sanctions to encourage cooperation and punish norm violators. While the former processes can be described as socialization or "learning" through persuasion, where states add new information to prior knowledge and beliefs to revise their assessments, the latter process is more akin to coercion. Finally, states may simply "follow the leader" and emulate, that is, imitate the most powerful or successful actors. Emulation may be deliberate, as a means to increase one's standing in the world without significant costs, or a process of more "blindly" adopting the "beliefs and behavioral patterns of the surrounding culture, without actively assessing the merits of those beliefs and behaviors or the material costs and benefits of conforming to them" (Goodman & Jinks, 2013, p. 22). Emulation can also work in the opposite direction: one country may be tempted to opt out of a practice if others have adopted it. In other words, emulation on the part of some states could lead, in the end, to greater policy divergence (Encarnación, 2017).

These processes of policy diffusion can simply involve government-togovernment transfers of ideas, norms, and best practices—after all, we know that policy decisions take place in conditions of state interdependence and decisions taken by one government alter "the conditions under which other governments make their decisions" (Elkins, 2009, p. 49)—or they can be explicitly facilitated by IOs. The sheer number of meetings of "first ministers" in intergovernmental forums has proliferated over the past fifty years (Greenhill, 2010), enabling policymakers from various countries to interact more than ever and facilitate government-to-government learning and policy convergence (Boushey, 2010; Drezner, 2005).

Beyond state-to-state policy transfer, however, are other principal agents of diffusion, including *transnational advocacy networks*, "distinguishable largely by the centrality of principled ideas or values motivating their formation" (Keck & Sikkink, 1998, p. 1) and *transnational epistemic communities*, communities of experts rooted in shared scientific models who articulate "cause-and-effect relationships of complex problems," and who share their ideas with each other and with policy makers (Haas, 1992, p. 2). *International organizations* are also key organizations that help to facilitate the spread of common practices amongst member countries. The question is what are IOs' sources of authority and what kinds of governance functions do they perform. After all, the traditional role of IOs and IGOs are to facilitate interstate interactions, promote collective interests, and provide services such as data collection on behalf of states, not to be domestic lawmakers. Thus, an important research question is whether and to what extent IOs affect domestic policy development.

Barnett and Finnemore (2004) outline the sources of IO authority, including the strongest source, conditionality, which requires states to alter domestic policy in exchange for funding. Such is the case with regard to the International Monetary Fund and World Bank. Other sources of authority include facilitating the negotiation of treaties and conventions that bind countries that ratify, such as UN human rights conventions. Finally, some IOs are authoritative because of their perceived legitimacy on economic and other policy matters and expertise, such as in the case of the OECD (see also Jenson in this volume).

In most matters to do with social and family policy, IOs rarely have the authority to exercise "harder" forms of power. Greenhill notes that the European Union has the power to motivate countries that wish to join the EU to, for example, comply with human rights practices, but that "this type of coercively induced human rights improvement seems to be more the exception than the rule" (2010, p. 129) with regard to IOs. De la Porte notes that the EU has the power to regulate—that is, set legal standards and norms at the EU level—and to facilitate "hard" coordination—that is, "EU-facilitated policy coordination that is legally binding and the involves high surveillance and enforcement of member state policy" (2018, p. 478). But it does not have the authority to "decide upon national social security, the principles of distribution, and the access to social rights or the level of generosity of various benefits" (De la Porte, 2018, p. 478). De la Porte argues that EU regulations—in the form of directives—can certainly contribute to policy coordination in the area of social policy, mobility rights and the movement of works, anti-discrimination, and even the labor market given EU directives on minimum maternity and parental leave standards. Still others argue that "softer" tools at IOs' disposal can be persuasive (Kollman, 2009). Soft power refers to the ability of IOs to exert influence through the establishment of norms that socialize actors toward compliance, as well as processes of persuasion and learning, that is, the *proffering of policy models* or "best practices." Furthermore, IOs such as the OECD are also sources of expert knowledge (Barnett & Finnemore, 2004) and can engage in softer forms of governance such as *peer review*—the assessment of policies—and *surveillance*—the monitoring of compliance with agreed to conventions and rules.

Another question is whether domestic policy advocates find the international arena useful in pressing their policy agenda. Poloni-Staudinger and Ortbals (2014) find empirical evidence that in fact women's groups still target domestic levels of government more so than the international level even when the issues for which they advocate are more transnational than domestic. They are only more likely to focus attention on the international level when the domestic political opportunity structure seems less hospitable because of electoral cleavages and low alliance possibilities.

International norms, though, can have a "boomerang effect" in that they can serve as resources for domestic and transnational policy actors to pressure governments through supranational forums (Krook & True, 2012, p. 107). Finally, norms are not simply constructed but also debated and challenged. There is thus a discursive element to their creation and evolution that can be "revealed by repeated speech acts" (Krook & True, 2012, p. 105). The process of international norm setting can therefore be a key domestic discursive instrument that through the very act of their discussion and debate change actors' perspectives on an issue.

# **How Easily Are Ideas Translated from the International to the Domestic Policy Arena? A Review of Empirical Research**

Empirical research demonstrates that norms tend to be imperfectly transmitted from the international to the domestic level and enacted into law and policy (Linos, 2013; Orenstein, 2008; Risse-Kappen, 1994; Weyland, 2006). One limitation is whether states regard transnational advocates, epistemic communities, and IOs as sources of legitimate policy advice. Some country governments may regard the authority of international organizations with suspicion. Other governments may contest the legitimacy of norms.

Domestic uptake of international and transnational ideas and norms is thus contingent on a number of interactive factors—what Risse and Sikkink (1999) label a "spiral model" of contingent change. Domestic institutions and electoral coalitions in government may create veto points that make the transfer of policy ideas difficult (Tsebelis, 1995). For example, the political elite at one level or branch of government may embrace a particular idea, but competing elites either within another branch of government or at another level of government in a federal system may act as "veto players"—someone whose agreement is needed in order to proceed (Tsebelis, 1995). So too can recalcitrant bureaucracies (Patashnik, 2008). Partisan shifts in key political decision makers, such as after an election, may stymie the adoption of a policy with which a previous electoral coalition found favor. Opposition interest groups may also respond negatively to policy initiatives that were generated at the supranational level. Policy makers may thus back away from their international commitments for fear of electoral loss. All these factors can impede the transfer and uptake of international ideas and norms in any domestic policy setting.

#### **International and Transnational Influences on Gender Equality**

What is the record of success in the area of gender equality policy generally and family equality in particular? A number of researchers have examined the domestic policy impact of international human rights norms negotiated between states and articulated in treaties and conventions,<sup>2</sup> such as the UN Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) (Raday, 2012; Zwingel, 2012). CEDAW was adopted in 1979 by the UN General Assembly and was ratified by 187 states. It establishes an international regime with both a regulative dimension—in the form of standard setting—and a constitutive dimension—in the sense that it identifies principles of "right" action for states. The goals articulated in the convention are the "elimination of discrimination against women, achievement of gender equality, and state responsibility" (Zwingel, 2012, p. 117). By ratifying, "states agree to eliminate direct and indirect forms of discrimination against women in any field of life" (Zwingel, 2012, p. 117). The expert committee of CEDAW monitors state performance with regard to the protection and fulfillment of women's rights.

The convention covers a wide range of gender-based inequality concerns including addressing formally discriminatory treatment and recognizes intersectionality brought by poverty, membership in minority groups, and so on. Raday argues that the ratification of CEDAW signals "considerable achievements" including "international recognition of women's rights as human rights, creation of a national machinery for the advancement of women,

<sup>2</sup>For research on the UN convention on the rights of the child see, for example, Boyle and Nyseth (2011) and Brehm and Boyle (2018). See White (2014) on Canada's abysmal implementation record.

collection of sex-disaggregated data by states parties, promotion of transnational social movements [TSMs] for women's political participation, and enactment of specific domestic violence legislation" (2012, p. 516). But the norms articulated in the Convention also compete with norms around states' religious and cultural autonomy and the structural realities of economic globalization where women remain disproportionately in occupations that are not as well remunerated, where they are more likely to live in poverty, and be subject to human trafficking. Furthermore, both Raday (2012, p. 516) and Zwingel (2012) note that CEDAW is one of the most "heavily reserved of the international human rights treaties" particularly regarding religious reservations. A process exists by which individuals can bring complaints and independent inquiries if a state is accused of violating their responsibilities under the Convention. But the remedy is soft pressure or shaming, not coercion.

Zwingel (2012) thus offers a more contextualized assessment of the influence of CEDAW. She argues that "transnational, national, and local dynamics need to be taken seriously to understand the relevance of international institutions" (Zwingel, 2012, p. 115). International human rights regimes are what Zwingel (2012, p. 115) calls "promotional"—i.e., fixing the "ought"; but rights implementation is largely a domestic affair with a number of domestic factors that affect implementation, including competing norms. Thus, ratification of an international human rights treaty is only "one step on a long path to the realization of these rights" (Zwingel, 2012, pp. 115–116).

Other examples of uneven implementation of international norms include women's right to equality of political office holding, and gender mainstreaming (Krook & True, 2012; Paxton et al., 2006; Swiss & Fallon, 2017). The UN, for example, has provided an important forum to push for women's equal rights to political office holding. States adopted the Convention on the Political Rights of Women in the UN General Assembly in 1946, although it was not unanimous and many states reserved the right not to abide by provisions in the Convention (Krook & True, 2012, p. 113). The UN Beijing Platform for Action in 1995 set more specific targets for women's representation at 30% in political office. Those targets gradually expanded to recognize the importance of diversity in office holding, including in the political executive (Jacob, Scherpereel, & Adams, 2014) and the composition of office holders in UN bodies themselves (Krook & True, 2012). Gender mainstreaming norms do not set specific targets but rather prescribe a normative standard and practice of assessing the gendered impact of planned laws, public policies, and programs. As Krook and True (2012) note, "regulative" norms such as gender mainstreaming, while prescriptive, are also fluid enough to allow for variation in adoption.

Research has identified some of the important factors that affect norm uptake domestically: whether domestic governments have an interest and principled commitment to their implementation, which in turn is affected by the partisan composition of governing coalitions and domestic cultural values (Fernández & Lutter, 2013); whether there are robust non-state actors and activity at both the domestic and transnational level to advance equality concerns (Hughes et al., 2015); and whether a significant number of women hold elected office (Avdeyeva, 2010). Boyle, Kim, and Longhofer (2015) find that domestic factors, such as the number of women in parliament, are at least as predictive to abortion liberalization as international factors such as the extent to which countries are deeply embedded in what Meyer et al. (1997) label "world society."3 Boyle et al. (2015) operationalize the latter as the extent of state participation in women's rights INGOs and treaty signing. Kollman (2009) observes, in contrast, that a transnational network of LGBT actors and organizations, along with various European institutions led to a dramatic change in family policy around relationship recognition in a number of European countries over the past couple of decades. Kollman, however, attributes the changes not to legally binding EU mandates but rather to the creation of a soft law norm "cobbled together from a serious of resolutions by the EP [European Parliament], the incorporation of sexual orientation into the Treaty of Amsterdam's anti-discrimination clause, provided by Nordic states, and key decisions by the ECHR [European Court of Human Rights] to eliminate other forms of sexual orientation-based discrimination" (2009, p. 51). Yet, the unevenness in adoption throughout Europe "suggests that pressure for same-sex relationship recognition is filtered through domestic mediating factors" such as governing coalitions (Kollman, 2009, p. 51). Kollman's research also highlights the difficulties faced in attempting to impose harder laws when domestic government coalitions have not internalized core principles "particularly in policy fields, such as family policy, that are highly politicized" (2009, p. 51). Hughes et al. (2015, p. 358) also find a "recoiling effect" by some states in the face of normative pressures and thus a lesser likelihood of adoption, especially in response to highly contested global norms such as gender equality.

<sup>3&</sup>quot;World society" scholars such as Meyer et al. (1997) posit that global norms are important drivers of state action because state actors that participate in those institutions are shaped and constrained by the global norms they encounter about what it means to be "modern".

## **International and Transnational Influences on Domestic Family Policy**

The verdict is similarly mixed regarding the international and transnational influences on domestic policies such as family policy. White (2017b), for example, finds that in the area of maternity and parental leave expansion over the past two decades in Australia, New Zealand, and the United Kingdom IOs played less of a role than expected in benefits adoption/expansion, despite an EU directive and policy attention by other IOs including the ILO and OECD. Indeed, the Australian government had ratified CEDAW in 1983 but reserved the provision regarding maternity leave provision (White, 2017b, p. 62). The UK did not opt into the EU Parental Leave Directive until after the election of the Labor government under Tony Blair. Instead, domestic policy pressures to increase maternal employment, as well as political factors such as the election of center-left governments and interest group activism were found to be more influential. IOs' influence appears to include setting international benchmarks through reports and other policy documents that could be used by domestic actors to push for policy change—what Blyth (2001) calls "blueprints" for action. But, unlike world society scholars such as Meyer et al. (1997) assertion, "IOs' work and the ideas promoted did not create cognitive locks in these countries, where the ideas would be emulated simply because they were perceived as normatively 'good'" (White, 2017b, p. 58).

Other research has demonstrated that domestic political institutions, such as decentralized decision making arising from federalism (Banting & Myles, 2013) or principles of subsidiarity (Gray, 2010), political coalition formation—particularly alliances with left parties and labor unions (Korpi, 2006)—and welfare regimes (Esping-Andersen, 1990) matter; as do institutional path dependencies created by electoral institutions and political economic structures (Iversen & Soskice, 2009). As the literature on gender equality in general notes, so too does governing party ideology, cultural norms regarding the role of the state vis-à-vis the family, and the strength of both conservative public opinion and mobilized opposition to public investments (White, 2017a).

This is not to say that transnational norm development has no role to play. As Jenson's chapter in this volume notes, particularly in the EU context, the European Commission and related organizations played a large role in developing and shaping family policy and programs in response to a variety of member state concerns: population decline as well as an ageing population, income security, maternal employment, and reconciliation of work and family life. The EU has issued a number of directives around pregnancy and parental leave and workplace anti-discrimination and women's employment and child care service targets. And De la Porte (2018) notes an explicit dual-earner labor market norm has permeated EU policies over the last two decades (in contrast to the male breadwinner norm of the past). These ideas and policy goals were not exclusive to EU institutions, however. The counterfactual question is whether, in the absence of the EU whether countries would be driven to respond to labor market, population, and income security concerns regardless.

Those domestic policy agendas—and the factors that influence them are also important to observe. Morgan (2005), for example, tracks the relationship between varieties of capitalism—that is, the organization of principal economic and labor market institutions and the ongoing relationships between the state, employers, and employees—and public investments in human capital supportive policies and programs such as child care. Coordinated market economies tend to provide more extensively subsidized child care for families, in contrast to liberal market economies/liberal welfare states. In the latter countries, "labour markets are deregulated, workers lack job protection and generous unemployment benefits" and there is high turnover in employment (Morgan, 2005, p. 245). Such employment conditions create a pool of low-wage workers for service occupations such as child care. In many European coordinated market economies, a more highly regulated labor market increases the costs of labor. Governments therefore generally take the path exemplified in social democratic countries such Sweden and publicly provide child care with labor market rules that generate a highly skilled and well-paid often unionized child care workforce. Or governments such as France take a more regulatory route and provide policies to help reconcile work and family life to ensure employment is compatible with care.

Traditional patterns of state investment in family policy are changing, however, in a number of OECD countries. Daiger von Gleichen and Seeleib-Kaiser (2018), for example, observe that in a number of OECD countries, public policies have generally shifted away from the male breadwinner model toward support for dual earner households and to help reconcile tensions between work and family. While the early adopters were largely driven by the normative aim of gender equality and assisted in realizing those goals with the support of a strong women's movement, and strong representation in parliament, other countries have adopted family-friendly policies without strong social democratic parties or a strong women's movement. Daiger von Gleichen and Seeleib-Kaiser (2018) argue that more instrumental reasoning, such as improving the use of human capital through increased female employment or addressing demographic challenges are key drivers (see also Ferragina & Seeleib-Kaiser, 2015; Fleckenstein & Seeleib-Kaiser, 2011).

An important research agenda is emerging on the expansion of social investment policies and programs across a number of countries around the world (Garritzmann, Häusermann, Palier, & Zollinger, 2017; Hemerijck, 2017). Less world society and more the functional imperative of labor markets, countries such as Germany are overcoming traditional reticence to maternal employment to enact sweeping family policy reforms (Blome, 2017; Morgan, 2013). Garritzmann et al. argue, "The social investment perspective emphasizes the necessity to invest in human capital, to mobilize and to renew it along the entire life course in order to accompany the demands of changing production systems, and to address the new social risks not met by the old welfare state" (2017, p. 2). As Morel et al. argue, they include such policies and programs that "invest in human capital development (early childhood education and care, education and life-long training) and that help to make efficient use of human capital (through policies supporting women's and lone parents' employment, through active labor market policies, but also through specific forms of labor market regulation and social protection institutions that promote flexible security), while fostering greater social inclusion (notably by facilitating access to the labor market for groups that have traditionally been excluded)" (Morel, Palier, & Palme, 2012, p. 2)

Some scholars argue that the adoption and/or expansion of this basket of policies and programs represents a new paradigm, complementing and in some ways competing with traditional compensatory policies (Hemerijck, 2018; Ha¨usermann, 2018). At the same time, a great deal of variation can be observed in the scope and depth of social investment between and within countries. For example, among the liberal welfare states, governments have expanded public funding for child care in the UK, Australia, and New Zealand much more than in the USA and Canada (White, 2017a). Despite the typical drag that federal institutions and conservative politics play in stymieing social policy investments, the German government under Angela Merkel has made significant child care investments (Blome, 2017; Morgan, 2013). These social investments are driven by electoral calculations as well as the growing significance of women in political parties, according to Morgan (2013). It is worth noting that international norms and standard setting are not considered as salient explanatory factors. Instead, the welfare state "is a power resources that politicians seek to deploy for electoral purposes" (Morgan, 2013, p. 74).

While women's elected representatives, interest groups, and center-left party coalitions have been key factors in the social investment turn and support for work-family reconciliation policies, the decline of labor, and labor's attachment to left parties, have led both center-right and left party coalitions in a number of countries in Europe and elsewhere to flirt with policies and programs that aim to gain the support of "natives" over newcomers (Gingrich & Häusermann, 2015; Schumacher & Van Kersbergen, 2016). At the same time, challenges have arisen to welfare models based on universal entitlement, as a changing pool of recipients draws on benefits; and on to the dual-earner model which has only recently shifted from the male breadwinner model of employment (Grødem, 2017). A future area of research is to examine the diffusion of ideas around has been labeled "welfare nationalism" or "welfare chauvinism," in conjunction or in competition with traditional coalition tactics and social investment policies, as well as the role that IOs such as the EU can play in tempering radical chauvinism.

## **Conclusion**

As Krook and True (2012, p. 106) note, the trajectory of norms is "fraught with contestation and reversals as state and non-state actors compete to identify, define and implement" norms. Thus, it is very helpful to think of norms as processes rather than as things. This chapter documents some of those norms around gender equality, rights around family formation, and family policy, especially around policies such as child care and parental leave that help reconcile work and family life. While these norms around the "oughts" of family support are increasingly advanced by IOs and EU member states, the attendant laws and policies are subject to continuing contestation and change and myriad sources of ideational influence, including ideas generated at the international level. But domestic coalitions of political actors and organizations are key to implementation, as other chapters in this volume document.

## **References**


liberal welfare states. *Social Politics: International Studies in Gender, State & Society, 24* (1), 55–80.

Zwingel, S. (2012). How do norms travel? Theorizing international women's rights in transnational perspective. *International Studies Quarterly, 56* (1), 115–129.

**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **5**

# **What Does the UN Have to Say About Family Policy? Reflections on the ILO, UNICEF, and UN Women**

**Shahra Razavi**

The United Nations (UN) system is not a monolith by any stretch of the imagination. Insiders often make a distinction between the Secretariat, on the one hand, and the specialized agencies, funds, and programs, on the other, largely based on their sources of funding and modes of governance. A more useful delineation for the purposes of this chapter is what the UN Intellectual History Project refers to as the "3 UNs"—the UN of governments and intergovernmental processes, the UN of staff members, and the UN of closely associated consultants, NGOs, and experts (Jolly, Emmerij, & Weiss, 2009). It is at their intersection that policy ideas—"arguably the most important legacy of the United Nations"—are spawned (Jolly et al., 2009, p. 39). Not only have these ideas shaped global debates on peace and security, human rights, and the international economic framework, they have also offered human rights-based alternatives to the neoliberal worldviews and policy prescriptions of the international financial institutions (IFIs), even if the reaction of the UN to the IFIs can be characterized as "too little and too late" (Jolly et al., 2009, p. 13).

The scope of this chapter is both selective and limited. It considers three entities with mandates that have particular relevance for family policy: The International Labour Organization (ILO), the United Nations Children's Fund (UNICEF), and the United Nations Entity for Gender Equality and Women's Empowerment (UN Women). The chapter asks whether the three

© The Author(s) 2020

S. Razavi (B)

Research and Data Section, UN Women, New York, NY, USA e-mail: razavi@ilo.org

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_5

entities have a policy agenda in relation to families, particularly its gender dynamics, and if so, what has its evolution been, and with what inputs from governments, agency staff, and epistemic communities and advocacy networks.

To answer these questions, we need to unpack the meaning of the term policy in the context of international organizations. Policies reside both in the norms and standards they issue, as well as the content of their flagship reports, the pronouncements of their senior leadership, as well as their programmatic work. A comprehensive picture of these agencies would also have to be garnered through a dual focus, both at the level of headquarters (HQ) and in their regional and country offices ("field") where interactions with governments and other actors shaping policy are most direct. However, due to limitations of space and time, this chapter does not delve into the workings of regional and country offices which would have required significant primary research, keeping its focus on norms and standards, while also referring to flagship publications that articulate agency positions on major policy issues of concern. The extent to which these global norms and flagship reports permeate policy work with governments on the ground is a question beyond the scope of this chapter (but see White in this volume). What this means concretely for each agency is explained in sect. "What Is Family Policy?" where I also briefly unpack the parameters of family policy. Sections "ILO: A Labourist and Maternalist Approach to Families?", "UNICEF: Children Rights at the Center, Women's Rights an After-Thought?", and "UN Women: Feminist Vision of Families, with an Achilles Heel?" then provide a case by case analysis of the three agencies, before sect. "Conclusion" concludes.

What I hope the chapter will show is that each agency looks at family policy through its own specific lens, shaped by its mandate and institutional context, which creates considerable continuity in their respective approaches. What this also means is that there is no "one UN" approach to family policy to date and significant room for crafting one that has gender equality at its center. However, continuity does not imply inertia. Agency positions do change, even if at the margins, through their norm-setting bodies as well as more indirectly in response to the "ideas and (non) decisions" (van Daele, 2010, p. 38, cited in Deacon, 2013) of their staff, and their interactions with the epistemic communities (Haas, 1992) and transnational advocacy networks (Keck & Sikkink, 1999) that bring new ideas and make claims.

The ILO, established in 1919 and the oldest of the three agencies, has had the longest-standing engagement with family policy with its focus on labor standards and decent work. This occurred largely through a laborist paradigm concerned with work/family conciliation and social protection, both central to most definitions of family policy. However, laborism itself has been underdoing important changes with implications for how the ILO sees family policy. Driven by its child-centric mandate, UNICEF has consistently shone a spotlight on children's well-being and rights, but in so doing has arguably left out the needs of working parents, especially mothers who provide the bulk of unpaid care for children. Here too, however, there are signs of incipient change. The youngest of the three, UN Women, has expanded the terrain of family policy beyond core social policies by centering key feminist concerns, such as domestic violence and reproductive rights, while also broaching the highly politicized topic of "diversity" of family forms. Its Achilles heel is that both family policy and social protection have yet to find a strong footing in the organization's programmatic work and strategic plan.

The growing global interest in the care economy, or "care crisis" according to some readings, alongside transformations in gender roles, may account for the recent turn to family policy by both UNICEF and UN Women. While ILO's interest in family policy is long-standing, what seems to be new is its expanded attention to the rights of all citizens or residents, going beyond its core constituency.

## **What Is Family Policy?**

Family policy, as several contributors to this volume suggest, has moveable boundaries. This is especially the case when considering international organizations that work in countries with highly diverse socioeconomic structures and social policy configurations. Two conceptual parameters, proposed by Mary Daly in this volume, are useful in defining the boundaries of family policy. One core consideration of family policy is the *resourcing* of the unit and the individuals that comprise it. This first dimension directs our attention to social policies that have been at the heart of comparative family policy, i.e., public interventions including leaves, social protection transfers such as child and family benefits, and public services such as health, education, and childcare. A second consideration, informed by feminist thought, is the *regulation* of individual behavior and intra-family social relations along gender and generational lines which shape the power dynamics and inequalities within the unit. This delves into the broader legal and institutional context that governs marriage and cohabitation, sexual relations and reproduction, and interpersonal dynamics and intimacies, including issues of violence and bodily integrity that are core feminist preoccupations.

UN agencies per se do not issue global policies since the UN is not a global government. More accurately, some UN agencies, ILO and UN Women among them, have intergovernmental bodies—the International Labour Conference (ILC) and the Commission on the Status of Women (CSW), respectively—that adopt normative standards. There are, however, important differences between the norm-setting mechanisms and procedures of the ILO and UN Women. The conventions adopted by the ILC are legally binding international treaties that may be ratified by member states, while the "agreed conclusions" reached by CSW constitute international policy recommendations.<sup>1</sup> In both cases, agency staff—technical experts and bureaucrats—function as secretariats to the norm-setting bodies: they prepare documents and reports, and refine concepts which are then taken to the ILC and CSW for deliberation by government representatives.<sup>2</sup> In the case of the ILO, its tripartite governance means that government representatives work alongside the representatives of organized labor and employers who sit on its governing body and attend the ILC.

UNICEF does not have a standard-setting body that is equivalent to the ILC or CSW, both of which convene annually. But its work is grounded in the Convention on the Rights of the Child (CRC), which specifically grants a role to UNICEF for the implementation of the Convention.3 Responsibility for monitoring the enforcement of the CRC by governments that have ratified the Convention or one of its Optional Protocols, however, is undertaken by the Committee on the Rights of the Child, made up of 18 experts in the field of child rights who are nominated and elected by States Parties but act in their individual capacity. Under the CRC, UNICEF can be present when the Committee reviews the implementation of the Convention in a particular country and can be invited to provide expert advice and submit reports.

The organizational footprints of the three agencies are also different. UNICEF has a budget that is almost 10 times that of the ILO and 16

<sup>1</sup>CSW is the principal global intergovernmental body exclusively dedicated to the promotion of gender equality and the empowerment of women. It is a functional commission of the Economic and Social Council (ECOSEC). https://www.unwomen.org/en/csw. Accessed 23 Aug 2019.

<sup>2</sup>While final decisions are made by Member States, the leverage that secretariats have in pushing certain agendas through requires more research, along the lines of Bob Deacon's (2013) fascinating book-length analysis of the adoption of Recommendation 202 on social protection floors by the International Labour Conference.

<sup>3</sup>This is exceptional; no other UN human rights convention gives an explicit role to a specific UN agency. Although UN Women and Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) have a common mandate and collaborate at multiple levels, there is only one mention of CEDAW (reaffirming its relevance) in the General Assembly resolution that founded UN Women (A/RES/64/289), and regrettably no mention of women's rights in the title of the entity.

times that of UN Women.4 Given its budget, UNICEF's main advantage, compared to both the ILO and UN Women, is its presence in nearly every country in the world and the support it can give through technical and financial assistance for the implementation of the CRC, as well as the preparation of national reports to the CRC Committee. UNICEF, along with the World Health Organization (WHO), also issues guidelines on various issues related to child well-being and development, including breastfeeding, which has direct relevance for women's rights and family policy as we shall see.

# **ILO: A Labourist and Maternalist Approach to Families?**

ILO's mandate is to strive for a better world of work for everyone. Since its inception in 1919, it has built a system of international labor standards aimed at promoting everyone's rights at work, to ensure that work is performed in conditions of freedom, equality, security, and dignity. The organization's steadfast message during the rocky decades of neoliberal ascendancy and consolidation, globalization, recurrent economic crises, weakened welfare states and the attendant "race to the bottom" in labor rights has been that international labor standards are an "essential component of the international framework for ensuring that the growth of the global economy provides benefits for all" (ILO, 2019a, p. 7). The 2019 report of the Global Commission on the Future of Work reinforces the same principle by proposing "a humancentered agenda for the future of work that strengthens the social contract by placing people and the work they do at the center of economic and social policy and business practice" (ILO, 2019b, p. 11).

As part of this concern for the social side of work, from its early days, the ILO together with women's rights organizations that operated "in its orbit," advanced regulations and policies related to women's work, including with respect to maternity and family responsibilities (Boris, Hoehtker, & Zimmermann, 2018, p. 5).5 It was at the first International Labour Conference (ILC) in 1919 that the Maternity Protection Convention (No. 3) was adopted. While the 1919 Convention was limited in scope, only covering women working in any public or private industrial or commercial sector, in

<sup>4</sup>In 2018, UNICEF had a total revenue of 6676 million US Dollars, compared to ILO's total revenue of 708 million US Dollars, and UN Women's 404 million US Dollars (ILO, 2018a; UN Women, 2018a; UNICEF, 2018).

<sup>5</sup>As Boris et al. (2018) observe, the lack of formal status in ILO's governance structure, never stopped international women's rights organizations from weighing in on ILO deliberations.

1952 the revised Maternity Protection Convention (No. 103) extended its reach to include women wage-earner homeworkers and domestic workers. It was also in 1952 that the landmark Social Security Convention (No. 102) was adopted which recognized maternity as one of nine contingencies requiring income protection through social security. The Convention promulgated family benefits and pensions, among others, for employees and their family members, financed through contributory social insurance systems. Further changes were brought about with the adoption of the Maternity Protection Convention (No. 183) in 2000 which broadened the scope of coverage to all employed women, including women employed in atypical forms of *dependent* work.

Two criticisms have been leveled against ILO conventions. The first concerns their differential treatment of women and men, at least historically, which speaks to the regulatory aspect of family policy mentioned above. The granting of special labor "protections" to women, which was prevalent for much of the last century, has been criticized for being discriminatory and for reinforcing the male breadwinner family that was hegemonic at the time. This includes the Universal Declaration of Human Rights which reads "everyone who works has the right to just and favourable remuneration ensuring for *himself* and *his family* an existence worthy of human dignity…" (Article 23(3), emphasis added). The other alleged shortcoming is the grounding of ILO's labor standards in an employer-employee relationship, which effectively excludes from their purview the significant cohort of workers, predominantly women, who are either self-employed or who work as contributing family workers on family farms and enterprises. For example, in 2018, 33% of all female employment in sub-Saharan Africa, compared to only 15% of male employment, was as contributing family workers on family farms and enterprises where they often receive no direct pay or remuneration for their work (ILO, 2019d).

Regarding the first charge, although special protections for women workers were prevalent during the early decades of the ILO, by the beginning of the twenty-first century the only convention that applied to women only was the revised Maternity Protection Convention (No. 183). Adopted in 2000, its aim is two-fold: to ensure that a woman's economic activities do not pose risks to her health and the health of her child, and that childbearing does not compromise the economic security of herself and her family. The Convention also stipulates important minimum standards concerning the occupational health and safety, duration of leave, the level of payment, and the funding modality.6

While woman-specific stipulations are highly discriminatory, the issue of childbirth poses a conundrum. As Sandra Fredman points out, "Substantive equality requires stereotypical expectations in relation to childcare to be dislodged while insisting that pregnancy and childbirth receive specific treatment" (2005, p. 29). ILO conventions have indeed focused on the latter aspect, i.e., the need for special measures to guarantee women's right to leave and compensation as a result of maternity. However, a crucial cornerstone of a transformative approach is to ensure not only that women are not prejudiced and penalized by pregnancy and maternity, but also that men are included in childcare. In other words, while pregnancy and childbirth are uniquely female, caring and parenting are not. The question is whether ILO standards have gone far enough in transforming care and family responsibilities into a shared endeavor—an issue that various human rights bodies, including the CEDAW Committee as well the Committee on Economic, Social and Cultural Rights have also recognized (ibid.).

Interestingly, a convention adopted in 1981—the Workers with Family Responsibilities Convention (No. 156)—does exactly that. It deals with a broad range of care responsibilities without representing women as the only ones in charge of care. The same broader and gender-neutral approach,7 however, was not applied when the ILC sought to revise a "fifty year old instrument on maternity protection" in 2000, for example, by including paternity and parental leave among its provisions as the Nordic delegates to the ILC had insisted (Murray, 2001, p. 39). As a result, the Maternity Convention 2000 continues to deal with "only one set of relationships and one mode of care: the mother/child relationship immediately before and after birth" (Murray, 2001, p. 36), even though the related Recommendation (No.

<sup>6</sup>The duration of leave is stipulated to be no less than 14 weeks, of which 6 are compulsory after childbirth; payment is set at a level that ensures an adequate standard of living for the mother and her child, but no less than two-thirds of prior earnings where under the law cash benefits are based on previous earnings; and the funding source is preferably through compulsory social insurance or public funds, rather than employer liability, in order to prevent discrimination against women in the labor market.

<sup>7</sup>A distinction needs to be made between gender blind and gender neutral. It is well-appreciated that in the context of structural gender inequalities, a gender blind approach can lead to the exacerbation of gender inequalities. Macroeconomic policies, for example, are often designed without specific reference to gender, and hence considered to be gender blind; macroeconomic policies interact with structural features of the economy, such as women's disproportionate share of unpaid care work and gender segregation of employment, to produce distinct outcomes for women and men (Heintz, 2019). In the context of care for children, as in this example, a gender neutral approach is one that does not assume women to be the default care providers, and hence can support gender equality by involving men in the provision of childcare.

191) of 2000 makes reference to parental leave. However, to its credit, the revised Convention established new grounds of protection which include an explicit guarantee of return "to the same position or an equivalent position paid at the same rate"—an important recognition of women's strong attachment to the labor market.

This is not to suggest that the ILO—its research, declarations, and pronouncements by senior managers—is still enmeshed in the "workermother" norm. In the past few years a number of publications, including *Maternity and Paternity at Work* (ILO, 2014) and the flagship *World Social Protection Report* (ILO, 2018b) have been documenting both maternity and paternity in law and practice. Furthermore, the landmark 2018 publication, *Care Work and Care Jobs*, is emphatic about the crucial importance of redistributing unpaid care within families if equality in the labor market it to be achieved: "No substantive progress can be made in achieving gender equality in the labor force before inequalities in unpaid care work are first tackled through the effective recognition, reduction and redistribution of unpaid care work between women and men, as well as between families and the state" (2018c, p. 38). The same message was boldly stated in ILO's Centenary Declaration adopted by the International Labour Conference in 2019, calling for "achieving gender equality at work through a transformative agenda," one that "enables a more balanced sharing of family responsibilities" and "provides scope for achieving better work-life balance" (ILO, 2019c, p. 4).

However, since labor standards continue to be the ILO's most important governance tool, what they say, and don't say, has considerable salience. Whether the ILC will seek to revise Convention 183 along the lines of the 2018 report on the care economy, and whether in the current climate of austerity such a revised convention (or recommendation) will endorse equal parental leave for both parents without diluting the crucial guarantees with regard to maternity that Convention 183 has already secured is an open question. However, as the following sections will show, the ILO may need to take some action on this front as other UN agencies move ahead with gender-neutral family leave guidelines.

As for the second charge, of the ILO being ensconced in an outdated employer-employee model, it is important to recognize that the organization has taken huge strides in adapting to the realities of the world of work in the twenty-first century, including the growing prevalence of informal and nonstandard work.<sup>8</sup> This was in great part due to the work of trade unions and

<sup>8</sup>The ILO has had a long-standing interest and engagement with issues of informality; the term 'informal sector' was first coined by the ILO in 1972, based on work carried out in Kenya (ILO, 1972).

organizations of informal workers, and a governance structure that provides a space for bottom-up contestations and inputs. The Self-Employed Women's Association of India (SEWA), for example, was pivotal in leading the effort for the adoption of the Homework Convention of 1996 (No. 177), which marked an important breakthrough by recognizing that the home is the place of work for large numbers of women. It is unfortunate that the ratification rate of this Convention remains among the lowest (Boris et al., 2018). The Domestic Workers Convention (No. 189), adopted in 2011, represents another significant step in creating international norms of legal protection for work long thought to lie outside the purview of the ILO.

With regard to family policy more specifically, one of the most significant breakthroughs came in 2012 with the adoption of the Social Protection Floors Recommendation (No. 202) which effectively applies "not only to the 20% of the world's workers who were formal employees but also to 100% of the world's residents" (Deacon, 2013, p. 34). This Recommendation, which enshrines *universal* access to basic income security and essential health care throughout the life course, has taken the ILO beyond its laborist worldview and contributory social protection systems.

With respect to family policy, not only does Recommendation 202 promulgate basic income security for *all* children, *all* persons in active age who are unable to earn sufficient income (due to sickness, unemployment, maternity, and disability) and *all* older persons, by extending coverage beyond the categories of workers falling under the scope of previous conventions, it also "completes and universalizes the principles of maternity protection established by previous instruments" (Addati, 2015, p. 74). In other words, the income security and maternity benefits that are promulgated apply to people performing all kinds of work, whether formal or informal, paid or unpaid—even though there is still no provision for paternity or parental leave, which effectively keeps the "mother-worker" norm intact.<sup>9</sup>

Despite this important breakthrough—revolutionary in the context of ILO—to get Recommendation 202 through the ILC, its scope had to be constantly managed, and the definition of social protection kept under tight reign. Important for family policy, and for women's rights in particular, while both transfers and services could have been included under the umbrella of social protection, this was not done. In other words, the new regulatory

<sup>9</sup>This broader and universal understanding of the ILO's mandate is in line with another important development: under the auspices of the ILO, the 19th International Conference of Labour Statisticians in 2013, re-defined the concept of work to include "activities that are carried out for the production of goods or services for one's own final use or for the final use of others," thereby bringing all forms of unpaid work, including unpaid care work, under the broad umbrella of work.

regime created through Recommendation 202, left out care services that are powerful enablers of gender equality in the world of work.

In response to the global financial crisis, by fortuitous circumstances,10 in 2009 the idea of a social protection floor had become part of UN policy through the United Nations System Chief Executives Board (UNCEB), where ILO worked with WHO on item 6 of the action plan, called "social services, empowerment and protection of people."<sup>11</sup> In August 2010, nearly two years before ILC adopted Recommendation 202, a Social Protection Advisory Group was established, chaired by Michelle Bachelet, former president of Chile and at the time, the first Executive Director of UN Women. This advisory group issued its own report in July 2011, entitled *Social Protection Floor for a Fair and Inclusive Globalization* (ILO, 2011), widely referred to as the Bachelet Report. The Report, in line with the UN, reflected a broad understanding of social protection, inclusive of both transfers *and* services, including childcare services that are an important component of family policy.

However, this broad definition did not make its way into the ILC Recommendation in 2012. As Bob Deacon recounts, the worry in the ILO Social Protection Department, which was the acting secretariat to ILC for the drafting of the Recommendation, was that the "prospects of a broad campaign for investment in drains and sewers and much more besides" would disrupt the "narrower and precise focus … involving only income guarantees and access to health" which the Department thought could make it through the ILC (Deacon, 2013, p. 45). Article 4 of the Recommendation (No. 202) thus reads:

"Members should, in accordance with national circumstances, establish as quickly as possible and maintain their social protection floors comprising basic social security guarantees. The guarantees should ensure at a minimum that, over the life cycle, all in need have access to essential health care and to basic income security *which together secure effective access to goods and services* defined as necessary at the national level." (emphasis added) (Article 4).

In other words, it is through the guarantee of income security and access to essential health care—the direct objects of the Recommendation—that access to goods and services can be secured. With respect to childcare more

<sup>10</sup>Not least, the role that Juan Somavia, the Director General of the ILO, played in winning over a range of UN organizations and international civil society to the concept of a global social protection floor (Deacon, 2013).

<sup>11</sup>In 2009, UNCEB also established an inter-agency collaboration mechanism on the social protection floors which would continue to coordinate work on social protection across the UN system, bilateral donors, and the Bretton Woods Institutions. The Social Protection Inter-Agency Cooperation Board (SPIAC-B) is co-chaired by the ILO and the World Bank.

specifically, on similar lines, the Recommendation specifies the guarantee of income security (i.e., child benefits) without making the provision of childcare services a direct concern: "basic income security for children, at least at a nationally defined minimum level, providing access to nutrition, education, care and any other necessary goods and services" (Article 5b).

While childcare services are left outside of the scope of Recommendation 202, in subsequent years the ILO has given attention to care services, both as a precondition or enabler of women's employment and as a source of employment. For example, in ILO's standard-setting work, Recommendation 204 on the transition from the informal to the formal economy encourages (2015) "the provision of and access to affordable quality childcare and other care services in order to promote gender equality in entrepreneurship and employment opportunities and to enable the transition to the formal economy" (Article 21).12 In its policy work, the ILO has also underlined the importance of regulating the conditions of work in the care sector. This was one of the strongest messages emerging from *Care Work and Care Jobs* (ILO, 2018c). The report provides estimates of the current and projected size of the care sector, and advocates for the feasibility of a "high road" scenario for the sector based on good-quality care employment that promotes gender equality and benefits all involved parties: care recipients, care workers, and unpaid carers. Indicative of the success of this report, the care economy is also featured prominently in the 2019 report of the Global Commission on the Future of Work (ILO, 2019b, p. 28), where it is listed, along with the digital economy and the green economy, as a key site of employment generation that needs to be transformed to create decent work. The same report also refers to parental leave and investments in public care services as crucial areas needed to "foster the sharing of unpaid care work in the home to create genuine equality of opportunity in the workplace" (ILO, 2019b, p. 11).

To summarize, given its mandate to strive for a better world of work, the labor standards issued by the ILC have had a long-standing focus on maternity, and more recently, parental leave as well as social protection transfers, such as family and child benefits. The scope of these standards has broadened over time to include a wider range of workers, and more recently under Recommendation 202, to reach all residents. Care services have also been given increasing prominence lately, both as an enabler of gender equality in the home and in the world of work, and a potential source of present and future jobs. Going against the tide of "private sector solutions" that has swept across the UN system, the ILO has continued to advocate for the regulation

<sup>12</sup>Significantly, article 21 appears under section 5 of the recommendation, which is on rights and social protection.

of care jobs to build a "high road" scenario that benefits all care recipients and their unpaid care providers, and to create quality jobs in the care sector. The "high road" strategy is premised on its capacity to provide *universal* provisions that are adequate and equitable across all social groups—women or men, poor or non-poor, urban or rural, citizen or non-citizen (ILO, 2018c, p. 116).

# **UNICEF: Children Rights at the Center, Women's Rights an After-Thought?**

With its significant financial and technical presence in low- and middleincome countries, UNICEF has been influential in shaping child-related policy and programming, especially in developing countries. This section briefly considers two specific areas of child-oriented family policy which have considerable bearing on women's rights and gender equality: breastfeeding and childcare services. Both are issues for which UNICEF has been actively advocating, in the former case, along with WHO and an active civil society network.

It is important to mention that apart from its field-based technical and programmatic work, UNICEF has also played a critical and broader policy role at specific junctures by issuing timely "wake-up calls." Its 1989 publication, *Structural Adjustment with a Human Face*, was a milestone documenting the devastating impacts on children of structural adjustment policies imposed by the international financial institutions (IFIs) on indebted developing countries (Cornia, Jolly, & Stewart, 1987). Twenty years later, in response to the wave of austerity measures being imposed by the same institutions in the aftermath of the global financial crisis of 2008, a similar call was issued (Ortiz, Chai, & Cummins, 2011), advocating alternative policies to bring about "recovery for all."

While there are synergies between children's rights and women's rights extensively documented in the 2007 edition of UNICEF's flagship publication, *State of the World's Children* (UNICEF, 2007)—tensions and trade-offs also exist that need to be surfaced to inform policy choices, rather than assuming that the synergies are automatic and "what is good for children is also good for women" as the default. Attention to potential tensions is particularly important given the long-standing tendencies and powerful cultural assumptions that have lumped women's and children's interests together. Women's incorporation into welfare systems, for example, has been "strongly influenced by their symbolic and social roles as mothers" (Molyneux, 2007, p. 2), evident today in relation to child-oriented cash transfer programs that have proliferated across diverse regional contexts, often targeting low-income women in their capacity as mothers. The Convention on the Rights of the Child (CRC) itself, which provides the normative foundation for UNICEF's work, has been criticized for stereotyping women as mothers, thereby limiting their life options. With this in mind, this section considers if UNICEF has been able to alter its maternalist lens and see women as actors with their own rights, rather than as a means or "policy conduit" to secure child welfare (Molyneux, 2007).

Exclusive breastfeeding for at least six months, was the key message of the 1990 Innocenti Declaration on the Protection, Promotion, and Support of Breastfeeding, which came out of a WHO/UNICEF policymakers' meeting on "Breastfeeding in the 1990s," a global initiative, held at the Innocenti Center in Florence in 1990. The message has been forcefully sustained not only through the Global Strategy for Young Child Feeding, jointly developed by WHO and UNICEF (2003), but also by an active global network of individuals and organizations dedicated to the protection, promotion, and support of breastfeeding worldwide, called World Alliance for Breastfeeding Action (WABA).

The Global Strategy, as it declares in its preface which is signed by the directors of the two organizations, is based on "the evidence of nutrition's significance in the early months and years of life" (WHO & UNICEF, 2003, p. 5). Lack of breastfeeding, it continues, "and especially lack of exclusive breastfeeding during the first half-year of life" constitute "important risk factors for infant and childhood morbidity and mortality that are only compounded by inappropriate complementary feeding" (p. 5). The life-long impacts include "poor school performance, reduced productivity, and impaired intellectual and social development" (p. 5). The "call for action" urges governments, international organizations, and others to provide "mothers and families the support they need to carry out their crucial roles" (p. 6).

The 2003 Strategy declares "mothers and babies" to be "an inseparable biological and social unit" (p. 3), with directives that at times border on compulsion, "The vast majority of mothers can and should breastfeed" (p. 10), as well as blaming mothers for uninformed feeding practices that result in child malnutrition. References to women's employment are largely negative: "Expanding urbanization results in more families that depend on informal or intermittent employment with uncertain incomes and few or no maternity benefits" (p. 6). The document displays little recognition that women in many low-income families need to earn an income, let alone any mention of the empowering potential *for women as women*, of having an income of their own. There are references in the Strategy to the ILO Maternity Protection Convention (No. 183) and the need for "day-care facilities and paid breastfeeding breaks" for all women employed outside the home. The ILO standard of at least 14 weeks of leave, however, is hardly enough to cover six months of exclusive breastfeeding that is called for. During the negotiation of the ILO Convention on maternity (No. 183), WHO and UNICEF, as "observers" at the ILC, had voiced a strong preference for six months of maternity leave, and to this day they continue to advocate for paid maternity leave for a minimum of 18 weeks, which is in line with ILO Maternity Protection Recommendation (No. 191), and preferably, for a period of six months, along with "paid paternity leave" to add some gender balance.<sup>13</sup>

Another critical area of intersection between children's rights and women's rights is with regard to early childhood education and care (ECEC) services. International organizations were not active players in the field of ECEC until the 1970s, and initially they were only interested in pre-school education; children under three were assumed to be cared for at home by their mothers (Mahon, 2016). However, in recent decades ECEC has moved to the center of global policy debate, given its fit with the contemporary discourse on the "knowledge-based economy" and as part of the push-back to the harsh neoliberalism of the 1980s (Mahon, 2016). A handful of International Organizations, most notably the World Bank, the Organization for Economic Cooperation and Development (OECD) as well as UNESCO and UNICEF, have framed ECEC for a global audience, drawing selectively on Northernbased research by neuro-scientists, economists and pedagogues to persuade governments in low-income countries to invest in early childhood, given its "high returns" and "in the interests of competitiveness" (Penn, 2019, p. 7).

While there is broad-based agreement on the importance of ECEC, there are differences among these major players in how they frame the issue—for some it is a social right while for others it constitutes an investment in human capital; there are also salient differences among them in terms of preferred modalities of service provision—formal and universal programs versus nonformal community programs targeted to the poor (Mahon, 2010, 2016; see also Vandenbroeck in this volume). For our purposes another important divide is between those looking at ECEC largely from the perspective of child development and those looking at it from the perspective of adult women's rights, both as unpaid care providers in families *and* as childcare workers in the delivery of ECEC services.

<sup>13</sup>UNICEF Executive Director Henriette H. Fore and WHO Director General Tedros Adhanom Ghebreyesus, World Breastfeeding Week, 2019 Message.

Not surprisingly, policy interventions in this area are often framed in terms of "the best interest of the child" to promote and optimize children's health and cognitive development. Gender equality and the rights of adult women—whether as unpaid family caregivers or childcare workers staffing ECEC programs—are all too often an after-thought (Staab, 2019). While the availability, affordability, and quality of childcare services, including their location and opening hours, are pivotal for women's ability to access paid work, ECEC services are not often designed with women's needs and aspirations in mind, though there are enough examples to show that it can be done. Apart from Nordic countries where children's rights and development have been center stage along with strong public support to promote gender equality (Mahon, 2016), there are also a handful of developing countries where efforts are being made to gradually transform ECEC provision in ways that respond to women's rights. In both Chile and Ecuador, for example, efforts have been made to up-grade service quality and adjust the schedules of childcare centers to better respond to the needs of working parents, and to improve the employment conditions and wages of their predominantly female staff (Staab, 2019).

UNICEF has been an avid advocate of early childhood development programs, largely from the perspective of child development, combining human rights and social investment discourses. While in its 2007 flagship report, childcare responsibilities were recognized as a constraint on women's labor force participation, it is not clear if this translated into "a focus on high-quality, center-based childcare services in the organization's programming efforts on the ground" (Staab, 2019, p. 75). Earlier research suggests that at the country level, UNICEF's interventions have tended to promote home- and community-based programs aimed at strengthening parenting skills for children under three, thereby spreading its efforts "wider but more thinly" (Penn, 2004, p. 25), while for three to six-year-old children it has supported ECEC services to enhance school readiness (Staab, 2019).

In July 2019 UNICEF launched a "family friendly" policy initiative which seems to signal something of a breakthrough, as it finally links its concern with child development to the needs of working parents. A series of evidence briefs—on paid parental leave (UNICEF, 2019a), childcare services (UNICEF, 2019b), child benefits (UNICEF, 2019c), and especially women's economic empowerment (UNICEF, 2019d)—strongly connect to the needs of working families, especially working women in low-income households, in the context of a "global crisis of care" and cognizant that care responsibilities often "compromise women's economic empowerment" (UNICEF, 2019b, p. 1). There is recognition of the gendered effects of time-related policies, and that "long maternity leave, with no commensurate paternity/parental leave, may reinforce the gendered division of care work within the home" (UNICEF, 2019d, p. 3). The brief on child benefits also raises the concern that conditionalities attached to child benefits may reinforce gender stereotypes while adding to women's unpaid work, and hence articulates a preference for making child benefits universal and unconditional (UNICEF, 2019c, p. 3). Referring to the integrated nature of the Sustainable Development Goals (SDGs), there is a call for action to connect the targets on early childhood programming (4.2), with valuing and supporting unpaid care work (5.4) and promoting decent work for all (8.5) (ibid.).14

It is too early to gauge whether the "family friendly" approach has filtered down to UNICEF's programming on the ground. The briefs foresee a major role for publicly funded childcare services because "private childcare remains expensive and restricts women in low-income families from engaging in the paid economy," while employer-led and employer-funded care services are considered less desirable because they are likely to put a strain on small- and medium-sized enterprises where the majority of women workers are located, thereby restricting coverage (UNICEF, 2019b, p. 5).

However, in the current climate of austerity, criticized by UNICEF for being short-sighted and misguided, a major expansion in public provisioning will need strong advocacy and support from UN agencies including UNICEF, not least vis-à-vis the international financial institutions that weight-in heavily on developing country governments, urging them to slash public expenditure. In the meantime, an "employer supported childcare" model is being advocated by the International Finance Corporation (IFC) of the World Bank Group. The IFC has been advising companies on how to improve work-family balance for their employees, as a means of attracting and retaining qualified staff and talent—"making the business case" (IFC, 2017), as the current lingo frames it. This is a far cry from childcare as a public good, available to *all* children as a right, regardless of family income

<sup>14</sup>The 2030 Agenda for Sustainable Development, adopted in September 2015, by 193 UN Member States, tackles a broad range of global challenges, aiming to eradicate poverty, reduce multiple and intersecting inequalities, address climate change, end conflict and sustain peace. It is comprised of 17 Sustainable Development Goals (SDGs). Goal 5, for example, is dedicated to gender equality and includes among its 9 targets a specific target (5.4) on recognizing and valuing unpaid care and domestic work. Goal 4, which is on quality education and life-long learning, includes a specific target (4.2) on quality early childhood development, care and pre-primary education, while Goal 8 which is on economic growth and employment includes a specific target (8.5) on full and productive employment and decent work for all women and men, and equal pay for work of equal value (see UN, 2015).

and parental employment status, and risks creating highly uneven and fragmented provision, reinforcing existing inequalities and leaving most informal workers and their children stranded.

To summarize, UNICEF has indeed seen families through the prism of child welfare and development, while women have invariably figured in maternal roles, responsible for breastfeeding and ensuring their children's nutritional and health needs. Through its extensive field presence, UNICEF has been supporting early childhood development through improved nutrition, breastfeeding, parenting programs, and play-based interventions rather than concerning itself with the needs of working parents through quality childcare provision. The recent turn to "family friendly policies" marks a breakthrough, responding to changed material circumstances—a "global care crisis" and women's increasing breadwinning roles—signifying a belated recognition that adult women too are right-holders. The extent to which UNICEF is able to connect children's right with women's rights in its programming remains to be seen, which is where it can make the biggest difference, but also where bureaucratic inertia and resistance are likely to be greatest. Rights-based universal ECEC services that meet the needs of working parents "allow mothers to work outside the home with tranquility and include women educators who become professionals, receive decent salaries, work in adequate places and produce good care for children" (Rosemberg, 2006, p. 82, cited in Mahon, 2010) must also confront the straightjacket imposed by fiscal austerity and the continued faith in private sector solutions and the "business case."

# **UN Women: Feminist Vision of Families, with an Achilles Heel?**

UN Women was created in July 2010 by the United Nations General Assembly to consolidate and strengthen the global drive for gender equality, and address the challenges posed by the fragmentation of responsibilities for gender equality across four different offices. The four predecessor offices that were merged into UN Women in 2010 included the Division for the Advancement of Women (DAW) which acted as the Secretariat to the Commission on the Status of Women (CSW), already referred to in sect. "What Is Family Policy" above; the United Nations Development Fund for Women (UNIFEM); the Office of the Special Adviser on Gender Issues and Advancement of Women (OSAGI); and the International Research and Training Institute for the Advancement of Women (INSTRAW). For our purposes the most relevant are DAW and UNIFEM, since they were responsible for the bulk of normative and operational work on gender equality.

Despite the role of UN Women's first Executive Director, Michelle Bachelet, in chairing the Social Protection Advisory Group (see Section "ILO: A Labourist and Maternalist Approach to Families?" above), there is no reference to UN Women in its 2011 Report. In the preface to the report, Juan Somavia, the Director General of the ILO at the time, explains how Bachelet's "achievements in successfully extending social protection in Chile where significant investments were made to enhance access to health, pensions, education, housing, water and sanitation, and especially to promote child development and improve gender equality" stand her in good stead as chairperson of the Social Protection Advisory Group.

There is little evidence, however, that the findings of the Report were brought back to UN Women. Why did the Bachelet Report not have any ripple effects within UN Women in a context where women persistently shoulder the lion's share of unpaid care and domestic work, comprise 65% of those above retirement age without a regular pension, and either juggle or miss out on employment opportunities because of a dearth of affordable care services and basic infrastructure to reduce the drudgery of domestic work?

Family policies—especially work/family conciliation through maternity and parental leave, investments in care services and social protection transfers—were not prominent themes in the work of DAW, with one important and significant exception. The 53rd session of the Commission on the Status of Women (CSW-53), which convened in March 2009, broached critical areas of family policy in the context of its priority theme, "The equal sharing of responsibilities, including caregiving in the context of HIV/AIDS."<sup>15</sup> This was a significant moment, as it was the first time that the issue of care was being placed on the agenda of CSW. The devastating consequences of the HIV/AIDS pandemic, especially for women and girls in Sub-Saharan Africa, had catapulted the issue of care onto the global agenda. In preparation for the session, as per usual practice, DAW convened an expert group meeting, inviting a range of external experts, many of them prominent feminist academics working on the topic (e.g., Mary Daly was the author for the background paper of that meeting), as well as relevant "sister agencies," including the ILO and the United Nations Research Institute for Social Development (UNRISD) to present their work. The Secretary General report

<sup>15</sup>The Commission on the Status of Women (CSW) adopts a multi-year program of work containing what are called the 'priority theme' for discussion and action for its annual sessions that take place in March in New York.

on the priority theme was drafted by DAW staff drawing on the expert papers and UN agency inputs.

The agreed conclusions of CSW-53 made copious references to the "equal sharing of responsibilities between women and men" for the care of persons, which strategically went beyond persons with HIV/AIDS to include both children and older persons as well (UN, 2009). The ILO Convention that was "duly noted" was Convention 156 on the equal sharing of responsibilities between women and men. In terms of policies, the agreed conclusions referenced a range of "family friendly policies" including maternity, paternity, and parental leave as well as "campaigns to sensitize public opinion and other actors on equal sharing of employment and family responsibilities between women and men." Social protection measures, including child and family allowances, and affordable, accessible and quality care services for children and other dependents, were also emphasized while underlining the need to ensure that such services meet the needs of both caregivers and care recipients. Investment in infrastructure, such as clean water, constituted another prominent theme, given its relevance in the context of caring for people with HIV/AIDS in many low-income communities. Many of these issues would resurface in later years after the creation of UN Women, as will be shown further below, including in the 2030 Agenda under target 5.4 (unpaid care and domestic work).

Nor was family policy a familiar terrain for UNIFEM. Its main areas of programmatic work included ending violence against women; governance, peace and security; and strengthening women's economic capacities and rights (UN, 2000). In 2000, the first issue of UNIFEM's flagship report, *Progress of the World's Women* (*Progress* for short), was launched focusing on women's economic empowerment in the context of globalization (UNIFEM, 2000). Authored by the prominent feminist economist Diane Elson, the report provided a "more complete view of how economies work" (p. 7), including a focus on unpaid care work and volunteer work done in homes and communities, and invisible informal paid work done in small workshops and on the streets. Another prominent theme in the report was the need for governments to be accountable to women for public expenditure (UNIFEM, 2000). Work on gender-responsive budgeting (GRB) as an instrument with which to scrutinize budgetary allocations from a gender perspective, was already underway in some countries. After the launch of *Progress*, it became one of the signature programmatic areas of UNIFEM. The work on economic rights focused on the informal economy while criticizing the structural adjustment policies that were devastating women's livelihoods.

Social policy and family policy, however, seem to have fallen through the cracks, perhaps on the faulty assumption that they were more relevant for high- and middle-income countries, than the low-income and fragile countries that most concerned UNIFEM and its donors.16 In subsequent years, research by a number of UN agencies, including the ILO and UNRISD, as well as prominent research networks has shown the relevance of social policies to all countries, including low-income ones, as an enabler of development, and not something that countries can only afford to do once they have developed.<sup>17</sup> The lack of attention to social policy and family policy may have also had deeper roots in the "women in development" thinking that associated social policies with a "welfarist" approach that feminist advocates were keen to eschew (Razavi & Miller, 1995). Whatever the reason, UNIFEM's work on women's economic rights remained squarely focused on women's paid work, especially in the informal economy.

Hence, when UN Women was created in 2010, with Bachelet at its helm, there was limited on-going work or expertise in the organization on social protection to seize the opportunity presented by the Bachelet Report. Furthermore, as the first executive director of a brand-new organization, Bachelet had the formidable task of making a new organization viable. The reluctance to bring the work on social protection into UN Women may have also been due to what one observer refers to as a "legacy problem, meaning the organization (and its staff ) sticking to inherited priorities from the four predecessor offices," and being reluctant to take on emerging issues, especially those relating to structural causes of gender inequality. It took another five years before the theme of social protection, including family policy, resurfaced, this time in the 2015 issue of *Progress of the World's Women*, focusing on transforming economies for gender equality by anchoring both macroeconomic *and* social policies in human rights (UN Women, 2015).

Under social policy, the report included paid leave (both maternity and parental); social protection transfers, preferably in the form of unconditional and universal child and family benefits and universal pension systems; and adequate investment in a range of public services, including early childhood education and care services and long-term care. In line with human rights principles, the need to ensure the accessibility, affordability, and quality of

<sup>16</sup>Personal communication with Anne-Marie Goetz (28 August 2019), chief of the Women, Peace and Security Section at the time of UN Women's creation, and previously leading the same area of work in UNIFEM.

<sup>17</sup>The case was made most persuasively by the UNRISD research programme on 'Social Policy in a Development Context' which showed how historically social policies were an important enabler of development both in the Nordic countries as well as in East Asia (Mkandawire, 2001).

services was underlined, along with decent conditions of work for serviceproviders. At a broader level, a key message of the report was the need for economic and social policies to work in tandem, seeing both unpaid care services performed in the home and paid care services as "investments" in human capabilities that contribute to economic productivity and dynamism (see also Heintz, 2019).

Almost in tandem with this issue of *Progress*, family policies, such as paid leave, child, and family allowances and in particular the urgent need for governments to invest in care services, also started appearing in the priority theme reports and agreed conclusions, of CSW, especially CSW-58 which took place in 2014 and reviewed the achievements and challenges of the Millennium Development Goals (MDGs) as a precursor to the Sustainable Development Goals (SDGs), and CSW-61 in 2017 which had women's economic empowerment in the changing world of work as its priority theme (UN, 2014, 2017).18 The reconciliation of paid work and care responsibilities, the sharing of family responsibilities, and the urgency of investing in social protection systems and care services were important themes in CSW-58 which paved the way for the inclusion of care in the 2030 Agenda.

The most in-depth treatment of family policy by CSW took place in March 2019, when the priority theme was specifically on social protection systems, access to public services, and sustainable infrastructure for gender equality (UN, 2019). Not only did the Commission recognize the importance of relevant ILO standards, it specifically referred to Recommendation 202 on social protection floors in its preambular paragraphs. Investment in accessible, quality, and affordable early childhood education and care services was recognized as crucial in enabling women to enter and remain in the labor market (UN, 2019, para. 19). Another important contribution was the emphasis it placed on the interlinkages between social protection systems and public services and the need for the two to be better integrated. This is particularly important at a time when child-oriented cash transfers—one specific instrument of social protection—are given considerable policy attention and donor funding, while adequate investment in care services, arguably a far stronger enabler of women's labor force participation and economic autonomy, elude most countries and donor priorities.<sup>19</sup>

<sup>18</sup>Within UN Women, the Intergovernmental Support Division (IGSD) assumes the main secretariat function to CSW as well as other intergovernmental processes. The substantive part of the function, which includes the preparation of Secretary General (SG) reports on the priority theme of CSW each year, is mostly assumed by the technical experts in the Policy Division. The SG reports for CSW 61 and 63 drew on the 2015 Progress of the World's Women report.

<sup>19</sup>There are well-known feminist concerns about the conditional cash transfer schemes: the conditionalities attached to the transfers tend to reproduce traditional gender roles and aggravate the unpaid

Building on this body of work, the 2019 issue of *Progress* focused squarely on the theme of families in a changing world, advocating for a "familysensitive" policy agenda. There are two important elements in this report: first, drawing on the latest available global data it demonstrates the diversity of family forms; and second, it proposes a comprehensive family policy agenda from a feminist perspective.

The first element is in many ways novel and addresses a major lacuna in global reports and policy discussions. At the same time, it responds to contentious political debates, including among Member States that attend CSW, on the diversity of family forms which sees right-wing populists internationally aligned with conservative religious interests in defense of the so-called "natural family" (Goetz & Irani, 2019). With this report UN Women boldly interjects in this debate by providing rigorous empirical evidence to show that families are indeed diverse everywhere (and thus need to be resourced), while the two-parent household with children, assumed to be the "norm," only makes up 38% of all households globally. Family policy therefore needs to respond to the diverse realities of family life which include single-person, single-parent, extended as well dual-parent households, including same-sex partnerships—requiring a major regulatory adaptation if family policy and the resourcing of families is to respond to the reality of how people are living their day to day lives.

The second element brings together the policy analysis already done by UN Women under the theme of social protection and care systems. However, not only did the family-friendly policy package include leave policies (maternity and equal parental leave), transfers (universal child benefits and pensions) and care services (early childhood education and care, and long-term care) the usual components—it also embraced key feminist issues that are not often included under family policy or social policy (Shaver, 1994). These include policies and measures needed to prevent and respond to domestic violence, and policies to guarantee sexual and reproductive health and rights, including comprehensive sexuality education. The last two elements were important additions from a feminist perspective, given their salience to women's human rights. The 2019 report also includes a costing exercise to show that the proposed policy package is affordable for most countries—and not something that only high- and middle-income countries can do.

While the steady rise of family policy in UN Women has been impressive, it has an Achilles heel. UN Women's work on social protection and family policy is yet to find a secure footing in the organization's programmatic work

care work that women already perform (Molyneux, 2007), while also exposing them to discrimination and coercion by programme staff (Cookson, 2018).

and strategic plan. At the time of writing, much of the work and the expertise, remain in the organization's research and normative sections, at the HQ level. However, for the issue to gain traction, especially at the country level, it needs to be translated into programmatic interventions, which is also where funding goes. The fact that both social protection and unpaid care work are visible components in the 2030 Agenda—under Goals 1 and 5, respectively—means that there are likely to be openings for Country Offices to pursue such work, especially if there is funding to support it and if UN agencies with the relevant mandates are able to work together.

At the ideational level, the care economy provides a more comfortable framing for some elements of family policy—most notably investments in childcare services—to gain traction in UN Women. The fact that investments in care services are seen as "productive"—enabling women's labor force participation—provides an easier fit with the "women's economic empowerment" mindset that remains dominant in the organization. In fact, following the publication of UN Women's first flagship report on SDGs that showcased care policies (UN Women, 2018b) several UN Women Country Offices have been costing early childhood education and care services, including the returns on investing in them, to persuade governments to take steps in this direction. Social protection, however, may still be regarded as smacking of "welfarism," even if there is considerable analytical work persuasively showing that it is "productive" (Cichon & Scholz, 2006; Mkandawire, 2007)—an argument that UN Women reports, including *Progress*, have also made.

To summarize, while family policy may not have been an area of work for UNIFEM and DAW, and hence for UN Women at the time its first Executive Director issued the Bachelet Report, in the last six years it has become a visible area of concern, both at CSW and in UN Women's major reports. The reasons for the increasing prominence of the theme is twofold. First and foremost, both social protection and the care economy, which draw attention to family policy, have seen a meteoric rise over the past decade, thanks to the epistemic communities and advocacy networks that have framed the issues for policy audiences. Second, having insiders is also helpful, to seize political opportunities, to domesticate key elements of the social protection and care agenda as relevant for family policy, and make it visible through reports and publications. As a feminist organization, UN Women has expanded the scope of family policy to include not only the issue of domestic violence and reproductive rights, but also has taken on the highly contentious issue of diversity of family forms. These positive developments notwithstanding, this section also raised a question about its uncertain status as a long-term strategic commitment and driver of its programmatic work. It was suggested that in UN Women the care economy, more so than social protection, provides the best ideational framing for family policy.

# **Conclusion**

Policies that can support families in all their diversity, recognize women's rights and embed gender equality are critical for social and economic development everywhere, even though the policy instruments needed to do so will vary depending on the specificities of labor markets, livelihoods and state capacity. To answer the question posed in the title of this chapter, there is no one family policy at the UN, but different policy approaches pursued by different UN entities, reflecting their respective institutional mandates and histories. As we have seen, there is considerable path-dependency in how different UN agencies approach family policy. But there is also learning and cross-fertilization, evident in the way in which issues of care and social protection have reverberated across the system. The recent turn to family policy, with issues of gender, social protection and care at its center, across the three entities, responds to both material and ideational changes: women's changing roles, a perceived crisis of care, and women's claim-making for equality and rights. UN agencies are also porous: transnational networks and epistemic communities have been able to frame family issues as compelling policy problems, while staff have domesticated them through research products, normative, and policy work.

The fact that these issues have also made their way into the 2030 Agenda provides an anchor for a sustained focus on key elements of family policy. The chapter has also indicated two concrete issues that require attention: the ball is now in ILO's court to issue gender-neutral guidance on family leave, but without losing the guarantees for maternity leave that Convention 183 has secured; care services, particularly for children under three, remain orphaned but are too critical and could benefit from more systematic coordination between all three agencies to ensure that they work for children, their parents and care workers.<sup>20</sup> Attention to long-term care services for frail elderly persons is another urgent family policy issue with significant gender implications, as women make up the bulk of those needing care and providing care (both paid and unpaid).

The 2030 Agenda provides a common frame and point of reference, urging UN agencies to connect the dots and think about the issues in an integrated

<sup>20</sup>UNESCO's mandate covers pre-school education, usually for children aged 3 and above, until school age.

way. The focus on synergies and interconnections is a boon for thinking about family policies. At the same time, there is also an on-going process of UN reform which is urging different parts of the system to work together more effectively, overcoming the territoriality around mandates, in order to better support countries to meet the SDGs. The SDGs are broad goals or desirable destinations, but without the policy roadmaps needed to get there. This is where UN agencies can step in to provide guidance to countries, and "best practices," to inform policy choices. Family policies also need financing, which makes it a difficult proposition at a time when austerity looms large. In this context, private sector provision and finance may seem luring, but history tells us that market-based solutions are unlikely to provide the kind of universal social and family policies that can reign in gender, class, and other intersecting inequalities.

**Acknowledgements** The views expressed in this chapter are in my personal capacity and do not reflect the position of UN Women. I would like to thank Laura Addati, Valeria Esquivel, Sylvia Hordosch, and Silke Staab for their comments on an earlier draft, though they are not responsible for any errors or weaknesses in the chapter.

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Part III**

**National**

# **6**

# **Conceptual Approaches in Comparative Family Policy Research**

**Hannah Zagel and Henning Lohmann**

Comparative family policy research relies on concepts that define the relationships between the welfare state and the family. The literature contains several proposals on how to conceptualize these relationships for understanding country differences. Depending on the research tradition, the conceptual approaches vary in their interpretation of what is the main challenge in these relationships. For example, whether the focus is on economic autonomy, care relationships, class differences, or on any combination of these.

Closely linked to the theoretical perspective on the family–welfare nexus is the methodological practice dominant in the respective research traditions, although the links are not deterministic. Researchers have developed their concepts within the epistemological logics of their respective research domains. Comparative historical research stresses the need for in-depth analysis of spatio-temporal configurations while quantitative comparative welfare state research attempts to achieve a high degree of measurability of general concepts across contexts. In comparative family policy analysis, there is arguably a good level of exchange of conceptual ideas between the different comparative approaches.

H. Zagel (B)

H. Lohmann University of Hamburg, Hamburg, Germany e-mail: henning.lohmann@wiso.uni-hamburg.de

Humboldt-University zu Berlin, Berlin, Germany e-mail: hannah.zagel@hu-berlin.de

This chapter provides a systematic overview of some of the most popular and widely used concepts in comparative family policy research. We situate each conceptual approach in their respective research tradition, define the concepts' main characteristics and demarcate them from similar ones. Our discussion is guided along six criteria, which highlight differences and similarities between concepts. These are criteria with regard to the scope of the concepts as well as to methodological issues. Questions related to the conceptual scope are: (1) Who is considered as the main addressees of family policy interventions? (2) Does a concept focus on gender and/or intergenerational relationships? (3) Is the concept defined from the perspective of the caregiver and/or the care receiver? (4) Does the concept consider the state and/or the market as welfare providers alongside the family? The main methodological distinction is whether a concept points to ideal types in family policy regimes or not (5). Last, concepts can be differentiated into unidimensional and multidimensional ones (6). All conceptual approaches we discuss are rooted in comparative welfare state research, which considers nation-states as the main unit of analysis. To foreclose some of our conclusions, the similarities between the concepts seem often stronger than the differences. However, we show that there are important nuances, which can be critical in drawing comparative conclusions about family policy and its expected outcomes. The nuances have been blurred over time, and it is rare to find explicit arguments for using one rather than another concept in empirical comparative family policy research. With the discussions in this chapter we provide the basis for such reflections.

The remainder of this chapter is organized in three parts. In the next section we discuss research traditions, dominant ideas, and methodologies in the field. What follows is the heart of the chapter, presenting definitions and discussions around two classes of concepts in comparative family policy research: categorical concepts (captured in typologies) and gradual concepts (measurable as indexes or scales). In the concluding section we propose criteria by which to classify and evaluate concepts as indicated above. These criteria can be used to inform decisions about which concepts to use in research.

# **Research Traditions, Ideas, and Methodologies**

Conceptual approaches in comparative family policy research are rooted in distinct research traditions. Arguably, today, the ideas once driving different conceptualizations are less visible in research applying them, and are often deemed less important. Understanding the theoretical origins of conceptual ideas however supports an informed use of concepts in empirical research. Further, research traditions are linked to particular methodological strategies. Understanding these links helps to contextualize conceptual approaches and to position one's own research in the field. It should be noted that research traditions are far from being homogenous, nor are the methodological practices anchored within them. There are however broad differences that can be identified for orientation in the field.

#### **Research Traditions**

There are at least three main research traditions in which concepts used in comparative family policy research emerged: *family well*-*being*, *feminist* and *mainstream comparative welfare state* research. The perspective on family policy differs across research traditions and who is regarded as the main addressee of policy interventions. The first tradition, which we here call the "family well-being research tradition," has a focus on the family as a social institution. It is often motivated by the recognition of "the importance of the family in society" (Kamerman & Kahn, 1978, p. 1). The family, usually defined by the presence of children and excluding kin relationships beyond the nuclear family, is understood to fulfill different social functions.<sup>1</sup> For example, it is the site of social reproduction, socialization, and child development, but also of parental challenges to maintain the family socially and economically. Family policy, then, is seen as state intervention "to and for the family" (Kamerman & Kahn, 1978, p. 3), allowing the family to realize their respective functions. Research in this tradition acknowledges the variation in the forms of intervention as well as in the goals of family policy. Kaufmann (2002) differentiates eight different motives used to legitimate family policy intervention: the institutional motive, the natalist motive, the eugenic motive, the economic motive, the societal motive, the sociopolitical motive, the women's issues motive and the children's welfare motive. The institutional motive legitimizes family policies by the need to preserve the family as a social institution with its own value. The natalist motive on the other hand centers on the question of demographic reproduction, while the eugenic motive pursues control over the genetic reproduction of the population. The economic motive follows the idea that family policy should guarantee the family's economic function in production and consumption,

<sup>1</sup>Although apparently influenced by sociological functionalism, which defined the (nuclear) family by its social functions, this discussed strand of the family policy literature does not operate with the same degree of normative reasoning.

and the societal motive is similar but considers reproductive functions as well. The sociopolitical motive then legitimizes family policies by the need to balance inequalities between parents and people without children, while the women's issues motive primarily aims to address mothers' disadvantages. Last, the children's welfare motive sees child well-being at the center of family policy intervention. Kaufmann argues that some of these motives overlap, and that some are used jointly, but that it would be short-sighted to assume one of the motives would cover the whole range of family policy issues (Kaufmann, 2002, p. 432). This reveals one of the key differences to the feminist tradition in comparative family policy research, the second one we consider here—although viewing feminism as a homogenous approach limited to "women's issues" is certainly too narrow a view.

The feminist tradition in comparative family policy research tapers in what has been called gendered welfare state research. This literature combines two research fields: feminist analysis of women's rights and comparative research of social policy and welfare states (Sainsbury, 1996). Gendered welfare state research has traditionally taken a critical stand. A common denominator of this literature was the underlying feminist ideas that patriarchal structures in society should be exposed and dismantled. A key concern was the omission of women's positions in mainstream welfare state research, and new concepts and theories were proposed to address this gap (e.g. Lewis, 1992; Lister 1994; O'Connor, 1993; Orloff, 1993; Pascall, 1996). Among the central goals was to unpack the various relationships between women and the welfare state. Theories addressed the male bias in theoretical explanations of social phenomena such as social policy provision and receipt. New concepts, then, enabled analyzing women as the targets and beneficiaries of welfare states.

However, rather than arguing that gender-centered measures of policies reflect what is good or bad for women, as often implied in critiques of this approach, this research highlights that the relationship between gender and the welfare state is complex (Lewis, 1997). A key theoretical contribution of the feminist perspective has been to understand the welfare state's role in affecting both paid as well as unpaid work. Including gender as an analytical category provided the conceptual tool for highlighting that work in the family was overwhelmingly done by women. Two concepts that enable a gendered analysis of welfare states are O'Connor's (1993) personal autonomy concept and Orloff's (1993) capacity to maintain an autonomous household or self-determination concept. While these can be applied to analyze the relationships between the welfare state and both men and women (Daly, 1994), they focus mainly on paid work (Lewis, 1997). In sum, many important concepts used in comparative family policy research evolved in the feminist research tradition, because it made reliance on the family to provide welfare visible as a distinguishing feature of welfare states.

Comparative welfare state research has traditionally lacked this focus on the family, looking instead at the relationship between capital and labor in welfare states with decommodification as a central concept. Decommodification provides wage earners with an alternative option to maintain a livelihood and is a power resource in wage bargaining processes. As a response to feminist criticism of the focus on paid work, representatives of this research tradition have later added concepts to account for work done in families to their initial frameworks (Esping-Andersen, 1999; Korpi, 2000). The initial classifications of welfare regimes that emerged in the power resources research tradition (conservative, liberal, and social democratic welfare regimes) were maintained even after adding the family as a welfare provider to the theories.

#### **Methodological Practices**

Comparative family policy research is a methodologically heterogeneous field (see Lohmann & Zagel, 2018 for a discussion on methodological approaches). Each of the above-discussed research traditions features a specific core methodological practice. In all traditions, comparative historical research of institutional development was important. This work is overall less concerned with defining quantifiable dimensions. In particular the feminist research tradition has taken a stance for comparative historical analyses. One argument in favor of this approach is that institutional settings are characterized by historical processes and complementarities which cannot be captured in single quantitative indicators.<sup>2</sup> The definition of welfare state models or welfare state types is based on an analysis of these historical processes. In the comparative welfare state research tradition both quantitative and comparative historical institutional analyses were used. In the past, quantitative comparative research has often relied on analysis of social expenditure data (Skocpol & Amenta, 1986), not least due to a lack of comparable institutional data. As a response, the establishment of the Social Citizenship Indicator Program (SCIP) at the Swedish Institute for Social Research at Stockholm University by Walter Korpi in the early 1980s was path-setting. From then on, other databases have emerged, including those on family policy indicators, although data availability remains an issue in the

<sup>2</sup>For example, Lewis argues that some quantitative studies fail to be convincing because their operationalization of central welfare state dimensions is too crude, such as operationalizing pronatalism with demographic variables (Lewis, 1997, p. 168).

field of comparative family policy research (Lohmann & Zagel, 2018; see also Chapter 24 by Sirén, Doctrinal, Van Lancker, & Nieuwenhuis in this volume). Quite often also quantitative research in the field aims at the identification of models or types. But here, country cases are assigned to types based on quantitative indicators or composite indexes. A well-known example is the use of a decommodification index to assign countries to welfare regime types (Esping-Andersen, 1990). In family policy research, typologies of familialism and de-familialism are examples for this approach (Leitner, 2003; Saraceno & Keck, 2010). Concepts such as familialism are, in principal, independent of methodological approaches. That means, most of the concepts discussed below cut through methodological divides and may be used in both caseoriented or variable-oriented approaches. However, it will become obvious that some are more usefully applied in one than in another research design.

# **Concepts of Comparative Family Policy Analysis**

Our discussion of concepts in comparative family policy research is structured around the distinction between two classes of concepts: categorical and gradual concepts. Categorical concepts characterize qualitative differences between countries, while gradual concepts are used to indicate (quantitative) degrees of institutional characteristics and policy intervention on an implicit or explicit scale.3 The former consider countries as cases, the latter as units of analysis for which values of variables are observed. First, we discuss typologies, such as the male breadwinner model and family support models, as a specific case of categorical concepts. Second, we give an overview of the conceptual discussion of gradual concepts such as defamilization, familization, and related ones. Figure 6.1 gives an overview of the concepts we selected for discussion in this chapter, illustrative for the differences between the approaches. It also shows the interlinkages between concepts. In particular, overlaps between categorical concepts (typologies) and gradual concepts will become clear throughout, as the latter is often used for generating the former.

<sup>3</sup>This distinction should not be confused with contrasting quantitative and qualitative methodologies.

**Fig. 6.1** Examples of categorical and gradual concepts in comparative family policy research (*Note* See text for further details and references)

#### **Typologies as Conceptual Devices**

Typologies are popular and used extensively for defining theoretical ideal types in comparative welfare state research. Mainstream welfare state typologies have widely been criticized for their ignorance of gender inequality. New theoretical concepts were needed that considered policy effects on the gender division of labor. According to Sainbury (1994), two approaches were used to respond to the critique of gender-blindness: the first one argued for integrating gender and family alongside class concepts into the mainstream theories; and the second one suggested to dismiss the country groupings in mainstream research while producing typologies purely based on gender and family (see Hook, 2015 for a similar argument). The two responses developed successively.

As an immediate critique of Esping-Andersen's typology, Lewis (1992) introduced the concept of the male breadwinner model for typologizing welfare regimes. Second, based on this work, Esping-Andersen (1999) and Korpi (2000) included family and gender into their comparative frameworks. While the former added the defamilization concept to refine his *Three Worlds* typology, Korpi (2000) defined different types of family support models to form distinct regime types in combination with class inequalities. We will now discuss the male breadwinner concept and the family support models in turn. Defamilization and related concepts will be discussed in the following subsection.

#### **Male Breadwinner Model**

Ever since the concept emerged in the 1990s, the male breadwinner model is a key reference point in comparative family policy research. As a precursor in this field, Langan and Ostner's (1991) study draws on the very ideas underpinning the concept, but was still lacking a systematic conceptual framework. Langan and Ostner assess Esping-Andersen's typology by asking how policies affect the gender division of labor, focusing on the degree of women's economic dependency on a male earner across countries. Lewis (1992) then introduces the concept of the *male breadwinner model* as an analytical framework for cross-country comparison. The concept builds on the 1980s discussion around women's rights and the family wage (Land, 1980). Analyzing four countries' adherence to the model, Lewis (1992) finds that Britain and Ireland are historically strong breadwinner countries, France a modified and Sweden a weak breadwinner state. This initial formulation of a breadwinner model typology counts as one of the key feminist responses to Esping-Andersen's welfare regime typology. Country groupings were later added and adapted (e.g., Lewis & Ostner, 1994).

The male breadwinner model is broadly defined as the "idea and norm of a 'male breadwinner' and a 'secondary' female wage earner (…) built into the welfare system" (Lewis & Ostner, 1994). The breadwinner model typology rests on the conviction that any classification of regime types must incorporate the relationships between paid and unpaid work and welfare. The argument is that a sole focus on paid work omits the gendered nature of welfare provision. The concept of the breadwinner model as used by Lewis (1992) reflects welfare states' different degrees of support to the gendered division of labor, by granting social rights to women primarily as wives or as individuals. It sometimes appears difficult to pin down exactly what the breadwinner model is. Although it is clear that the core of the concept is about differences in policies, it sometimes seems to describe the state of gender relations in society more broadly. In this aspect it overlaps with Pfau-Effinger's (1996, 2005) gender arrangement concept, which is however critical of "institutionalist" perspectives, and more explicitly based on cultural differences between countries.

The breadwinner model concept is used to describe country differences and change over time within countries in the relationship between paid and unpaid work and the welfare state. Since the initial formulation of the concept, the erosion of the male breadwinner model in European policies has been noted (Lewis, 2001; Lewis & Giullari, 2005). In ensuing discussions, the model has been juxtaposed to what is called an *adult worker model* . Policies reflecting the adult worker model address adults as employable individuals rather than men and women as spouses with specific roles and responsibilities within the family. Governments promote an adult worker model in the hope of it fostering economic growth, tackling poverty, increasing gender equality in the labor market, and addressing the problem of deteriorating dependency ratios (Lewis, 2006, p. 9). Hence, adult worker model policies assume that men and women should claim social entitlements on the basis of "sameness" rather than on the basis of "difference" as is the case in a strong male breadwinner state (Lewis & Ostner, 1994).

One of the main questions raised by the observers of potential shifts from male breadwinner to adult worker model is whether policies that are designed to support an adult worker model are also gender-sensitive (Lewis, 2009). That means, for example, whether work–family reconciliation policies with the aim of bringing women into employment consider gender equality. Or whether such policies go at the expense of women who are now expected to take on the same level of paid work without being relieved from unpaid work. Lewis and others find that "gender equality" has been used mostly rhetorical and instrumental rather than as a genuine policy goal (Daly, 2011; Lewis, 2009; Lewis & Giullari, 2005). This discussion again points to the difficulty of using the breadwinner model concept for comparing institutional settings across countries, because often it has more in common with an ideational concept.

This fuzziness notwithstanding, research continuously refers to the male breadwinner model concept.<sup>4</sup> It has proved to be of some relevance in the comparative family policy literature. Most empirical studies explicitly analyzing the breadwinner model typology take a historical perspective focusing on a small number of country cases (Lewis, 1997). There are however several large-N studies that group countries based on indicators of women's paid work. Few studies actually look at household level work–family arrangements (Hook, 2015), as would be adequate for a breadwinner model perspective. One exception is Lewis, Campbell, and Huerta (2008), who descriptively examine empirical patterns of paid and unpaid work in couples across Europe, but without attempting to identify clusters.

<sup>4</sup>Lewis (1992): 162 citations in 2018 according to google scholar as of March 2019.

#### **Family Support Models**

The second prominent conceptual perspective in research using typologies for comparing family policies across countries explains welfare state differences with variations in *family support models* (Korpi, 2000). This strand originated in the comparative welfare state research tradition. It is characterized by an attempt to incorporate gender as an analytical category into existing comparative welfare state frameworks. Korpi's (2000) approach is strongly influenced by the feminist critique of Esping-Andersen's 1990s study. One of his main contributions is, however, to consider gender inequality by simultaneously accounting for class inequality in the classification of welfare regimes. Korpi argues that "effects of welfare states on patterns of inequality in terms of gender and class can be asymmetrical" (2000, p. 142). That means, for example, welfare states with generous redistribution between high- and lowincome groups are not necessarily those that are most effective in reducing gender inequality.

The family support model concept distinguishes three distinct family support models: the general family support model, the market-oriented support model, and the dual-earner support model. Korpi applies a social rights perspective situated in the state-market-family triangle (Korpi, 2000, p. 144). The general family support model fosters the traditional gender division of labor while dual-earner support enables women's employment and the socialization of care work. Welfare states that provide neither support to the nuclear family, nor to the dual-earner family are assumed to have opted for family support to be provided by the market. Korpi argues that these theoretical ideal types reflect both policies and policy goals regarding gender equality (ibid.).

Two characteristics are central in Korpi's (2000) approach, the focus on institutions and its multidimensionality. The focus on institutions is justified by the alleged need for welfare state typologies to focus on institutional indicators rather than outcomes (Korpi, 2000). From this point of view, in particular some approaches in the feminist tradition are guilty of conflating institutional and outcome perspectives. For example, the male breadwinner model reflects both, policies (institutions) and women's employment (outcome). Further, Korpi and his colleagues stress that their approach is multidimensional (Korpi, 2000; Korpi, Ferrarini, & Englund, 2013). In his early study, Korpi explains that family policy has to be conceptualized and measured focusing on more than one dimension, because policymakers face the fundamental dilemma of whether to grant women the same or different kinds of rights than men (Korpi, 2000, p. 140). Hence, if one policy dimension considers support to women's employment, another one has to account for support to unpaid family care. Family policy multidimensionality is consequently conceptualized as the position of a country in the space between the two dimensions of general family support and dual-earner support (Korpi, 2000; Korpi et al., 2013). With this, Korpi's typology results from crossing two gradual concepts (see Fig. 6.1).

Ferrarini (2006) adopts the framework proposed by Korpi precisely for the advantages of a multidimensional perspective. He adds the "Contradictory family policy model" to the ideal types of family policy models where both general family support and dual-earner support are generous (Ferrarini, 2006). The notion of contradiction (also termed pluralistic, ibid., p. 13) reflects the assumed underlying policy ideology. Hence, while Ferrarini (2006) builds his typology on the same gradual concepts as Korpi, he deduces an additional type from the combination of specific values on these dimensions.

#### **Gradual Concepts**

In contrast to using such concepts as stepping stones for creating typologies, gradual concepts are also often used without higher order aims (typologies). In this section we consider defamilization and its offshoots as a representative of a second class of concepts in comparative family policy research. Defamilization shares theoretical roots with other concepts such as the male breadwinner model. As the latter, the concept of defamilialization (Lister, 1994) or de-familization (McLaughlin & Glendinning, 1994) emerged from the feminist critique of Esping-Andersen's welfare state analysis and its central concept of decommodification. In this section, we map the discussion around the concept of defamilization, and we review other concepts that evolved from it. Underpinning the concept of defamilization is (again) the idea that the family is a welfare provider, but that the family is also characterized by strong social and economic dependencies among its members. By providing certain policies, welfare states can alter the degree of family dependence and with it the relative importance of state, market, and family as welfare providers. These links are at the heart of the defamilization concept.

The interpretation of the defamilization concept varies, but the different possible readings go rather unnoticed in the literature. The relative lack of systematic discussion of the concept's theoretical roots and conceptual facets is striking given its popularity in comparative research (but see e.g. Kurowska, 2018; Leitner & Lessenich, 2007; Lohmann & Zagel, 2016, Zagel & Lohmann, (forthcoming). The concept seems to polarize into those who use it unquestioned and those who are skeptical about its use. The latter position manifests in studies that take defamilization as a starting point for defining alternative or additional concepts. These concepts are generally less comprehensive, or focus on a certain aspect of defamilization, like for example the concepts of dedomestication (Kröger, 2011), degenderization (Saxonberg, 2013), and demotherization (Mathieu, 2016). Often, the motivation for defining these offshoot concepts seems to result from differences in uses and understandings of the defamilization term. The systematic discussion of such differences, to which we contribute in the following, is likely to resolve some of the issues.

## **Defamilization**

There are at least three influential definitions of defamilization. The first was proposed by Lister who defines defamilialization as "the degree to which individual adults can uphold a socially acceptable standard of living, independently of family relationships, either through paid work or through the social security system" (1994, p. 37). McLaughlin and Glendinning coined the second definition: "[D]e-familization is constituted by those provisions and practices which vary the extent to which wellbeing is dependent on 'our' relation to the (patriarchal) family" (1994, p. 65). In contrast to Lister, the definition by McLaughlin and Glendinning does not put such a clear emphasis on economic independence. A third definition was proposed by Esping-Andersen who adopted the concept as a reaction to the critique of his earlier focus on decommodification for welfare regime comparison: "I shall use 'de-familialization' […] to capture policies that lessen individuals' reliance on the family; that maximize individuals' command of economic resources independently of familial or conjugal reciprocities" (1999, p. 45). While Lister and McLaughlin and Glendinning originally introduced the concept to comparative social policy research, Esping-Andersen transferred it into what has been referred to as mainstream welfare state literature. Later uses of the concept are often strongly influenced by Esping-Andersen's reading of the earlier proposals. Clearly, Esping-Andersen's definition stresses the aspect of economic independence while, in particular, McLaughlin and Glendindinning's definition includes policy effects on other family dependencies as well. Drawing on this wider definition Leitner and Lessenich (2007) argue that defamilization is a multidimensional concept. That is, state and market would not only affect economic, but also social dependencies, which they understand as relationships between caregivers and receivers of care. This multidimensional definition is not widely used in the literature, nor is there a consensus on differentiating the social and economic dimensions of defamilization. The analytical distinction makes sense because social and economic outcomes of policies are often intertwined.

There are two indications of the lack of conceptual clarity of defamilization. First, the incoherent spelling of the term in the literature, and second, the lack of precision in quoting the original sources of the concept. As regards the spelling, neither the occasional deviation from the original spellings, nor the intentional use of a specific spelling that an author makes explicit would be so surprising. It is the uncommented use of particular ways of spelling that suggests arbitrariness. For illustration, we found four different ways of spelling: defamilialization (Lister, 1994; Mathieu, 2016; Saxonberg, 2013), de-familialization (Esping-Andersen, 1999; Leitner, 2003; Saraceno & Keck, 2010), de-familization (McLaughlin & Glendinning, 1994; Saraceno, 1997), and defamilization (Cho, 2014; Kröger, 2011; Lohmann & Zagel, 2016). We risk over-interpreting the differences if we were to assume different intended meanings by these choices of spelling, especially regarding the use of a hyphen. And yet, we would argue that a systematic discussion of these differences in terms of substantive meanings could contribute to a more coherent use of the concept. Second, the imprecise reference to the original sources does not help to reach an agreement on the concept. For instance, Esping-Andersen (1999) does not refer to either of the two original proposals by Lister (1994) and McLaughlin and Glendinning (1994). Other authors seem to mix up sources and quotations of the early publications on the concept (comparing, for example, Mathieu, 2016, p. 578; Saxonberg, 2013, p. 28f ). As a consequence, in these discussions, differences in the initial definitions by Lister (1994) and McLaughlin and Glendinning (1994) were blurred. Furthermore, new aspects and interpretations have been added to the concept.

Most importantly, a number of scholars have pointed out that familization and defamilization should not be framed as negatives or as opposite poles on a continuous scale (Eggers, Grages, Pfau-Effinger, & Och, 2020; Leitner, 2003; Lohmann, 2009; Lohmann & Zagel, 2016), but may be regarded as two related but distinct concepts. The fact that a number of welfare states combine familizing and defamilizing policies shows that the two are not mutually exclusive (e.g., in the form of optional individualism or optional familialism; Leitner, 2003; Lohmann & Zagel, 2016). Related to this, the concepts of familization and defamilization have been used jointly for defining types of familialism, de-familialism, or individualism (e.g., Leitner, 2003, see Fig. 6.1). Defamilization then came to capture (quantitative) degrees to which certain policies provide relief from family dependencies, while types of de-/familialism characterize qualitative differences between combinations of such policies. Gender and intergenerational (i.e., children–parents–grandparents) dependencies are at the heart of the debate around defamilization. And although other characteristics such as sibling order, matri- or patrilineal lineage are usually sidelined, the concept of defamilization is principally open to them.

A crucial but unresolved question in the discussion of defamilization is whether only the state or also the market is considered as sources for independence from the family. Economic independence from the family is often construed as commodification, i.e., reliance on the market for acquiring own income as an alternative to depending on family members' economic support. In contrast, social independence is more often discussed as being achieved via state provision of services. That means, statutory provision of care services such as childcare is defamilizing because it relaxes social dependence between caregiving and care receiving family members. However, the distinction between state or market as alternative providers to the family is not as clearcut. For instance, Lister (1994, p. 37) explicitly mentions "the social security system" as alternative to "paid work" as source of economic independence. A second example is Leitner, who focuses not only on economic but also social independence, discusses—with reference to Esping-Andersen (1999)—also "de-familialization through market driven service provision" (2003, p. 356f ). For example, childcare services are often provided by for-profit organizations.

Against the background of this wide, multidimensional understanding of defamilization we now discuss more recently proposed concepts such as dedomestication (Kröger, 2011), degenderization (Saxonberg, 2013), and demotherization (Mathieu, 2016). While the first and the third are construed as additional concepts to defamilization, the second is proposed as an alternative one.

#### **Dedomestication**

Kröger's (2011) proposal of an additional concept labelled "dedomestication" is based on a thorough discussion of the literature on defamilization. He recognizes in detail the differences in the concepts as proposed by Lister (1994) and McLaughlin and Glendinning (1994). Consequently, he borrows from Leitner and Lessenich (2007) the distinction between economic and social independence as well as their critique of a narrow, economic understanding of defamilization in the mainstream welfare state literature. Although Kröger adopts a wide notion of defamilization that includes both economic and social aspects, and argues strongly for the integration of the perspectives of caregivers and receivers of care, he rejects the use of the term defamilization. Instead, he argues that defamilization tends to be used more narrowly in the literature, and that using it in a different, more encompassing way would cause "conceptual confusion" (Kröger, 2011, p. 429). Therefore,—despite his own understanding of the concept—he opts for the use of "defamilization" to describe the degree of economic independence and proposes "dedomestication" to describe the degree of social independence only (see Fig. 6.1).

Dedomestication "is understood as the degree to which social care policies make it possible for people to participate in society and social life outside their homes and families. Dedomestication thus refers to freedom from a confinement to the domestic sphere or, using the term of Lewis (1997), to citizens' right to limit their engagement in unpaid informal caring. From the care receiver's perspective, dedomestication is about not being dependent on informal care from close persons, in other words, about citizens' right to receive formal care (Knijn & Kremer, 1997)" (Kröger, 2011, p. 429). With its explicit reference to the right of participation within the public sphere and of limiting unpaid informal caring, Kröger (2011) addresses the gendered consequences of caring arrangements. Still, dedomestication is not predominantly about gender. By stressing the rights of receivers of care the concept also explicitly addresses intergenerational dependencies. Although analytically convincing, the concept has not been picked up widely. Its measurement has seemingly never evolved from the rather tentative sketch in the original article based on data on childcare for zero- to two-year olds—leaving out indicators for the care provisions for older children, disabled and older people due to a lack of data. Kröger himself critically comments on the value of the proposed index: "The end result of this empirical experiment is another ranking order of welfare states, based on the level of dedomestication of their formal childcare provisions for children under three years. It can be questioned if such league tables have any theoretical or policy significance" (2011, p. 436).

#### **Degenderization**

In contrast to Kröger, Saxonberg (2013) applies a narrow interpretation of defamilization in his proposal of replacing the concept with that of degenderization. A selective reading of Lister (1994, 2003) and Esping-Andersen (1999) results in a purely economic understanding of defamilization. Saxonberg overlooks the social dimensions of defamilization despite referring to Leitner (2003) whose main focus is on the caring function of families. He interprets defamilization as a concept addressing primarily gender relations, and even more, as a normative concept directed at creating gender equality. Overall, Saxonberg arguably takes an unnecessary normative stance in what claims to be an analytical discussion. Arguing from this position, he states that researchers "are not even able to reach an agreement as to whether familialization is bad and defamilialization is something good!" (Saxonberg, 2013, p. 29). This bold statement disregards the careful discussions about the complex relationships between gender and the welfare state in this research area (Daly, 1994; Lewis, 1997). An illustrative example is the concept of "optional familialism" (Leitner, 2003), which is based on the understanding that preferences for caring do not need to be in conflict with policies lowering the burden of care. In other words: contrary to Saxonberg's claims, single policies are not to be judged unidirectionally but in terms of the options they offer. Narrowing down previous proposals to gendered economic dependence (see Fig. 6.1) leads Saxonberg to a general critique of the concept of defamilization as "simply too ambiguous" and that it "does not really clearly describe the main goals of mainstream feminist scholars writing on gender" (2013, p. 27). Instead, he proposes "degenderization" as a more adequate concept.

In our reading, the concept of degenderization fails to be convincing, because it is a narrowed down, normative version of defamilization. Kurowska comes to a similar, slightly more favorable conclusion. She "supports most of Saxonberg's critical arguments towards (de)familialization when it is understood (and applied) as a comparative and evaluative perspective focused on gendered aspects of welfare states." (2018, p. 30). However, Kurowska also stresses that (de)familialization may be interpreted more comprehensively, and as such is able to provide useful insights into the role of the welfare state that go beyond gender issues, particularly for securing the welfare of children and the elderly. She therefore argues in favor of the mutual irreplaceability of the two concepts.

### **Demotherization**

A third new concept—demotherization—was recently proposed by Mathieu (2016). It is defined as follows: "The concept of demotherization refers to the degree of independence mothers enjoy from the necessity of performing care work, and most specifically childcare" (Mathieu, 2016, p. 577). Although Mathieu shares the focus on gender equality with Saxonberg (2013), she acknowledges the multidimensionality of defamilization. And although the title "From the defamilization to the 'demotherization' of care work" seemingly refers to demotherization as an alternative, Mathieu is clear in proposing it as an additional concept. Mathieu—like Kröger (2011) with his proposal of dedomestication—adds a concept which is focused on care relations and, thus, on the social dimension of defamilization (see Fig. 6.1).5 In contrast to Kröger, Mathieu explicitly genders dedomestication as it "does not account for the fact that care work can be performed within the family, by other relatives than by the mother" (Mathieu, 2016, p. 588).

## **Conclusion and Outlook**

Comparative family policy research relies on concepts by which differences in countries' reliance on the family as a welfare provider can systematically be evaluated. The relationship between the welfare state and the family is a key aspect in any conceptual approach to comparative family policy research. Concepts however vary in their focus on different aspects of this relationship.

Our discussion was guided by six criteria we laid out in the introduction. (1) Who is considered the main addressee of family policy interventions in the concept? The main distinction is whether families or individuals living as families are considered as main addressees. Where individuals are addressed, these can be children, women, men, mothers, fathers, parents, and grandparents. Most prominently, concepts arising from feminist welfare state critique have brought forward a perspective of individual's—men and women's social rights, a perspective which all recent concepts share to a stronger or lesser degree. Children as family members with individual rights are less explicitly discussed but have obtained more attention recently (Daly, 2020; Palme & Heimer, 2019). (2) Does the concept focus on gender and/or intergenerational relationships? This question relates to the first but focuses on social relationships and the dependencies arising of those. While some concepts focus explicitly on the relationships between women and men, others integrate intergenerational relationships too, such as defamilization in a broad reading or family support models. (3) Is the concept defined from the perspective of the caregiver and/or the care receiver? Concepts, such as Leitner's types of familialism or Kröger's concept of dedomestication, are rooted in the analysis of care relationships. These concepts are more sensitive to include the perspective of caregivers and receivers of care. (4) Does the

<sup>5</sup>Mathieu uses gradual concepts—de-/motherization and de-/familization—to define different types of maternalism, an approach similar to Leitner (2003). For the sake of readability, this is not depicted in Fig. 6.1.

concept consider the state and/or the market as welfare providers alongside the family? All concepts share the view of family, state, and market as welfare providers. The question whether the state or the market are sources for independence from the family is not addressed explicitly in all concepts. (5) Does a concept point to ideal types in family policy regimes? We distinguished between gradual concepts and categorical concepts. The latter categorize countries into types which may also be interpreted as ideal types (although the distinction between ideal and real types is often blurred). Some categorical concepts use the combination of gradual concepts as basis for a typology. (6) Are there are any subdimensions to the concept? Categorical concepts are by definition multidimensional as they consider countries as cases, i.e., as units defined by a combination of, often complementary, characteristics. Thus, the question of multidimensionality arises only in the discussion of gradual concepts. We argued that defamilization is understood as unidimensional as well as multidimensional concept, which is a root for misunderstandings in the reception of the concepts since it was proposed.

In this chapter, we have shown that the different concepts developed in distinctive research traditions, which featured specific methodological practices. We further discussed the specifics of some of the most prominent concepts and how they developed over time. One of the conclusions that we draw from this evaluation of the literature is that there are many similarities in the underlying ideas behind the concepts. However, there are also important differences, which so far have partly gone unnoticed mostly because of a lack in systematic discussion.

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **7**

# **Conceptualizing National Family Policies: A Capabilities Approach**

**Jana Javornik and Mara A. Yerkes**

Comparative family policy research has advanced significantly in recent years. The growing availability of more and better data have improved our understanding of cross-national similarities and differences in family policies, as well as how they shape the lives of different families and children, also evidenced by various chapters in this handbook. Despite advancements, comparative family policy research continues to face difficulties. For example, empirical analyses largely include "standard" measures tailored to dual-earner heterosexual couples because the multifaceted nature of family policies makes cross-country comparisons complex. Such analyses fail to reflect policy logics, and few studies use legal formulations to capture the state's underlying assumptions and differential statutory entitlements (Javornik, 2014; but see Chapter 12 by Skinner & Hakovirta and Chapter 16 by Evertsson, Jaspers, Moberg in this volume). Moreover, comparative policy analyses often face problems conceptualizing family policies as well as operationalizing them into measurable indicators (Yerkes & Javornik, 2019; see Chapter 6 by Zagel & Lohmann in this volume).

J. Javornik (B)

M. A. Yerkes Department of Interdisciplinary Social Science, Utrecht University, Utrecht, The Netherlands e-mail: m.a.yerkes@uu.nl

© The Author(s) 2020

Leeds University Business School, University of Leeds, Leeds, UK e-mail: J.Javornik@leeds.ac.uk

In this chapter, we offer an innovative approach to comparing family policies using the capability approach (capability approach). From a capabilities perspective, individuals are embedded in broader, relational contexts; these ecological and social contexts shape the real opportunities individuals have (Javornik & Kurowska, 2017). Thus, what individuals are really able to do and be is a reflection of their capabilities, their agency, and choice (Robeyns, 2017), within the diverse contexts in which individuals are embedded (Hobson, 2014; Hvinden & Halvorsen, 2018). Applying the *capability approach* (Nussbaum, 2000, 2011; Robeyns, 2005, 2017; Sen, 1992, 1999a) to family policy analysis offers a promising evaluative perspective for comparatively investigating the extent to which policy design empowers parents to freely use its instruments, or, to what extent its design enhances the capabilities of parents; in other words, what parents are truly able to do and be (Sen, 1992).1 Using examples of childcare policies in Sweden and the UK, a policy area particularly prone to conceptual challenges, we discuss the difficulties involved in conceptualizing family policies in comparative research. Childcare services, as a key component of family policy (Yerkes & Javornik, 2019), intertwine with other policy domains such as employment, gender equality, demography, and the tax system. Such policy complexity creates comparative challenges for empirically investigating childcare services across countries. It also creates difficulty in disentangling what key aspects of childcare policy mean for different groups of parents. The latter is crucial, as childcare service provision shapes the conditions under which parents can access and engage with employment or other opportunities (e.g., education). The two countries investigated here, Sweden and the UK, differ in one salient aspect: while Sweden uses a public-provision mechanism through demandpriming approach, the UK has been reforming its supply-led, marketized childcare system (Yerkes & Javornik, 2019). This, as shown in our earlier work, has significant ramifications for parents' childcare capabilities.

The capability approach is increasingly used in social policy research (Yerkes, Javornik, & Kurowska, 2019) and in family policy scholarship in particular (see also chapter 19 by Schober in this volume). For example, it is applied in relation to work-family policy (Chatrakul, Ayudhya, Prouska, & Beauregard, 2017; den Dulk & Yerkes, 2016; Fahlén, 2013; Hobson, 2014; Korpi, Ferrarini, & Englund, 2013; Yerkes & den Dulk, 2015), more recently to parental leave policy (Javornik & Kurowska, 2017; Javornik & Oliver, 2019; Koslowski & Kadar-Satat, 2019; Kurowska & Javornik, 2019) and in reconceptualizing the (de)familialism perspective in comparative family

<sup>1</sup>While we recognize that family policy affects children's well-being, our focus here is on parental perspective.

policy research (Kurowska, 2018). To date, only one study has applied the capability approach to comparative childcare policy (Yerkes & Javornik, 2019), and this chapter builds on this earlier work. We start by outlining the capability approach, and the advantages of using the capability approach for evaluating family policies in comparative perspective. We then conceptualize key policy aspects taking the capability approach perspective and discuss their various implications along gendered and classed lines.

# **The Capability Approach: Background and Application**

Sen developed the capability approach by drawing on the moral and political philosophy of Aristotle, Smith, and Marx (Sen, Hanžek & Javornik, 2002). This perspective shows a dual concern: a philosophical concern for social justice and the human good, and an economic concern for measuring life quality, as well as promoting autonomy and pluralist individual life choices. The political philosophy of Aristotle, whose key principle was human flourishing as ethically fundamental, lies at the heart of the capability approach (Nussbaum, 1987). Further developed through moral and political philosophy, the capability approach values pluralist life courses (Robeyns, 2017), promoting the idea of individuals "in need of a totality of life activities" and real freedoms to pursue those (Nussbaum, 1987). Ultimately, the freedom to achieve well-being is of moral importance and viewed in relation to people's capabilities—their real opportunities to do and be what they have reason to value (Yerkes et al., 2019). Emphasizing capabilities, or individual freedom to achieve a wider range of valued outcomes, shifts the focus away from purely economic measures of utility toward other valued outcomes and individual capabilities to pursue these activities. Work-family scholarship, for example, shows couples may value relationship harmony over gender egalitarian divisions of labor (Baxter, 2000; Thompson, 1991). Similarly, mothers may be willing to sacrifice career opportunities in the short term because they value flexible working arrangements upon returning to work after childbirth (Yerkes, Martin, Baxter, & Rose, 2017). The capability approach is a flexible and multi-purpose *framework* (Sen, 1992, p. 48), useful for moving comparative analysis beyond childcare policy as a means for work-family reconciliation toward understanding childcare policy in relation to parents' capabilities to live the life they have reason to value. To apply the capability approach effectively in family policy research, further specification of the framework is needed. Namely, the capability approach is "*open*-*ended* because the general capability approach can be developed in a range of different directions, with different purposes, and it is *underspecified* because additional specifications are needed before the capability approach can become effective for a particular purpose" (Robeyns, 2017, p. 29; emphasis in original).

# **Key Elements of a Capability Approach to Family Policy**

The capability approach (Sen, 1992, 1999a) centers around multiple concepts, with five being key to its application: means, capabilities, functionings, conversion factors, and agency (cf. Robeyns, 2005). *Means* are the various resources (e.g., economic, social) to which individuals have access. Inequality can arise due to variation in means. However, even when individuals have equal access to means, they may not have the same *capabilities*, i.e., real opportunities or potential, to live the life one has reason to value. Capabilities, what people are able to do or be (Robeyns, 2017) thus differ from what individuals actually achieve (*functionings, or achieved functionings*). Inequalities can arise out of variation in means but also because individuals are not equally able to translate means into capabilities. The translation from means into capabilities is shaped by *conversion factors*, i.e., the multiple contexts in which individuals are embedded at the personal, community, and societal level (e.g., Robeyns, 2005, 2017; Yerkes, Hoogenboom, & Javornik, 2019). Personal conversion factors include aspects such as gender, class, race/ethnicity, age, and health. Community level conversion factors refer to relationships at the local level (e.g., social networks) as well as the places and spaces in which individuals are embedded (Yerkes et al., 2019). Societal level conversion factors include social norms and social movements (Hobson, 2014). Combined, these conversion factors enhance or constrain the set of real opportunities individuals can choose from (the *capability set* in capability approach terms). Inequalities in what individuals are able to achieve can also be attributed to inequalities in *agency*. *Agency* or the way in which individuals perceive, interpret, and respond to their social situation (Giddens, 1984; Mead, 1934) can be seen as the freedom to act, and reflects individual's active response in social situations (Sen, 1999b; Shaw, 1994). The reflexive interaction with the world (agency—structure) can be seen as a mutual constitutive process of structuration (Giddens, 1984). For example, men and women from different social classes may use family policy differently because gender and class inhibit or enhance their agency. At the same time, gender and social class as a social structure may shape individual behavior. The capability approach emphasizes such relational aspects, seeing individuals with differing freedoms to act (inequalities in *agency*) as relationally embedded in personal and social contexts (*conversion factors*).

A successful application of the capability approach in comparative family policy research requires first and foremost accounting for the role of policy itself. At present, social policies, including family policies, are largely interpreted and applied in three ways: as conversion factors and as structural constraints (Hobson, 2014; Robeyns, 2017), and/or as a *means* to facilitate capability (Javornik & Kurowska, 2017; Kurowska, 2018; Yerkes & Javornik, 2019; Yerkes et al., 2019). Traditional applications of the capability approach (e.g., Robeyns, 2005, 2017) view policy as a structural constraint. In this scholarship, policy is viewed as an interdependent set of measures and instruments aiming to change human behavior and/or improve quality of life and well-being. Initial applications of the capability approach to family policy, such as Hobson's (2014) capabilities framework for work-life balance, treat policies as part of the social context, whereby parents' use of policies is seen to differ based on their perceived set of alternatives available and the sense of entitlement to use these policies (Hobson, 2018). Later applications by Hvinden and Halvorsen (2018) and Kurowska (2018) suggest policies are *both* conversion factors and means. For example, parental leave and childcare policies can function as a means that facilitates parent's capabilities to arrange care in a way they have reason to value (Javornik & Kurowska, 2017; Kurowska, 2018; Yerkes & Javornik, 2019; Yerkes et al., 2019). What role policy plays depends upon the context of the research question at hand (Hvinden & Halvorsen, 2018; Kurowska, 2018), similar to the differing roles variables can take on in empirical research (e.g., mediator, moderator).

For the purpose of this chapter, we see family policy as a *means* for parents to arrange childcare in a valued way. The translation of this means into real opportunities is shaped by gender and class, thereby affecting parents' capabilities (Yerkes & Javornik, 2019). These capabilities will differ across parents. For some parents, public childcare creates the means to better reconcile work and care (Gornick & Meyers, 2003; Grönlund & Javornik, 2014); for others, it creates the means to use childcare as an aspect of children's development needs and socialization (OECD, 2017b; Saraceno, 2011; van Huizen & Plantenga, 2015). Equally, it reduces the need for family care to enable other meaningful contributions to society (e.g., education, training, job search, or volunteering) (Javornik & Ingold, 2015). Conceptualizing policies in this way means understanding family policies in diverse, historical, and political contexts (Ginsburg, 2004) as value-laden, developed based on culturally informed, dominant ideas (Béland, 2005, 2016). Crucially, policies provide normative reference points (Goerne, 2010; Javornik, 2014) that set the "rules of the game" (North, 1990). That is, policies define what means are available to parents in a specific country, and the capability approach helps to analyze whether they help individuals to achieve that normative reference point.

However, individuals do not have the same real opportunities (*capabilities*) to achieve varying life pursuits, which leads to inequality in outcomes, or *achieved functionings* (Sen, 1992). Social policy scholars largely view capabilities in Sen's (1992) terms of *valued functionings*, or the real opportunities individuals have to pursue a life they have reason to value (Yerkes et al., 2019). Similar interpretations can be useful for applications to family policy. For example, an individual may value being a carer and places greater value on providing care than on participating in paid employment. The capability approach rests on the idea that individuals have an array of valued functionings, reflecting diverse needs and desires. Individuals are not equally able to pursue these valued life activities, which leads to inequalities in outcomes or *achieved functionings*. As outlined by Robeyns (2017, pp. 41–45), capabilities and functionings can be either positive or negative, and thus must be viewed as essentially value neutral in the abstract sense. While in some cases we might be able to distinguish positive functionings (e.g., gender equality at work) or negative functionings (e.g., female double burden), the value of functionings is often ambiguous. Assuming that family policy aims to facilitate a positive functioning for most parents, this same policy can lead to unintended negative functionings for some groups or individuals.

How childcare policy translates to parents' capabilities depends on conversion factors at multiple levels (as above). For example, parents wishing to reconcile work and care often do so within the context of their workplace. Organizational practice, with dominant norms of an ideal worker, influences one's "sense of entitlement" to make use of available policies (Grönlund & Javornik, 2014; Hobson, 2014). While we recognize the role of cultural, societal, community, and organizational contexts as well as personal history and circumstances as conversion factors in parents' capabilities, in this chapter we focus on gender and class, two key individual-level conversion factors operating in intersectional ways (e.g. Hook, 2015; Korpi et al., 2013). Namely, childcare costs constrain childcare capabilities of lower socio-economic classes, limiting their ability to choose high-quality childcare (Morgan, 2005). Similarly, low accessibility or availability can make childcare exclusive, maintaining gendered patterns of care (Kreyenfeld & Hank, 2000). Given educational homogamy among couples (Steiber & Haas, 2009), childcare capabilities relate to the intersection of class and gender. With education as a proxy for class in work-family arrangements (Hook, 2015), mothers with higher levels of education have more opportunities to arrange childcare, relative to mothers constrained by economic need (Steiber & Haas, 2009). In contrast, highly educated mothers generally have stronger labor market attachment and higher opportunity costs from opting out of work and staying home to provide care (Hook, 2015). Our approach allows us to highlight variation and any tensions between key aspects of childcare policy and how this affects parents in relation to gender and class.

Evaluating childcare policies from a capabilities perspective means not only viewing individuals in relation to the social spaces in which they are embedded, but also the relational nature of social rights embedded in childcare policies. In the former, individuals may be supported or limited in their capacity to access childcare as a means to achieve a valued outcome given personal, social, or environmental factors (Hvinden & Halvorsen, 2018). Thereby, childcare policies inherently create inequalities through a process of exclusion given varying degrees of selectivity. However, policies are relational and interdependent and often accessible only through other policies. In the UK, for example, subsidised childcare for under-3s is only available to families in receipt of low-income benefits, i.e., childcare is dependant upon social policy criteria. It is thus unavailable/unaffordable to many parents who may wish or need childcare sooner, which affects the demand for parental leave (similarly to other countries). These examples demonstrate the relational aspect of the capability approach, focusing on the interconnectedness of parents' decision-making about childcare. It demonstrates that policy can be part of a broader social context that shapes individuals' access to other policy instruments. Against this backdrop, the question of accessibility becomes central because it shapes parents' freedoms, and is thus indicative of *distributive justice*. Namely, to fully use the policy, one needs to be aware of a web of policy options and be able to navigate the legal landscape. This entails understanding (1) the policy process (*functional literacy*), (2) the relational aspects of social rights (i.e., how one affects another), and (3) the power dynamics between them (which and whose right superceeds) (Yerkes et al., 2019).

Our suggested application of the capability approach in family policy research builds on previous work in this area. Earlier applications of the capability approach have been framed as challenging a dominant utility perspective and highlighting differential constraints on choice and agency. Specifically, in relation to parental leave, Javornik and Kurowska's (2017) analytical framework aims to enable a comparative analysis of parents' real opportunity to take parental leave in the context of given policy entitlements and the impact of social norms in relation to both gender and class. Javornik and Oliver's (2019) legislative analysis of the UK shared parental leave seeks to understand leave in a broader legal context, which created "material and discursive opportunities" for shared parenting. Hobson, Fahlen and Takacs (2011, p. 169) consider discursive space in the context of workfamily policies and how policy entitlements translate into an individual "sense of entitlement". Authors argue that by challenging gender norms at a cognitive level (of agency), policy framing facilitates agency and challenges gender norms, making the use of policy "possible." Chatrakul et al. (2017) further demonstrate how this "sense of entitlement" is shaped by social and economic circumstances, functioning as a cognitive filter to influence one's agency and capabilities in reconciling work and family life. To exemplify, when a policy instrument is removed from cultural norms, familial values, or organizational practice and culture (Grönlund & Javornik, 2014), individuals may not value or use it.

Earlier studies highlight the multiple interpretations and applications of the capability approach in family policy research, in particular with varying emphasis on policies, workplaces, the "sense of entitlement," or conversion factors, as well as variation in interpreting policy as an institutional conversion factor or a means. Some of this divergence can be explained by two varying approaches to the capability approach: Sen's and Nussbaum's. Sen's perspective is essentially economic and philosophical, emphasizing questions around the pursuit of the "good life" (that which people have reason to value) and measurements of life quality. Nussbaum's perspective, in contrast, is moral-legal-political philosophical, arguing that governments should guarantee a given set of "basic" human capabilities (Nussbaum, 2000).

Our application of the capability approach to family policy follows Sen's perspective. While we recognize the value of both approaches, we use Sen's for its emphasis on the role of agency in creating inequalities in capabilities. Furthermore, we demonstrated above how domain-specific knowledge in the family policy field is required to inform the application of the capability approach in comparative family policy research. Similar to the application of the capability approach to countries' national childcare policies in our earlier study (Yerkes & Javornik, 2019), such domain-specific knowledge provides building blocks for further specification of capability "theories" within the family policy field.

## **Operationalizing and Evaluating Childcare Capabilities**

To apply the capability approach as an evaluative empirical framework, we focus on childcare policy as a *means*, and how aspects of childcare policy design interact with the *conversion factors* of gender and class to create varying childcare capabilities. A similar approach can be taken in future family policy studies, using domain-specific knowledge to identify the relevant aspects of policy design that potentially interact with conversion factors to create variation in capabilities. For childcare, based on our earlier research and established gendered welfare state scholarship, we argue that direct public service provision offers parents across socio-economic groups the best opportunities to arrange childcare in ways they have reason to value because it provides *real* opportunities (Leitner, 2003; Lister, 1997; Saraceno & Keck, 2010; see also Chapter 8 by Vandenbroeck in this volume). Against this background, we focus on national childcare services across five most salient features of its potential as a means: availability, accessibility, affordability, quality, and flexibility (Bonoli & Reber, 2010; Ciccia & Bleijenbergh, 2014; Daguerre, 2006; Eydal & Rostgaard, 2011; Gislason & Eydal, 2011; Gornick & Meyers, 2003; Gornick, Meyers, & Ross, 1998; Javornik, 2010, Javornik, 2014; Plantenga & Remery, 2005, 2009; Saraceno, 2011). Our underlying assumption is that availability, accessibility, and affordability are key to childcare capabilities, and that quality and flexibility become an issue once childcare is available and accessible, and are often a reflection of affordability (Kreyenfeld & Hank, 2000; Morgan, 2005; Verhoef, Tammelin, May, Rönkä, & Roeters, 2016).

Ideally, we would use data that best allow us to illustrate the capability approach applicability to comparative childcare policy analysis. However, available and commensurable data have several shortcomings, such as a focus on the public sector, and hence an absence of private sector data, spending profiles which cover multiple services and programs, and variation in what actually gets measured (Fagan & Hebson, 2005; Javornik, 2014; Keck, Hessel, & Saraceno, 2009; Lambert, 2008; Mätzke, Brokking, Anntonen, & Javornik, 2017; OECD, 2018; Plantenga & Remery, 2005). Considering these issues, we use multiple data sources (Eurostat, 2017; Eurydice, 2018; Multilinks, 2011; OECD, 2017a, 2017b, 2018; Plantenga & Remery, 2005, 2009, 2013, 2015; SPIN, 2019) and single country reports to provide a comprehensive analysis; arguably, better data will, in the future, enable improved analyses, and our work seeks to contribute to these efforts (see also Chapter 24 by Sirén, Doctrinal, Van Lancker and Nieuwenhuis in this volume).

# **Availability**

Childcare availability is integral to parents' employment, particularly mothers', but it also plays a key role in child development (OECD, 2017b; Plantenga & Remery, 2009; Saraceno, 2011; van Huizen & Plantenga, 2015). Analyzing availability is complex because of different types of care provision available in each country. Moreover, countries combine formal and semi-formal provisions, offer subsidized arrangements outside nurseries, registered playgroups, and a mix thereof. In earlier research, availability is largely operationalized using enrolment rates. We find this problematic because it conflates structural differences in care provisions by focusing on a single outcome (Javornik, 2014). A more useful approach distinguishes logics of care provision (Brennan, Cass, Himmelweit, & Szebehely, 2012), including logic of non-familial care through market provision (for profit), state provision, and associations, whereby childcare is offered by formal private or non-profit organizations. We argue that public-and market-provision mechanisms have different consequences for parents' capabilities, and address this by combining data on the problematic yet most commonly used measure for availability (enrolment rates for 0–2-year olds and 3–6-year olds) with a classification of countries based on their prevailing provision mechanism.

## **Accessibility**

Comparative studies on childcare typically rely on capacity (the number of places available) as a proxy for service accessibility (Plantenga & Remery, 2015). However, to identify the potential for social inclusion/exclusion from public service, we need to understand admission criteria, i.e., the construction of a right to childcare (Jensen, 2009). Namely, selective practices (using preferential criteria) and provider autonomy can create tensions and diminish childcare capabilities (Javornik, 2014). In contrast, national-level capacity planning could reduce disparities between geographical units (OECD, 2018; Plantenga & Remery, 2005, p. 35). Furthermore, childcare is accessible when it is offered as a right attached to children rather than conditional (Saraceno, 2011). Childcare subsidies tied to eligibility criteria (e.g., parents' employment) create an opportunity gap for parents in education or training, or seeking a job, starting a business or volunteering. Having childcare in place is essential before parents can undertake such activities (Javornik & Ingold, 2015). When childcare placement is provided as a child's right, policy "stresses societal responsibility to grant all children adequate family care and time, as well as non-family resources for the full development of their capabilities," which improves parents' childcare capabilities, particularly across class (Saraceno, 2011, p. 92). Furthermore, and drawing on Gornick and Meyers (2003) and Javornik (2014), policy that guarantees a place to all children (child's right) without delays and gaps between paid leave and childcare improves childcare capabilities. When a child can be refused a place based on preferential criteria (e.g., family composition, income, parental employment status), service provision is likely to vary (Fagan & Hebson, 2005), hindering access to public service. Lastly, policy that enhances childcare capabilities opens care service before the end of paid leave or at least coordinates the two (Javornik, 2014; Javornik & Kurowska, 2017). Thus, a measure of accessibility should focus on admission requirements, including the allocation of places and admission age, to be evaluated in relation to parental leave and its intertwining with childcare service accessibility.

#### **Affordability**

From a policy perspective, providing affordable childcare is seen as key to promoting mothers' employment. Understandably so, as childcare costs are the equivalent of a regressive tax on mothers' labor supply and reduce financial returns from employment (Esping-Andersen, 2009). In practice, free childcare services practically do not exist. As a result, funding mechanisms are crucial for childcare capabilities. Earlier studies predominantly measure affordability by public spending. Such measures are problematic as they reflect national funding streams to finance services (OECD, 2016a). These measures generally do not capture other funding sources, e.g., when local governments do not report on childcare spending (e.g., Gornick & Meyers, 2003; Javornik, 2014; Lambert, 2008) or when services are funded through collective bargaining agreements (Yerkes & Tijdens, 2010). Measuring affordability using the cost of childcare as a percentage of net family income (e.g., Keck et al., 2009; OECD, 2017b) is similarly problematic. Tax allowances presume parents have a taxable income (e.g., Immervoll & Barber, 2006; Jaumotte, 2003); such figures therefore do not consider parents not in employment.

A salient aspect of affordability often not taken into consideration is the funding mechanism, i.e., the use of "supply-led" direct funding streams to providers, and/or the use of a "demand-priming approach," whereby parents receive financial help directly. For the latter, policy design regulates funding through means-testing or limits to parental/childcare fees using a sliding-fee scale (Immervoll & Barber, 2006; Javornik, 2014). A "supply-led system" opens space for parents across socio-economic groups to access childcare via direct funding (Javornik, 2014; Leitner, 2003). In contrast, the demandpriming approach creates gaps in childcare capabilities as operating rules are set by providers to maximize profitability (Brennan et al., 2012). Parental fees increase, and the level of household income and/or childcare subsidy, tax allowance, and employers' assistance become crucial (Blackburn, 2012; Mamolo, Coppola, & Di Cesare, 2011; Morgan, 2005). Another issue related to the demand-priming approach is that child-related tax deductions in some countries are not available at the time when parents incur childcare expenses, but in the following fiscal year (Immervoll & Barber, 2006). Affordability is therefore particularly relevant for low-income parents (Capizzano & Adams, 2004) and single mothers (Kreyenfeld & Hank, 2000), whose purchasing power is comparatively lower to other groups. We address these issues by combining established data on childcare costs with an analysis of funding rules.

## **Quality**

Quality is a measure of service experience but can be seen as subsidiary to service availability, accessibility, and affordability; unless childcare is accessible and affordable, quality is arguably less important. When there are issues with service quality, parents may not easily switch between childcare providers or an exit could be too costly (OECD, 2018). We know little about these interrelationships but high-quality service generates incentives for using nonfamilial childcare (Plantenga & Remery, 2005). Measuring childcare quality is problematic because a standard definition is missing and there can be considerable differences between measurable qualities and how parents assess and experience quality (Janta, van Belle, & Stewart, 2016; Keck et al., 2009; OECD, 2018).

Family policy research typically distinguishes between *process* quality at the organizational/staff level (e.g., what happens in the setting: the play and learning environment, child–teacher and child–child interaction) (Janta et al., 2016; OECD, 2018) and *structural* quality (e.g., maximum group size, childto-staff ratios, and the educational level of childcare staff ) (Penn, 2013). Comparative data on process quality remains limited (Janta et al., 2016; OECD, 2018). Recent reports and a meta-analysis from the OECD (2017b, 2018) suggest few countries structurally gather data on process quality. The limited evidence available suggests that measures of process quality and its effects on children's development are inconsistent. Given these limitations, we focus on structural quality measurements of maximum group size, childto-staff ratios, and the educational level of childcare staff. Where possible, information on process quality is provided (OECD, 2018).

#### **Flexibility**

Flexibility in childcare service provision seems essential for childcare capability but suffers from limited understanding of what it entails, in addition to a lack of standardized data allowing for cross-country comparison. Flexibility is generally defined in relation to opening hours of childcare services. These determine the extent to which parents can use services (Gornick & Meyers, 2003, p. 227) and deal with the constraints of time and distance (Emlen, 2010). Childcare services are not commonly available during nonstandard hours, which makes parents reliant on informal or commercial care. Lack of flexible childcare affects parents working non-standard or variable work hours, in education, looking for jobs (Grönlund & Javornik, 2014; Javornik & Ingold, 2015; Verhoef et al., 2016). However, flexibility in childcare services presumes usefulness for parents, rather than children. Whereas developmental studies suggest children do well in structured, consistent care environments, inflexible childcare services can make it difficult for parents to cope with the parameters of employment, schooling, or other activities (Javornik, 2014). The focus here is on parents' capabilities; to enable their childcare capabilities, services should be available on a full-time basis for at least 30 hours per week (Mills et al., 2014) all year-round, while providing parents sufficient freedom to choose the hours (OECD, 2007; Plantenga & Remery, 2005, pp. 38–42). Ideally, we would develop an indicator using parents' self reports to capture this. However, such data are unavailable across countries, and thus we analyze guidelines and statutory frameworks related to the annual timetabling and opening hours, using Eurydice (2018) and single country data.

For all five policy aspects, we focus on publicly funded (non-familial) childcare for children from birth to compulsory school age in center-based day care. We evaluate the situation from a dual-earner couples' perspective while acknowledging that families are much more diverse.

# **Empirical Applications to Childcare Services**

Analysing childcare policies along the key aspects of childcare services from a capability perspective demonstrates how diversified national childcare policies are and how their policy designs vary more within our de-familialized country cluster than suggested by "varieties of familialism" scholarship (see Yerkes & Javornik, 2019). While some patterns are discernible within the public provision and marketized groups of countries, considerable differences exist within and between subgroups, which can have significant consequences for childcare capability. Building on an earlier empirical analysis of six countries (Yerkes & Javornik, 2019), we go more in-depth in our analysis here to highlight the potential consequences of variation in childcare service availability, accessibility, affordability, quality, and flexibility in two systems characteristic of a public provision (Sweden) and marketized provision (UK) (for more countries see Yerkes and Javornik, 2019). The data are summarized in Table 7.1.

In Sweden, childcare is primarily offered as a public service. The provision is decentralized as the legal responsibility of local municipalities, but is nationally regulated, coordinated, and monitored (Gislason & Eydal, 2011; Javornik, 2014). Admission criteria are nationally regulated as well, with central capacity planning, which ensures uniformity and coherence across settings and municipalities. Children are largely enrolled in full-time childcare (30+ hours a week). While few children under 1 year attend childcare given the Swedish system of generous parental leave, nearly all children aged 3 to compulsory school age attend formal care (96.2%). Unlike Sweden, part-time childcare is the norm in the UK, and market provision is the dominant form of childcare service delivery (Penn, 2013). Responsibility for childcare services is less clear-cut than in Sweden, and is divided between central and local governments dependent upon the age group concerned (Gislason & Eydal, 2011, p. 73). There is an array of actors operating across private, voluntary, and non-for-profit organizations and local council services in the UK (Lloyd, 2015), resulting in significant regional variation in childcare provision, with London and the South East offering the most expensive under-5 childcare (Javornik & Ingold, 2015). The expectation is that the market mechanism creates incentives for providers to offer more choice and competitive pricing, leading to a better balance between supply and demand (Brennan et al., 2012). However, 30% of parents report insufficient childcare in their area (DfE, 2014). By and large, the dominance of market mechanisms, especially for children under 3, leads to inequality in capabilities. Although not explicitly cast in gendered terms, variant availability negatively




*Data sources* CESifo/DICE (2010); Eurostat (2017); OECD (2014, 2016b); Plantenga and Remery (2015). \*The structure of parental leave in Sweden means data include children 12 months–2 years. Data are from 2016. \*\*Data refer to 2012 full-time childcare costs for a dual-earner couple with full-time earnings (at 100% and 67% of average earnings, respectively). \*\*\*Data on quality are from 2015

affects women's capabilities who need to reduce working hours or withdraw from the labor market in order to provide childcare. Namely, in this one-anda-half earner model, women take on higher parenting and care responsibilities than men (Lewis, 2009). From a class perspective, policy incoherence across settings reduces childcare capabilities because private businesses are focused on profitability (OECD, 2018); this can diminish childcare capabilities of low-income parents in particular. In Sweden, in contrast, public provision that is centrally regulated and overseen improves childcare capability along gender and class lines.

Sweden similarly offers better accessibility than the marketized UK, providing childcare as a child's right and with a guarantee (without reasonable delay), although this is exceptional, even among public-mechanism countries (Yerkes & Javornik, 2019). This is not the case in the UK. Sweden offers childcare services to children aged one to compulsory school age on a fulltime basis. In the UK, childcare (early education) is essentially available for children aged 0–14, and up to age 18 for disabled children (OECD, 2015). However, children under two are not entitled to public childcare and children aged 3–4 have the right to 30 hours for 38 weeks/year since 2018 (Javornik & Ingold, 2015).

In both countries, however, inequalities in childcare capabilities can arise from limitations in childcare accessibility. In Sweden, paid parental leave and childcare services are contiguous in principle. But in practice, most places become available in September, when older children go to school (Grönlund & Javornik, 2014; Javornik, 2014; Javornik & Kurowska, 2017). This gap can disrupt childcare capabilities, an issue pertinent for low-income parents. Similarly, the absence of well-paid parental leave in the UK can lead to reduced childcare capabilities across gender and class. The UK's maternity leave offers the lowest replacement rates in the OECD, of around one-third of gross average earnings (OECD, 2017a) and Shared Parental Pay (the statutory payment provided to eligible parents taking up Shared Parental Leave during the first 37 weeks of leave) adopts the same basic rate. While employers can top-up government payments by offering extra-statutory (enhanced) benefits (e.g., sickness, maternity, paternity, and shared parental pay) up to one year, these are at the discretion of employers and seldom offered to fathers (Javornik & Oliver, 2019). Thus, fathers in particular are missing out on a potential conversion factor to draw on parental leave to achieve their preferred childcare arrangement. This widens the gap between parental leave and public childcare, further diminishing capabilities of mothers in particular. This aggrevates class divisions considering a recourse to paid parental leave is not a real choice when alternative childcare is unavailable (Kurowska & Javornik, 2019).

A significant difference exists between public-mechanism and marketized countries in relation to affordability (OECD, 2017a). Childcare is affordable in Sweden, where dual-earner couples spend a marginal share (4.4%) of net family income on childcare services. Sweden uses an income-based sliding-fee system; maximum fees are set by the state, with a lower payment ceiling and discounts for certain groups (e.g., low-income parents, single parents, and large families). Municipalities can introduce discounts and charge differently within the national guidelines. Such regional variation could diminish parents' childcare capability across municipalities, particularly for middle-income families: while high-income families benefit from the ceiling, low-income families can have fees waived completely. Unlike Sweden, the UK combines part-time, universal free places with demand-led funding through the tax and benefit systems for pre-school children. This results in parents paying an exorbitant 33.8% of family net income per month for outof-pocket childcare expenses, versus the 12.6% paid on average by parents in OECD countries. In the UK open market, fees are set by providers to maximize profitability. They can receive financial help directly; other subsidies go directly to childcare providers (Javornik & Ingold, 2015). Retrospective reimbursement through the tax and benefits system is inefficient and a deterrent for many families, resulting in diminished childcare capability. In reality, prohibitive childcare costs further hinder capabilities across social class and gender, when low-income families are forced to opt for family care, usually taken up by mothers.

In relation to structural quality, there is a clear division between Sweden and the UK. Sweden has smaller playgroups led by trained staff. All staff are required to have at least secondary or tertiary education, with the focus on a pedagogical service centered on children's development. While the UK has larger group sizes (30–35 children per group), staff-child ratios are lower (Table 7.1). Staff are not required to have secondary or tertiary education, and there are no mandatory requirements for in-service training (Penn, 2013). In the profit-driven childcare market, providers are motivated to minimize costs. With regulated child-staff ratios, cost minimalization can be achieved by employing a less qualified workforce (Brennan et al., 2012; Penn, 2013). The focus on profitability and childcare as an employment instrument leads to lower qualification standards, resulting in serious quality issues. When childcare is of lower quality, childcare capabilities of parents are reduced across gender and class. Attitudes toward the use of formal childcare—influenced by perceptions of quality—are more positive in Sweden than in the UK. While the use of formal childcare is becoming increasingly accepted in the UK, lower socio-economic groups are less positive toward public childcare, leading to gendered patterns of care, and thus inequality in capabilities (Fagan & Norman, 2012). Additionally, low-income parents may have fewer skills needed to "navigate the system" in assessing differences in quality and ultimately lack the financial resources to purchase higher quality childcare (Brennan et al., 2012). Both countries' regulatory systems outline process quality regulations and standards; this generally includes health and safety requirements, space, staff training, staff-child ratios, and curricula. In sum, unlike Sweden, the UK faces significant childcare quality problems, with an uncomfortable relationship between market provision and quality (Brennan et al., 2012; Lloyd, 2015). Childcare providers perform below the OECD average on quality, particularly in relation to staff skills and parental involvement (Taguma, Litjens, & Makowiecki, 2012). In Sweden, childcare quality is higher, with communal obligations to delivering reliable, high-quality childcare, with equity more pronounced than choice (Penn, 2013).

Lastly, our analysis suggests that limited service flexibility affects childcare capabilities across both countries, with notable differences. In Sweden, childcare is generally typified and standardized, i.e., largely compatible with national standard work hours (Grönlund & Javornik, 2014). Statutory frameworks set no requirements about opening hours, which are set by municipalities. Childcare centees run on a full-time basis, for 11 hours a day on weekdays throughout the year. Providers can offer weekend or overnight services at the discretion of local councils. In the UK, opening hours are not covered by statutory frameworks but are left to providers. Statutory guidance does support flexibility in service delivery and local authorities are expected to offer flexible packages of free hours but of not more than 10 hours (no minimum session length). The absence of childcare outside standard hours (there is no service before 6 am or after 8 pm) limits parents' capabilities, especially in non-urban areas, where out-of-hours care is most limited (Grönlund & Javornik, 2014; Javornik & Ingold, 2015). In the UK, this leads to high demand for flexible working arrangements, in particular among women and also negatively affects parents in education or training.

## **Discussion and Conclusion**

Contemporary comparative family policy research continues to face difficulties in conceptualizing and operationalizing family policies for cross-national research given their variations and complexities. This chapter offered an alternative analytical framework by conceptualizing family policy using the capability approach. We focused here on childcare services, in particular service accessibility, availability, affordability, quality, and flexibility. We suggest that family policy designs shape parents' capabilities in distinct ways. In relation to childcare, public service provision (e.g., Sweden) generally improves parents' childcare capabilities. In marketized countries (e.g., the UK), parents' capabilities are significantly diminished along gender and class lines. Additionally, our analysis highlights tensions, such as potential regional variation in childcare affordability in Sweden. We note that limited flexibility is an overarching problem across both systems, which is particularly problematic for parents in non-standard jobs (Verhoef, 2017), in training or education, or seeking jobs. A further advantage in applying the capability approach to comparative family policy research is the ability to conceptualize potential impact of policy across key individual-level factors. As our analysis demonstrates, distinct public and market approaches to childcare service provision have different gender and class consequences, which the capability approach is particularly apt to capture. Similarly, future studies could emphasize other *conversion factors*, such as ethnicity, age or individual health and well-being. A further potential for future comparative family policy analysis lies in the capability approach's ability to highlight the salient aspects of policy design for comparative purposes (as shown here), or to focus on conceptualizing the linkages between such policy design features.

In sum, the capability approach provides a valuable analytical instrument for comparative family policy analysis. Its analytical power is in moving beyond the more established approaches such as defamilialism, which mask key distinctions between public and market service provision. It facilitates a more nuanced conceptualization of family policy, thus revealing parents' real opportunities in relation to policy design. Furthermore, it recognizes multiple lifestyles and values, thus moving beyond implicit commodification assumptions; this opens up analytical space to include parents' appreciation of multiple care arrangement opportunities. Conceptually, the capability approach provides a promising way forward in comparative family policy research. Empirically, however, challenges remain. Optimal commensurable measures are not equally available across countries, which limits more nuanced cross-national comparative family policy analysis. As countries continue to invest more in childcare (Pavolini & Van Lancker, 2018), it is necessary to also invest in measuring the effects of childcare, both in terms of outcomes as well as parents' capabilities. Similar investments in other areas of family policy are needed as well.

**Acknowledgements** The research leading to these results has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 730998, InGRID-2—Integrating Research Infrastructure for European expertise on Inclusive Growth from data to policy.

## **References**


DfE. (2014). *Childcare and early years survey of parents* 2012–13. London: DfE.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **8**

# **Early Childhood Care and Education Policies that Make a Difference**

## **Michel Vandenbroeck**

The history of formally organized early childhood education and care (further: ECEC) is as long as the history of compulsory schooling. In several countries early childhood care and education from birth to compulsory school is integrated in one unified system (e.g. in Denmark, Sweden or Jamaica). However, a majority of countries have a historically constructed split system in which the "care" for the youngest is separated from the "education" of the older children (Eurydice, 2019; Kaga, Bennett, & Moss, 2010; Moss, 2013; Urban, Vandenbroeck, Lazzari, Peeters, & Van Laere, 2011). Provision for the 3 to 6-year olds (further: preschool) seems to be considered—historically and almost universally—as focused on education and is increasingly viewed as a preparation for later education or as the first step in life-long learning (e.g., European Commission, 2011; Naudeau, Kataoka, Valerio, Neuman, & Elder, 2011).

In contrast, provision for the youngest (further: childcare) has been the subject of very diverse policies in the course of its history. In several countries, formally organized childcare for babies and toddlers emerged in the nineteenth century as a means to combat child mortality as well as an instrument to allow cheap female labor in the period of the first industrial revolution (Vandenbroeck, 2006). In that period, children enrolled in childcare were

© The Author(s) 2020

M. Vandenbroeck (B)

Department of Social Work and Social Pedagogy, Ghent University, Ghent, Belgium e-mail: michel.vandenbroeck@ugent.be

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_8

exclusively from poor working-class families, and their parents were considered to be in need of being civilized, among others by childcare workers and preventive health care services (De Wilde, Vanobbergen, & Vandenbroeck, 2018). Throughout history and into the present, childcare has been and continues to be a subject of labor policies. Under the influence of feminism in the 1960s and 1970s, the focus of these labor policies shifted from the needs of poor working-class mothers to ensuring equal opportunity policies on the labor market for men and women alike. As a result, the socio-economic profile of childcare users changed singificantly since the 1970's. In those countries where childcare was established as a working-class facility, more and more middle-class families started to enroll their babies and toddlers (e.g., in Belgium, France, the Scandinavian countries) and governments consequently invested in increasing the number of available places. Sweden was the first country where childcare was considered a universal right for families. These investments came to a halt during the economic downturn of the 1980s, especially in continental Europe.

Typically, countries that did not invest substantially in the sixties and seventies (e.g., the United Kingdom or The Netherlands), realized only later that a childcare policy was a necessary condition for a labor policy. Therefore, these countries attempted to increase the number of places only later, after the economic crisis, with limited public means through privatization and a market-oriented system, assuming that the market would create places where the needs are higher and that competition for the parents as customers would also increase the quality of childcare. A typical example of this model is the Netherlands, where since the 2005 childcare reform private (for profit) companies took over small private childcare facilities in what was presented as a promising market, and six years later almost half of the childcare places in The Netherlands were owned by only 5% of the providers. It was believed that the market would solve the Dutch problems of accessibility and quality. However, we now know that the expectations of such policies were not met: access to childcare became more unequal (Noailly, Visser, & Grout, 2007), quality deceased and the levels of quality only recovered after substantial interventions by the government (NCKO, 2005, 2013).

By the end of the twentieth century, childcare policies were increasingly influenced by iconic longitudinal studies, such as the Abecedarian project, showing the beneficial effects of childcare participation on later development (e.g., Barnett & Masse, 2007). Under the influence of Third Way and neoliberal policies, ECEC increasingly became politically framed as a means to realize equal opportunities and to combat the intergenerational transmission of poverty by national (e.g. Field, 2010) and international policymakers (e.g., Paes de Barros, Ferreira, Molinas Vega, & Saavedra Chanduvi, 2009). In that vein, the educational function of not only preschool but also childcare became increasingly recognized next to its economic labor market-oriented function. As a result, it is now not only the quantity of childcare (numbers of available places) but also its quality that is a matter of political concern (e.g., European Commission, 2018). In sum, childcare for the youngest children originated as an instrument to combat child mortality, exclusively reserved for the poorest of working-class mothers. It evolved into an instrument to reconcile employment and parental responsibilities for all parents and consequently, the users became more middle-class families. While childcare retains its economic function, it gradually was also considered an educational environment, just as preschool already was, and subject to policies that aim to combat the intergenerational transmission of poverty. In many countries this paradoxically means that governments stress the social and pedagogical functions of childcare for poor families, while at the same time these families hardly have access to places of above-average quality (Vandenbroeck, 2019).

In this chapter, this paradox will be further developed. In the next section, we elaborate more on the evidence for the educational and social functions of preschool and childcare. Subsequently, it will be discussed how quality matters for the pedagogical function of ECEC, and how accessibility matters for the social function of ECEC. This chapter ends with a discussion what types of policies successfully combine social and pedagogical concerns, ensuring good quality and accessibility for all.

### **How ECEC Can Make a Difference**

After the Second World War, three major longitudinal projects started in the United States (US), one of which addressed childcare (the Abecedarian project in 1972) and two focused on preschool (HighScope Perry Preschool in 1962 and the Chicago Preschool project in 1985). The projects were part of a broader policy to address what was believed to be the "socio-cultural handicaps" of the black population in cities as Chicago and was in official documents labelled as "the negro problem" (Beatty, 2012). Together with other studies, these projects showed that children benefitting from highquality ECEC were better off later in their lives than children who did not attend (Barnett, 2011; Burger, 2010; Camii, Vargas, Ryan, & Barnett, 2010). The studies have been used to argue for considering ECEC as a means to close the achievement gap. However, as Morabito, Vandenbroeck, and Roose (2013) explain, as these first studies compared poor children with equally poor children, their results do not really learn anything about the equalizing potential of ECEC. Yet, since these early US-based studies, European studies have confirmed the positive individual benefits of attending high-quality ECEC in the domains of cognitive development, socio-emotional development, and school career for all children and for disadvantaged children in particular (Lazzari & Vandenbroeck, 2012). We give a few examples, without aiming to be exhaustive. For a more complete overview of European studies, see Vandenbroeck, Lenaerts, and Beblavy (2018). In Norway, the 1975 childcare reform was used to study the impact of childcare availability on children's educational attainment by Havnes and Mogstad (2011). The study concluded that childcare attendance was significantly associated with educational attainment in the long-term (30 odd years later). Subsample analyses indicated the largest effects on education for children with low educated mothers. In Northern Ireland, the team of the EPPNI study (Melhuish et al., 2006) concluded that preschool enhances cognitive development, social development, and behavioral competences, and that higher quality preschool is associated with better intellectual outcomes. Felfe and Lalive (2011) conducted a longitudinal study on a national data set of (former) East and West Germany, following 800 children from age two to ten. Similar to the Northern Irish study, Felfe and Lalive (2011) found that children who attended ECEC centers achieved significantly higher scores on all cognitive and non-cognitive indicators. Children from lower SES benefitted more when accessibility to ECEC centers was improved. The results are consistent with German research by Spieß, Büchel, and Wagner (2003) as well as Becker and Tremel (2006). In Italy, Brilli, Del Boca, and Pronzato (2011) confirmed that an increased accessibility for vulnerable children (i.e., by childcare rationing) had positive effects and that the impact is highest for children with mothers who had the lowest formal educational levels. In Switzerland, Lanfranchi, Gruber, and Gay (2003) looked at the impact of ECEC in migrant children with Albanian, Turkish, Portuguese, and Italian backgrounds, as well as Swiss backgrounds and suggest that children who attended ECEC were significantly better assessed by their kindergarten and primary school teachers in first class in terms of their linguistic, cognitive, and special skills than children who grew up exclusively in the circle of their own family. One of the largest studies in Europe is the Effective Pre-School, Primary and Secondary Education (EPPSE) project in the United Kingdom (UK). The study started in 1997 and followed more than 3000 children since then (Sylva, Melhuish, Sammons, Siraj-Blatchford, & Taggart, 2014). From these and other studies, there is now robust evidence that all children benefit from ECEC, as ECEC has the potential to increase language development, pre-academic skills, cognitive outcomes, as well as social and emotional development. Evidence on the equalizing potential of ECEC is less conclusive, but it is clear that the developmental benefits are most salient for children from more disadvantaged families. The EPPSE study sheds more light on this. It shows that beneficial effects of ECE are present for all children, and as such the gap between higher and lower social statuses is not diminished (Sylva, Melhuish, Sammons, Siraj-Blatchford, & Taggart, 2004). However, it also indicates that children from lower SES who attended high-quality preschool score above what primary school teachers estimate to be the minimum level to be successful in school, while children from lower SES who did not attend ECEC did not reach the same level. In contrast, all children with higher SES attain the minimal level, regardless of their attendance of preschool. In sum, while high-quality preschool does not necessarily close the achievement gap, it does make a salient difference, particularly for children from lower social statuses.

Research, however, also suggests that not all ECEC matters to the same extent. A famous example is the study of Caille (2001) in France, who showed that children attending *école maternelle* (preschool) at the age of two did not fare better in primary school than those enrolled at the age of three, when controlled for socio-economic status. The study of Driessen (2004) as well as the meta-study of Fukkink, Jilink, and Oostdam (2015) also suggests that the beneficial effects of preschool for 2 to 4-year old underprivileged children of the Dutch *peuterspeelzalen* were disappointing. These counter-intuitive results show that, while ECEC matters, not all ECEC matters in the same ways and that quality plays an important role.

The impact of ECEC is not limited to its direct impact on children's development. There is abundant literature showing that living in poverty is harmful for children's development (for an example from neuro-research, see Neville, Stevens, Pakulak, & Bell, 2013; for a broader overview see for instance Zaouche-Gaudron, 2017). Poor families tend to live in neighborhoods with poorer provision, more frequently employing staff that shows inadequate behavior toward children (Webster-Stratton, Reid, & Hammond, 2001). Living in adverse circumstances also has a negative impact on children's health and learning opportunities. In addition, job insecurity has a negative impact on the well-being, stress, and marital conflicts (Brotman et al., 2013) and may therefore negatively influence parental skills. It is well documented that social support is one of the more universal and salient forms of parenting support (Jack, 2000; Sarason, Sarason, & Pierce, 1990). ECEC can be an important source of social support for all parents and for parents in more vulnerable situations in particular, when it can serve as a meeting place where parenting experiences are shared. It is increasingly a focus of international policy makers that ECEC contributes to social support as well as to social cohesion. While empirical studies indicate that this may be the case, they also show that ECEC does not necessarily fulfil these roles (Geens, Roets, & Vandenbroeck, 2015; Geens & Vandenbroeck, 2013) and that professional support of the workforce is necessary to develop a vision on their role in relation to peer support of parents.

A special note needs to be made about the growing number of immigrant families with young children in European countries. ECEC has the potential to address essential needs of these vulnerable young children at a critical stage in their development as it represents the first and most intimate point of interaction between receiving societies and immigrant families with young children. They can (and often do) also play a central role in supporting their families' long-term integration outcomes (Vandenbroeck, Roets, & Snoeck, 2009). However, research examining successful strategies in serving young immigrant children and their families in early childhood remains scant (Park, Katsiaficas, & McHugh, 2017; Vesely & Ginsberg, 2011). In order to enable ECEC provision to fulfil these manifold expectations, ECEC policies need to properly address the issues of accessibility and quality.

# **Quality Matters**

All studies that not only looked into the general impact of ECEC on children's outcomes, but also took account of the issue of quality, demonstrated that the impact of ECEC on children, parents, and communities is moderated by quality (e.g., Loeb, Fuller, Kagan, & Carrol, 2004; Melhuish et al., 2006, 2015; NICHD, 2002). The core of the quality of ECEC resides in the educational and emotional support that childcare and teaching staff provides for the children as well as on their relations with parents. In other words, quality is a function of interpersonal interactions, which is called *process quality* (La Paro, Wiliamson, & Hatfield, 2014). Process quality, in turn, is conditioned by *structural quality*. While there is no one-on-one relationship between single structural quality dimensions and process quality (see Slot, 2018, for a systematic literature review), some structural quality conditions are necessary (but not always sufficient) conditions to realize process quality. One of these conditions is related to the workforce: staff competences and qualifications, and working conditions (Urban et al., 2011; Vandenbroeck, Laevers, et al., 2016). Cross-sectional as well as longitudinal studies indicate a relation between staff qualifications and competences and childcare quality as well as with children's developmental outcomes (Clarke-Stewart, Vandell, Burchinal, O'Brien, & McCartney, 2002; Early et al., 2007; Fukkink & Lont, 2007; Sylva et al., 2004). Also, a literature review conducted by the Organization for Economic Co-operation and Development (OECD) concluded that "staff who have more formal education and more specialized early childhood training provide more stimulating, warm and supportive interactions with children" (2006, p. 158). Several EU Member States have qualifications at a bachelor level (ISCED 5), such as the Danish *pedagog*, the Swedish teacher, the French *Éducateur de Jeunes Enfants,* or the Belgian *Pedagogie van het Jonge Kind* . However, the OECD (2006) also found that in many European Member States the actual qualification levels are much lower.<sup>1</sup> Better qualified staff is often not recruited by ECEC providers because of lack of funding, while in most countries a large proportion of the workforce also consists of unqualified assistants (Van Laere, Peeters, & Vandenbroeck, 2012). This relates to a second crucial structural quality criterion: working conditions. As Early et al. (2007) found, staff qualifications only make a difference if working conditions are adequate. If not, turnover rates are usually high which jeopardizes process quality. Adequate working conditions include decent salaries, paid time off from the children to document, reflect and plan, and the opportunity for continuous professional development. A systematic literature review (Peleman et al., 2018) as well as practical experiences in several countries (Vandenbroeck, Urban, & Peeters, 2016) demonstrated that continuous professional development compensates for a lack of pre-service qualification, provided it is long-term and sustained. In reality, however, these conditions are seldom met (Eurydice, 2019). Particularly in split systems, qualifications and working conditions of the workforce caring for the youngest children are less than adequate (see also European Commission, 2018). Next to the workforce, there are of course other structural conditions for optimal process quality, including adult child ratio's, space, and safety requirements.

In addition to workforce characteristics and working conditions, the pedagogical curriculum is the second most important aspect of process quality. There are many differences in how ECEC is conceived of. Some countries consider ECEC merely as a preparation for compulsory school and focus on adult-centered activities aiming at reaching specific developmental outcomes. Others adopt a more holistic social pedagogical approach, favoring play based and child-centered experiences (Bennett, 2005; Samuelsson, Sheridan, & Williams, 2006). In addition, curricula differ in how they conceive of

<sup>1</sup>See also http://seepro.eu, a website with a comprehensive overview of the different degrees and qualifications in ECEC being organized in EU Member States.

the relations between parents and staff (Janssen & Vandenbroeck, 2018), as well as in how to deal with the increasing diversity of children and families (Janssen, 2020). In-depth observational studies (Slot, 2018; Slot, Lerkkanen, & Leseman, 2016; Vandenbroeck, Laevers, et al., 2016) show that process quality is best served by an *educare* approach, integrating care and education, where adult-centered and child-centered initiatives are balanced and where there is room for children's perspectives, while adults scaffold and extend the children's experiences. This obviously also requires higher levels of staff competences.

Sadly, those observations time and again show that while the emotional support of babies and toddlers is moderate to good, the educational support is often significantly lower and may well be below the quality level that is necessary to expect long-term positive outcomes. Two US-based studies (La Paro et al. 2014; Thomason & La Paro, 2009), for instance, found that language support was on average inadequate. In the Netherlands, Slot and colleagues (2016) found similar results. A Belgian study of 400 childcare facilities (Hulpia et al., 2016) showed that educational support varies from moderate to low, with quality of feedback given to toddlers even being completely insufficient. Portuguese (Cadima et al., 2016) and US-based (e.g., Jamison, Cabell, LoCasale-Crouch, Hamre, & Pianta, 2014) observational studies show that also in the youngest baby-groups the quality of educational support is significantly lower than the quality of the emotional dimensions. What the levers are for the improvement of educational quality may differ from one country to another. Yet they most probably include adult-child ratios, qualifications, and professional support for staff, curriculum development, and monitoring.

It needs to be noted here that demand-side funding (i.e., funding parents who "choose" a place for their child on the ECEC market) has predominantly led to lower quality than supply side finding (funding provision), as marketization usually entails budgetary cuts on staff and limits the space of reflection and experimentation (Cleveland, Forer, Hyatt, Japel, & Krashinsky, 2007; Moss, 2009). In the Dutch case, the regular quality monitoring by NCKO (2005, 2013) showed that after the 2005 reform quality of both emotional and educational support dropped and concluded that even the physical safety of children in childcare settings was not safeguarded. In subsequent measurements, the emotional quality increased after substantial investments form the government, yet the educational support remained unsatisfactory. In her Ph.D. research on the hybrid Dutch childcare system (0–4 years), based on the comprehensive pre-COOL study, Van der Werf (2020) found that process quality in value-based non-profit providers is highest while it was lowest in large-scale for-profit providers, and this was the case for educational support, emotional support as well as for how staff experienced their working conditions. However, the numbers of high-quality value-based non-profit providers have decreased between 2012 and 2017–2018 under the influence of the aforementioned process of marketization.

The commodification of ECEC is often legitimized by framing parental *choice* as an important issue. It is believed that parents—as consumers will drive providers to ensure quality. However, time and again, research shows that parents cannot be considered critical consumers, picking the better childcare for their children. They lack the necessary information to do so (Marangos & Plantegna, 2006), may use other criteria than process quality and tend to overestimate ECEC quality compared to experts' ratings (Barros & Leal, 2015; Bassok, Markowitz, Player, & Zagardo, 2018; Grammatikopoulos, Gregoriadis, Tsigilis, & Zachopoulou, 2014; Howe, Jacobs, Vukelich, & Recchia, 2013; Mocan, 2007). Indeed, process quality—by definition—materializes when parents are absent. It does not come as a surprise that a recent sutdy in Flanders found that there is hardly any relation between quality as measures by experts and parents' appreciation of the childcare used (Janssen, 2020). Moreover, in most countries there is a shortage of places in ECEC in general and in childcare in particular and as a result, parents hardly have a choice. As Burman (1997) rightly argued, the term "consumer" masks practices of coercion within the language of "choice," and it implies equal access to the market that ignores actual structural positions of disadvantage.

### **Accessibility Matters**

Obviously, there is no point in increasing quality when it only serves those who are already privileged. As the knowledge about the beneficial impact of ECEC use on children's outcomes increased, so did the concern among researchers as well as policy makers about inequalities in access. Children from vulnerable families (e.g., families in poverty, migrant or refugee families, Roma, families with children with special educational needs) are less often enrolled in high quality childcare services than their more privileged peers. While detailed figures are not available for all countries (e.g., France does not officially record ethnicity), there is abundance of evidence that this is a global phenomenon. Unequal enrolment has been demonstrated in the US (Hernandez, Takanishi, & Marotz, 2009) and in several European countries, including France (Brabant-Delannoy & Lemoine, 2009), Germany (Büchel & Spiess, 2002), Italy (Del Boca, 2010), the Netherlands (Driessen, 2004; Noailly et al., 2007), Belgium (Ghysels & Van Lancker, 2011), England (Sylva, Stein, Leach, Barnes, & Malmberg, 2007), and some Nordic countries (Wall & Jose, 2004). Reports from country experts in all 28 European Member States in the framework of the European Child Guarantee project show that children from ethnic minorities, refugee children, children with special needs and children from poor families are underrepresented in childcare (Frazer, Guio, & Marlier, 2020; Vandenbroeck, 2019). Inequalities are most outspoken for the youngest children and this is particularly the case in split systems. While differential take-up between high- and low-income groups (or the so-called Matthew effect) is a general feature of ECEC in general and childcare in particular, the degree to which the take-up differs, varies significantly across countries (Van Lancker & Ghysels, 2016). In countries with high overall enrolment rates (Denmark, Estonia, Iceland, Malta) take-up of childcare by vulnerable groups is generally higher, illustrating that universal rights-based policies are more effective in reaching vulnerable families than targeted policies.

The first and main reason for the unequal enrolment is the lack of places and the geographical inequalities in how the available places are distributed. The lack of available places entails waiting lists (e.g., in metropolitan areas in The Netherlands, Belgium, or Latvia) that affect especially those who have more difficulties in subscribing a long period of time before the actual enrolment date. This is the case for parents with precarious work, as well as for immigrant parents. The lack of available places often also entails privatization and commodification of ECEC: in times of budgetary restrictions, legislators may count on private investors to fill the gaps. Smaller or larger companies open up places where the gap between supply and demand is high, but tend to raise parental fees and therefore increase inaccessibility for vulnerable groups. In her Ph.D. research, Van der Werf (2020) also showed that larger for-profit providers tend to be significantly less inclusive to vulnerable families and to cultural and ethnic diversity issues.

The shortage of places for the youngest children may also be influenced by a historical legacy of family policies that favored family models where one parent (i.e., the mother) stayed at home and childcare was considered unnecessary and not educational in nature (Vandenbroeck, 2019). Shortage of places most often goes hand in hand with significant geographical disparities. The disparities may signify a gap between urban and rural areas (e.g., France, Hungary, Lithuania, Romania, or Slovakia). Yet, they may also occur within cities. Van Lancker and Vandenbroeck (2019) calculated that in many (yet not all) major municipalities in Flanders, higher income neighborhoods have more accessible childcare than lower income neighborhoods. Local municipal policies can make a difference when they carefully consider where to provide or expand the number of publicly provided places, yet they not always do so. This does not mean that public childcare necessarily remedies shortcomings of market-based provision. In the city of Vienna, for instance, non-profit private providers tend to fill the gaps that public provisions leave in underprivileged neighborhoods (Pennerstorfer & Pennerstorfer, 2020).

Furthermore, Van Lancker and Vandenbroeck found that over the last decade, the Flemish government predominantly invested in new childcare places where women's employment increased, and less in neighborhoods with lower average incomes, thus favoring labor objectives over the social and educational functions of childcare. Geographical disparities tend to discriminate also against Roma families and refugee families, as childcare centers are usually established at substantial distance from their settlements. As an example, in the Netherlands, only 35% of municipalities with an asylum center have ECEC available for refugee children (Vandenbroeck, 2019).

Another major barrier to accessibility is affordability. Overall, in the 27 European member states plus the United Kingdom, 50% of non-users mention costs as the reason not to enroll their child in ECEC. In Cyprus for instance, the cost of ECEC represents on average 15% of net family income, but for poor and single parent families this is above 60%. In Spain, ECEC represents 5.6% of the disposable income for dual earner families, but over 15% for single-parent families. In Croatia, parental fees may vary between 8 and 16% of net income, according to varying municipal standards (Vandenbroeck, 2019). Some countries offer free preschool (e.g., Ireland, albeit only one year). Yet, in England, the introduction of the entitlement for free years of preschool has not been accompanied by sufficient subsidies, resulting in an increase of parental fees above wages and inflation as well as an increased closure rate of nurseries and childcare providers. Moreover, the English policy of offering a free year of ECEC has mainly reduced fees for ECEC provision, yet failed in attracting a more precarious population (Campbell, Gambaro, & Stewart, 2019).

Over the last decade or so, it has become increasingly popular among policymakers to justify the organization of childcare in terms of parental *choice*. The language of choice often went hand in hand with deregulations and privatization of childcare services and consequently, with demand-side funding. Demand-side funding, meaning funding the parents instead of providers (supply side funding), takes different forms. It may be in the form of a tax deduction or other fiscal measures, as well as through various forms of voucher systems. As Cleveland and Krashinksy (2004) noted, those who favor demand-side funding typically believe that childcare markets work relatively well and that it is important to preserve parental choice. However, studies in countries or regions so diverse as California (Whitebook, Kipnis, & Bellm, 2007), Canada (Cleveland et al., 2007), Taiwan (Lee, 2006), Hong Kong (Yuen, 2007), or The Netherlands (Noailly et al., 2007), showed that demand-side funding tends to increase inequalities in enrolment, despite the rhetoric of choice and despite the use of vouchers for poor families. As an example, the Dutch *Planbureau* (planning bureau) calculated that the marketization of childcare, introduced in The Netherlands in 2005, led to a decrease of providers in rural areas and poor neighborhoods of urban areas, and to an increase of providers in more affluent urban areas, leading to inequalities in actual choice (Noailly et al., 2007).

Another salient example is Finland that in the 1980s introduced a Home Care Allowance (HCA) for parents who decide not to use ECEC. While the rationale is free choice of parents, low-income and single parent families use the HCA more often and for a longer period of time than on average, and the introduction of this policy will therefore probably lead to increasing the educational gap (Repo, 2010).

In relation to the shortage of places, managers of ECEC facilities are forced to set priorities of whom to enroll and who to refuse a childcare place. In many cases, the general rule is "first come first served." This means that waiting lists are created that favor those with regular jobs that can predict their ECEC needs well in advance and that more vulnerable families tend to be excluded. In regions with split systems, preschool is often considered as education for all, while childcare for the youngest bears a historical legacy of serving female employment. As a result, governments may tend to prioritize childcare places for dual earner families. This inevitably results in favoring higher income groups and negative redistribution of public money to those who are already privileged.

In relation to the paradigm of parental choice, it is often assumed that the use of childcare is molded by parental preferences, suggesting that when certain populations do not use childcare, this may be the result of cultural differences and preferences, rather than structural barriers. In an important study, comparing structural barriers and cultural norms about motherhood, Pavolini and Van Lancker (2018) found however that cultural norms that favor mothers as exclusive carers entail lower childcare use overall, but only structural barriers are associated with widening the socio-economic gap in enrolment.

### **Discussion: Policies that Make a Difference**

Childcare for the youngest children has historically been and continues to be a substantial part of family policies, albeit for different reasons. Originating as embedded in hygienist and labor policies, it is now a substantial element of social investment policies (Pavolini & Van Lancker, 2018) and of policies aiming at promoting equality of opportunities (Morabito et al., 2013). As a result, ECEC now entails much more than only caring for infants and toddlers while their mothers are at work. Beyond these economic and laborpromoting functions, it bears a social and educational mission. Consequently, two relatively new concerns have emerged: the concern of educational quality and of accessibility.

By the end of the twentieth century, many policymakers and officials believed that ECEC markets were a reasonable answer to these new emerging needs. It was expected that well-informed *consumers* would make critically informed choices; that supply would follow demand in a system of demandside funding; that vouchers for low income parents could close the enrolment gap, and that the market would be able to level the playing field (Moss, 2008). However, we now know that the promises of marketization in ECEC did not materialize. To give but a few examples, staff turnover appeared to be significantly higher in voucher centers than in contracted centers in California (Whitebook et al., 2007); the number of childcare places was reduced in low-income areas in the Netherlands (Noailly et al., 2007); also in the Netherlands quality decreased (NCKO, 2005: Van der Werf, 2020); and the enrolment gap grew with the establishment of a voucher system in Taiwan (Lee, 2006). This is not surprising. ECEC in general and childcare in particular are, after all, not simply "markets" First, parents are often illinformed and their evaluation of observable quality has hardly any relation with what we know about quality criteria that predict children's outcomes (Janssen, 2020), as parents are by definition absent when emotional and educational supportive relations take place. Moreover, in most countries, childcare places are scarce and even scarcer in poorer areas, so the concept of choice is masking structural inequalities, as Erika Burman (1997) warned us a decade ago. Market-oriented provision in general and for-profit organizations in particular tend to reduce costs and the most salient cost of childcare is the workforce. That explains the phenomenon that the commodification of childcare tends to lead to downwards qualifications, decreased professional support, and lowering working conditions, resulting in decreasing staff satisfaction and increasing staff turnover.

Successful policies combine high quality and access for all. Such policies are policies that consider ECEC as a public good, rather than a commodity. They succeed in balancing the economic, social, and educational functions. Such policies consider childcare as an inextricable part of the educational system, be it with a holistic view on education, meaning a balanced curriculum in which care and education are combined, and adult-centered initiatives are balanced with child-centered approaches. That may mean that childcare and preschool are unified under the umbrella of education policies (as in Sweden) but it may also mean that they are part of a broader welfare policy (as in Denmark). However, it always means that there is a comprehensive policy about continuity between birth and compulsory school age and that ECEC is a public good and publicly financed. Childcare as a public good includes democratic discussion on what ECEC is for (e.g., Lazzari, 2012; Moss, 2014; Vandenbroeck, 2020).

In countries successful in ensuring accessibility and affordability for all families, rights-based ECEC policies overcome the binary opposition between universal and targeted services in a so-called "proportionate universalism" approach. Universal policies often lead to Matthew effects, meaning that ECEC provisions are more often used by higher income families than by lower income families and that, consequently, public money favor the already favored (Pavolini & Van Lancker, 2018). In contrast, targeted provision may reach disadvantaged families, yet often lack public support and all too often services for the poor are poor services. Proportionate universalism, then, a term originating from public health (Marmot, 2010) means that ECEC services are universal and offer additional support for additional needs within the mainstream provision. An example is that (universally accessible) preschools in Flanders may receive additional funding when enrolling more children from vulnerable families. It requires that ECEC places are seen as an entitlement for families, and also that there is an alignment between national and local policy makers (i.e., on the municipal level) that are familiar with local needs. Monitoring the balance between needs and services requires local knowledge about where to implement additional places. In Hungary, for instance, ECEC services receive 105% of normal funding for a disadvantaged child and up to 150% for children with additional needs. In Croatia, Roma parents are exempt from paying kindergarten fees. Successful policies also have introduced income-related parental fees and pay due attention to indirect costs such as costs for meals and transportation. In the case of shortages of places, such policies have implemented priority criteria that balance the economic and social functions by implementing quota for dual earners as well as for vulnerable families. In some Flemish cities, for instance, (e.g., the city of Ghent), all childcare centers have one central enrolment policy that sets quota for different target groups, ensuring that the population using childcare is representative for the entire population of the city. In addition, several regions have engaged social professionals with expertise on the specific needs of the target group of vulnerable families to increase the sensitivity of childcare workers for working with vulnerable families.

Repeated detailed observations have demonstrated that quality in general and educational support in particular needs to be strengthened if we wish to include ECEC as part of equal opportunity policies. Therefore, central quality criteria are necessary, but insufficient if not met by central monitoring systems. Longitudinal studies in diverse countries have demonstrated that monitoring structural quality is not only a necessary means to ensure that vulnerable families have similar quality than more affluent families, but also that it serves to enhance the overall level of quality (Litjens, 2013; OECD, 2015). Standards may include structural quality criteria such as adult-child ratios, staff qualifications and remunerations, professional development opportunities, building expertise in working with children with special needs, or to work in contexts of diversity and multilingualism). The recent quality framework of the European Commission (2018) offers an outstanding example of a comprehensive set of standards.

Yet, we should always bear in mind that not everything that is measurable counts and that not everything that counts is measurable. Central quality standards also need to involve shared values (Urban et al., 2011). Inspiring examples in this vein are the Swedish curriculum (Skolverket, 2010) with its emphasis of the values of democracy; the Danish curriculum (Ministry of Children, Gender Equality, Integration and Social Affairs, 2007) with its focus on participation; the Australian curriculum and its value of belonging (Council of Australian Governments, 2009); the Berlin curriculum respecting diversity (Preissing, 2004) or the New Zealand curriculum on ecology (Ministry of Education, 1996), among many other possible examples. These examples show that how a community constructs the notion of childhood and the child is fundamentally implicated in the practices and policies of that community.

Developing a comprehensive vision on ECEC, that includes its historical economic mission; that balances this mission with contemporary insights about the potential long-term impact on children, families, and communities; and that includes shared democratic values can make ECEC into a powerful part of family policies. However, in order to be more than lip service, in a majority of countries, this will need increased investments in both quantity and quality.

# **References**


Zaouche-Gaudron, C. (2017). *Enfants de la précarité.* Érès.

**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **9**

# **Family Policies and Family Outcomes in OECD Countries**

**Willem Adema, Chris Clarke, and Olivier Thévenon**

Families are a cornerstone of society. Families and the way they function have huge effects on the well-being of their members. Families provide social support networks, offering love, care, and friendship. They also play an important economic role in the production of household goods, and in the provision of protection against hardship.

Families are also a source of economic and social externalities that have major effects on wider society. For example, it is individual adults who decide when and how to establish formal partnerships and when and how to have children. These choices are important to family members themselves, but also have implications for countries as a whole; today's birth rates strongly influence the size of the future labor force, for instance, with knock-on effects on future economic performance, tax revenue and the sustainability of

Organisation for Economic Co-operation and Development (OECD), Paris, France e-mail: olivier.thevenon@oecd.org

W. Adema e-mail: willem.adema@oecd.org

C. Clarke e-mail: chris.clarke@oecd.org

**Disclaimer** The views expressed in this paper cannot be attributed to the OECD or its Member governments: as with any remaining errors, they are the responsibility of the authors alone.

W. Adema · C. Clarke · O. Thévenon (B)

social protection systems. In addition, family decisions vary with the socioeconomic status and contribute to the transmission of inequalities from one generation to the next (OECD, 2018a).

All member countries of the Organisation for Economic Co-operation and Development (OECD) have policies in place to support families with children, though the types of support and the underlying goals differ widely. Family policy is often complex and multifaceted and supports families across a range of overlapping and interdependent objectives (Adema, 2012; Thévenon, 2011). These include: combating child and family poverty; promoting child development and enhancing child well-being; helping parents balance work and family life; mobilizing female labor supply and promoting gender equality to foster economic growth; and, promoting conditions that help adults have the number of children they desire at the time of their choosing. The importance of each of these objectives in the family policy mix varies across countries and over time. Moreover, not all of these objectives have the same function: some seek to "resource" families while others aim rather to "regulate" family behavior and align it to social norms (see Chapter 2 by Daly in this volume).

Since the early 2000s, many countries have increased their support for balancing work and family life (OECD, 2007). Recent cross-national research on family policy has identified family policy models, differentiating in particular by the extent to which countries aim to "de-familialize" or even "degender" child care work (Gauthier & Koops, 2018; Kang & Meyers, 2018; Saxonberg, 2013; for a detailed discussion on these and related concepts, see Chapter 6 by Zagel and Lohmann in this volume).

Overall, the literature assessing the effect of these policies highlights positive effects on family and work outcomes at population level (Thévenon, 2014). In particular, the expansion of parental leave rights and the provision of early childhood education and care (ECEC) has helped increase women's employment (Thévenon, 2016). Together with improvements in public financial supports, this has helped reduce poverty risks for families (Gornick & Nell, 2018; OECD, 2011; Thévenon, 2018). A better work-life balance also has a positive impact on fertility and the well-being of parents and their children (Collins & Glass, 2018; Luci-Greulich & Thévenon, 2013).

However, family policies may also have unintended consequences. Several authors have suggested that while some family policies can strengthen female labor force participation, they may at the same time encourage gender job segregation and reinforce "glass ceilings" on women's careers (Mandel, 2011; Mandel & Semyonov, 2006). Many female workers end up in *feminized occupations* in the public and/or care sectors, where conditions are family-friendly but opportunities for earnings progression are often limited compared to the private sector. For example, the Nordic countries have extensive family supports and high levels of female employment, but there is a high concentration of women in feminized occupations and low female representation in managerial occupations (Datta Gupta, Smith, & Verner, 2006; Mandel, 2011; Mandel & Semyonov, 2006; OECD, 2018b). The effectiveness of family policy measures often depends on the degree of coherence with other policies. For example, participation in ECEC by children under three appears more effective at boosting women's employment when tax and leave policies are also supportive (Thévenon, 2016). Effectiveness also depends on the cultural context. For instance, supportive cultures amplify the relationship between parental leave length, the provision of childcare services for very young children and earnings (Budig, Misra, & Boeckmann, 2012).

Finally, family living arrangements are changing. For example, cultural changes in many countries mean that parenthood outside of marriage is more socially acceptable now than in the past, as are family dissolution and reconstitution. This leads to a diversification of family living arrangements, to which family policies must adapt.

This chapter first provides an overview of changes in family and work behaviors. It highlights the diversification of family models and points to the inequalities associated with these changes. It then presents how policies have developed to reconcile work and family commitments and reduce gender inequality, with results that can vary with the socio-economic status of families. An important challenge remains to better support the most vulnerable families and address the inequalities that are widening with the transformation of family living arrangements.

# **Families and Their Work Arrangements Are Changing**

Families in OECD countries have changed in many ways over the past half-century or so. Only a few decades ago, most families in most OECD countries followed the traditional married couple male-breadwinner model. Birth rates were high by today's standards (if already falling) and separation and divorce were relatively uncommon. Many women left work on marriage or parenthood, and often did not return to the labor market until after children had left education, if at all.

Things look very different today. Across most OECD countries, partnering behaviors have changed, and families are living in increasingly diverse arrangements. Many children now live in "re-constituted" families or "move" between two homes. In the labor market, for most couples, dual-earner families have become the norm.

Marriage, divorce, fertility, and employment patterns differ considerably between (and within) OECD countries. For example, marriage remains a much stronger and more important social institution in the OECD's two East Asian countries—Japan and Korea—than in other OECD countries (OECD, 2019c), while the commitment to full-time dual-earning remains stronger in the Nordic countries than almost anywhere else.

### **Partnership, Marriage, and Divorce**

Partnership patterns have changed dramatically in OECD countries in recent decades. Women's gains in educational attainment—and the associated narrowing of the gender gap in education—is one of the main drivers of change, with major implications for gender equality, the formation and/or dissolution of partnerships and childbearing decisions (Van Bavel, 2012). Women's rising levels of education reduce the economic need for (early) marriage, raise minimum standards for acceptable partners and, at least in theory, increase the chances of women not marrying at all. Women's gains in education have also contributed to an increase in partnership between individuals with a similar level of education (so called "educational homogamy") and a broader shift toward "assortative mating"—i.e., the pairing of individuals with similar socio-economic and educational characteristics (De Hauw, Grow, & Van Bavel, 2017; Esteve et al., 2016). Less-educated men are likely to be the main losers, since they are more likely to remain single and be in low-paid employment than any other group (De Hauw et al., 2017). More "homogamy" in the marriage market also leads to increased income inequalities between couples with high and low educational attainment and income potential (see also Chapter 25 by Nieuwenhuis in this volume).

Marriage is still the dominant form of partnership in the vast majority of OECD countries, but its prevalence has declined considerably. On average, marriage rates have fallen by almost half since 1970; in some countries, including the Netherlands, Portugal, and Slovenia, they have fallen by around two-thirds. Increasingly, many young people are choosing to marry later or not at all. Data from the OECD Family database show that, between 1990 and 2017, the OECD average mean age at first marriage increased by almost six years (OECD, 2020a). Women are now, on average, 30 years old when they marry for the first time; men are almost 33 years old.

Cohabitation is becoming increasingly popular, especially among young people. As of the last round of major population and housing censuses (for most countries around 2011), on average across the OECD 17% of 20- to 34-year-olds reported living as part of an unmarried cohabiting couple. This is not far off the 23% living either as part of a married couple or in a civil or register partnership (OECD, 2020a). Cohabitation is most common among young people in several of the Nordic countries (Denmark, Norway and Sweden), and Estonia, France, the Netherlands, and New Zealand (OECD, 2020a). In addition, many people are choosing new forms of partnerships, including relationships that involve partners keeping their own place of residency, "weekend-relationships", and couples "living apart together".

Alongside declining marriage rates, the frequency of divorce has risen considerably. Between 1970 and 2017, the OECD average crude divorce rate—the number of divorces per 1000 people—increased from 1.4 per 1000 people to 2.0 per 1000 people (OECD, 2020a). In several OECD countries, including Belgium, France, Portugal, and Spain, each year there is now more than one divorce for every two marriages. The increase in divorce rates has opened up new opportunities for re-partnering. In 2017, on average across OECD countries, almost one in five marrying persons had previously been divorced (OECD, 2020a).

Divorce behavior has also changed across socio-economic lines. While divorce rates increased across all types of couples during the 1960s and 1970s, trends have since leveled off and then reversed for highly educated couples but continued to rise among couples with low levels of educational attainment (Härkönen & Dronkers, 2006). The impact of divorce on families' standard of living is typically high, and just under one-third of single-parent families are income-poor, which is three times higher than the poverty rate for two-parent families. Moreover, children who have experienced divorce have on average lower psychological well-being, more behavioral problems, and lower educational achievements than children who grew up in "intact" families (Amato & Boyd, 2013). The higher prevalence of income poverty among single-parent families is one explanation, but the loss of income experienced by more affluent families also contributes to the lower achievements of children with divorced parents (Bernardi & Boertien, 2016).

### **Parenthood and Fertility**

Alongside changing partnerships, fertility behaviors have changed radically. Improved access to contraceptives has given more adults control over the timing of births and birth rates have fallen everywhere in the OECD over the past half-century: the OECD area as a whole now records two million (or 15%) fewer births each year than it did in the 1960s (OECD, 2019d). The OECD average total fertility rate—measuring the average number of children born per woman given current age-specific fertility rates—has almost halved over the same period, falling from 3.17 children per woman in 1960 to 1.65 in 2017. Chile, Mexico, Turkey, and especially Korea recorded the sharpest declines (OECD, 2020a).

One major driver of the long-run decline in fertility is that couples are increasingly postponing parenthood until after they have established themselves in the labor market (Luci-Greulich & Thévenon, 2013; OECD, 2011). Across OECD countries with available data, the average age at which women first give birth has increased by roughly three years since the mid-1990s, from 26.2 in 1995 to 29.1 in 2017 (OECD, 2020a); in ultra-low fertility countries such as Italy, Japan, Korea, and Spain, the average age of a woman at first birth now exceeds 30. In Japan and Korea, childbirth remains closely associated with marriage; not so in other OECD countries, where the decline in marriage rates has been accompanied by an increase in the number of births outside marriage (OECD, 2020a).

The postponement of the first birth leads to a narrower age-interval in which women can have children, leaving less space for multiple births and large families. Between 1960 and 2017, on average across the 18 European OECD countries with available data, the share of births that were third or higher order births fell from just over one-third to just under one-fifth. Today, only about 5% of households with children contain three or more children on average across the OECD (OECD, 2020a).

Postponing the first birth also increases the risk of involuntary childlessness. In addition to those women who cannot conceive or decide not to have children, the upper limit to the childbearing years, set by the so-called biological clock, makes it difficult for women who delay parenthood to have children at later ages. Despite advances in fertility-related medical procedures, definitive childlessness has increased in many (although not all) OECD countries in recent years (MPIDR & VID, 2019).

Changes in parenting styles coincide with the postponement of births. Older parents usually have more resources to invest in children than younger parents, and parents with high levels of educational attainment devote more material resources to parenting and child education than parents with lower education usually can do (Doepke & Zilibotti, 2019; Kalil, 2015). These differences contribute to the transmission of disadvantage and the limited degree of inter-generational social mobility that is observed across OECD countries (OECD, 2018a).

#### **Family Living Arrangements**

Changing partnership patterns and falling birth rates have had a profound effect on the types of families in OECD countries. Declining fertility and population aging have led to a growing share of households without children. In 2015, on average across OECD countries, more than two-thirds of households did not contain any children (OECD, 2020a).

Most children in most OECD countries still live with two married or officially registered/partnered parents (Fig. 9.1). However, the share of children living with cohabiting parents has increased by 50% in the last decade, on average across OECD countries with available data (OECD, 2020a). Living with two cohabiting parents is most common in countries where marriage occurs late and births outside marriage are common (e.g. France, Iceland, Slovenia, and Sweden). Countries where births to unmarried couples are rare (e.g. Greece, Italy) or divorce is common (e.g. Lithuania, and the United States) have much lower shares of children living with cohabiting parents (Fig. 9.1).

In 2017, on average across OECD countries, about 17% of children (aged 0–17) were living with one parent (Fig. 9.1). The share of children living in single-parent families is highest (at around or above 25%) in countries with historically high divorce rates (e.g. Belgium, Latvia, Lithuania and the United States). It is lowest (at less than 10%) in countries like Greece, Poland, and Turkey, where both divorce and births outside marriage remain relatively uncommon.

Increasing divorce rates and the growth of re-partnering mean that "reconstituted" families, including step-, mixed- and blended families, are becoming more frequent (Bernardi & Mortelmans, 2018). The complex nature of reconstituted families, together with the limitations of traditional household surveys, means that comparable information on the frequency and structure of reconstituted families is scarce (Miho & Thévenon, 2019). However, data from the Health Behaviour in School-age Children (HBSC) study suggest that across participating OECD countries, close to 10% of adolescents (11- to 15-year-olds) live in a step family of some form (Inchley et al., 2016). In some OECD countries, such as Belgium, Estonia, and Finland, this increases to

**Fig. 9.1** On average across the OECD, one in five children live with one parent or less. Distribution of children (aged 0–17) by presence and marital status of parents in the household, 2018 or latest (*Note* "Parents" generally refers to both biological parents and step-, adoptive parents. "Living with two married parents" refers to situations where a child lives in a household with two adults that are considered parents and these parents are married to each other. "Living with two cohabiting parents" refers to situations where a child lives in a household with two adults that are considered parents and these parents are not married to each other. "Living with a single parent" refers to situations where a child lives in a household with only one adult that is considered a parent. "Other" refers to a situation where the child lives in a household where no adult is considered a parent. For Japan and Mexico, children aged 0–14. Data for Mexico refer to 2010, for Australia to 2012, for Japan to 2015, for Canada and Iceland to 2016, and for France, Hungary, Ireland, Luxembourg, Turkey, the Slovak Republic, and Switzerland to 2017. Exact definitions vary for some countries. See the OECD Family Database [http://www.oecd.org/els/ family/database.htm] Indicator SF1.2 for more detail. *Source* OECD Family Database, http://www.oecd.org/els/family/database.htm)

14% or more (Inchley et al., 2016). The rise of shared custody also increases the likelihood of children moving from one household to another on a regular basis. According to data from the Children's Worlds survey—which covers 15 countries worldwide, including eight OECD countries—more than 5% of 10- to 12-year-olds in Israel, Germany, and Spain, and over 10% of 10- to 12-year-olds in Estonia, Norway, and the United Kingdom, report "regularly sleep[ing] in two homes with different adults" (Rees & Main, 2015).

Re-partnering and the establishment of a reconstituted family often takes place gradually. Re-partnered couples frequently continue to live apart for at least a certain length of time, perhaps with one partner "commuting" from one household to another or with the couple living together only on a parttime basis.

Re-partnered couples make up a significant share of non-cohabiting couples. In Belgium (28%), France (27%), the Netherlands (33%), and Norway (28%), more than one-quarter of all non-cohabiting couples are partners aged around 40 who have had previously been in a union of some form, and who intend to start living together within the next three years (Mortelmans et al. 2015).

#### **Family Work Arrangements**

Women are increasingly engaged in the labor market, as related to women's gains in educational attainment, enhanced family-friendly policies, and changing attitudes toward women's roles in society. Ireland and Spain have seen some of the largest increases in recent years: in both, the employment rate for 15- to 64-year-old women has grown by well over two-thirds since 1990 (OECD, 2020b). In contrast, in the Nordic countries, female employment gains have been limited—or even negative—as female employment rates in these countries were already high in the 1980s (OECD, 2018b).

A large part of the increase in women's employment has been driven by a growing number of women staying in paid work after becoming parents. On average across OECD countries in 2014, two-thirds of women with children (aged 0–14) were in paid employment (Fig. 9.2). Maternal employment rates are highest in countries like Denmark, Slovenia, and Sweden, where extensive public ECEC policies allow both parents to return to work after a few months of paid leave. They are lower in many of the Eastern European OECD countries, where family policy provides financial incentives for mothers to stay at home at least until children enter pre-primary education at around age three, and in several of the Southern European and Latin American OECD countries, plus Turkey, where working parents are offered little work/life balance support in general.

Highly educated women are particularly likely to continue working after becoming mothers. In 2014, on average across OECD countries, more than three-quarters (78%) of highly educated mothers (with children aged 0–14) were in paid employment, compared to less than half (42%) of mothers with low educational attainment (OECD, 2020a). A particularly concerning trend is that, in several OECD countries, the employment gap between mothers with high- and low levels of education is growing, not shrinking. In Canada, for example, the gap has increased by one-third since the mid-2000s, largely because employment rates for mothers with low education fell sharply following the financial crisis and they have not yet recovered (OECD, 2020a). Similar patterns can be found in Belgium, France, Ireland, the Netherlands, the United Kingdom, and the United States.

**Fig. 9.2** The level and intensity of maternal employment varies considerably across OECD countries. Employment rates for women (15- to 64-year-olds) with at least one child aged 0–14, by part-time/full-time status, 2014 or latest (*Note* Part-time employment is defined as usual weekly working hours of less than 30 hours per week in the main job. In some countries, it is possible for individuals to report that they do not have usual set hours in their main job. Where this is the case, the individual's actual hours worked in their main job during the survey reference week are used in place of their usual weekly working hours. For Chile, the distinction between part-time and full-time work is based on actual hours worked in the main job in the previous week, rather than usual weekly working hours in the main job. No distinction between part-time and full-time employment in Australia, Israel, Japan, Mexico, Sweden, and Switzerland. Data for Denmark and Finland to 2012, and for Chile, Germany, and Turkey to 2013. For Japan, women of all ages, and for Sweden women aged 15–74. Children aged 0–15 for Canada, 0–14 for Japan, 0–18 for Sweden, and 0–17 for the United States. The OECD-31 average excludes Japan. *Source* OECD Family Database, http://www.oecd.org/els/family/database.htm)

In several OECD countries, many mothers use part-time work as a means of combining work and family responsibilities (Fig. 9.2). Part-time maternal employment is most common in Austria, Germany, Switzerland, the United Kingdom, and especially the Netherlands, where over two-thirds of working mothers work part-time (OECD, 2019b). Short working hours are less common for mothers in other countries, especially the Nordic countries, where comprehensive childcare supports help mothers work full-time, as well as many of the Baltic and Eastern European OECD countries like the Czech Republic, Hungary, Latvia, and the Slovak Republic, where part-time opportunities in general are rare.

Increasing maternal employment has contributed to the growth of dualearning. Most two-parent families in most OECD countries are now two-earner families, even if there are still large gaps in earnings between parents (OECD, 2017b). In some countries, especially Austria, Germany, the Netherlands, and Switzerland, one-and-a-half earner families (with one partner working full-time and the other part-time) are most common (OECD, 2020a). In others (e.g. Denmark, Slovenia and Sweden), the majority of two-parent families are full-time dual-earner families. Only a minority of families continue following the traditional single-earner model. Indeed, on average in 2014, fewer than one-third of couples with at least one child (aged 0–14) had one partner working full-time and the other not working at all (OECD, 2020a). In Belgium, Denmark, the Netherlands, and Sweden, this was below one-fifth.

Despite the gains made by women in the labor market, women in OECD countries still do far more unpaid work than men. On average across OECD countries with recent data, women spend just over two hours more each day on unpaid work than men (OECD, 2020d). In some countries (Mexico and Turkey), the gender gap remains as large as four hours. In most OECD countries, women's disproportionate hours spent on unpaid work result in women spending more hours in total on combined paid and unpaid work (OECD, 2020d). In all but three OECD countries (the Netherlands, New Zealand and Norway), women devote more time than men on combined hours of paid and unpaid work.

## **Family Policies in OECD Countries**

Family policies have become a core part of national social protection systems in OECD countries over recent decades. Public spending on families has grown almost everywhere in the OECD, reflecting shifting priorities and the greater emphasis many countries now place on childhood. In dollar terms, on average, public spending per head on family cash benefits and services has more than doubled since the early 1990s, rising from USD 417 in 1990 to USD 844 in 2015 (OECD, 2020c). In some countries, including Chile, Japan, Korea, Mexico, and Spain, it increased by more than four times over the same period (OECD, 2020c).

There are many differences in countries' family policy approaches (Fig. 9.3). All OECD countries provide family support in at least some form, but differences in countries' histories, their attitudes toward marriage and gender roles, the role of government, and the relative weight given to the various underlying family policy objectives mean that each takes their own approach to family policy. An insightful and well-known literature has grown up around trying to describe these differences (e.g. Eydal & Rostgaard, 2018; see also Chapter 2 by Daly in this volume).

**Fig. 9.3** The level and type of public family support differs strongly across OECD countries. Public expenditure on family benefits by type, as a % of GDP, OECD countries, 2015 (*Note* Public spending accounted for here concerns public support that is exclusively for families [e.g. child payments and allowances, parental leave benefits and childcare support], only. Spending in other social policy areas such as health and housing support also assists families, but not exclusively, and is not included here. Spending on in-work benefits such the Earned Income Tax Credit in the US can be important to families [worth around 0.5% of GDP], but the program is also open to workers without families and therefore not categorized as a family benefit. The data in Panel A cover public expenditure on family cash and in-kind benefits only, and do not include spending on tax breaks for families. Data for the Netherlands and New Zealand refer to 2011, and for Poland to 2014. For Lithuania, data on tax breaks toward families are not available. The OECD-32 average excludes Lithuania, the Netherlands, New Zealand, and Poland. *Source* OECD Family Database, http:// www.oecd.org/els/family/database.htm)

Some OECD countries, most notably the Nordic countries, provide service-heavy family supports to families with young children aimed primarily at encouraging full-time dual-earning and fostering child development. These countries provide parents with a continuum of support, from birth up to when children leave school. Parents can access generous paid parental leave for at least a few months after birth, followed by an entitlement to a place in a local public ECEC center, and, particularly in Denmark and Sweden, out-of-school-hours care services for children in full-time education up to at least age 11–12.

Other countries focus more on providing families with financial support through family cash benefits and tax breaks (Fig. 9.3). In some countries (e.g. the Czech Republic and Hungary), this is done largely through universal cash benefits provided to all families. These benefits are often structured in such a way as to encourage one parent (typically the mother) to care for children at home, at least until they enter pre-primary education at around age 3. Other countries (e.g. Australia, New Zealand, and the United Kingdom) put greater emphasis on targeted benefits aimed at achieving specific objectives or directed at specific groups such as single-parent families or families on low incomes. France is something of an outlier, in that it combines both generous (and largely universal) cash and tax benefits with relatively high spending on family services, especially for families with children in pre-primary education.

Some OECD countries (e.g. Greece, Mexico, and Turkey) provide relatively little in the way of either cash benefits or public services for families (Fig. 9.3). However, in several of these countries, limited public spending on families is balanced by comparatively low effective tax rates when entering work and, in some cases, considerable tax advantages for dual-earner families. In these countries, the emphasis is on delivering family income through market earnings rather than state support.

In many respects, cross-country differences in family policy continue to be as distinct today as they were when first discussed in the 1990s. OECD countries have not 'converged' on one single family policy model (Adema, Ali, & Thévenon, 2014; Gauthier, 2002); both the level and types of support provided to families remain diverse.

At the same time, however, there are clear common trends and developments in family policy that stretch across most if not all OECD countries. In particular, in many OECD countries, policy has shifted in recent years to become more supportive of women's employment, as well as to be more encouraging of a more equal gender division of labor. The following provides three specific examples of these developments: the growth of public support for ECEC; the rise of fathers-only paid leave; and the introduction in several countries of new measures to support flexible working.

#### **Public Support for Early Childhood Education and Care**

Public support for ECEC has grown substantially in OECD countries over the past few decades. Recognizing the significance of good-quality ECEC for child cognitive and social development (Camilli, Vargas, Ryan, & Barnett, 2010; Havnes & Mogstad, 2011; OECD, 2013), as well as the importance of ECEC for parental employment (Del Boca, 2015; Jaumotte, 2003; Olivetti & Petrongolo, 2017; Thévenon, 2013), OECD countries are increasingly introducing measures to improve access, affordability, and quality in ECEC (see also Chapter 8 by Vandenbroeck in this volume). Since around 2000, per head public spending on ECEC has increased in all OECD countries other than Denmark (where public spending on ECEC was already very high) and

**Fig. 9.4** On average across OECD countries, public spending on early childhood education and care has almost doubled since the turn of the century. Public expenditure per head on early childhood education and care, constant (2010) USD PPP, 2000 and 2015 (*Note* In some countries, local governments play a key role in financing and providing childcare services. Such spending is comprehensively recorded in Nordic countries, but in some other [often federal] countries it may not be fully captured by the OECD social expenditure data. For Poland, data refer to 2014 rather than 2015. *Source* OECD Family Database, http://www.oecd.org/els/family/database.htm)

the United States (Fig. 9.4). Over the same period, the OECD average per head spend grew by more than three-quarters, from USD 146 in 2000 to USD 272 in 2015 (Fig. 9.4).

Germany provides one of the clearest examples of growing public support for ECEC (OECD, 2017a). Beginning with increased public investment during the mid-2000s, the German federal government has since (in 2013) introduced a Nordic-style legal entitlement to ECEC for all children age one and over, followed by a series of major spending programs, including on all-day ECEC (the KiTaPlus program) and through the recent Gute-Kita Gesetz (the 'Good Child Care Law'). Partly as a result, the number of children enrolled in ECEC increased by over 25% between 2008 and 2017, from 1,565,000 to 2,024,000 (Destatis, 2018). Germany still faces many challenges in ECEC—supply continues to fall short of demand, for instance, and providers are likely to face severe staff shortages over the next decade or so (BMFSFJ, 2019; OECD, 2019a)—but further reforms and investment are in the pipeline.

Korea provides a second example. Since the early 2000s, the Korean government has introduced a series of major reforms aimed at increasing the availability and affordability of ECEC services, including extensive subsidies for childcare providers and generous cash benefits for parents using ECEC (OECD, 2019c). Korea backed these reforms with heavy public investment: real-terms per head public spending on ECEC increased by over 18 times between 2000 and 2015, from USD 17 to USD 326 (Fig. 9.4). The result is some of the lowest out-of-pocket childcare costs in the OECD, and ECEC enrolment rates that are now well above the OECD average (OECD, 2020a).

The growth in public support for ECEC had a considerable impact on enrolment. On average across European OECD countries, participation in formal ECEC by children under age 3 increased by well over one-third in the decade to 2017, from 26% in 2007 to 35% in 2017 (Eurostat, 2019). In Korea, the participation in center-based services by children under age 3 has grown by more than ten times since the early 2000s, from 3% in 2001 to 40.9% in 2018 (Statistics Korea, 2018).

There are, however, still many challenges in ECEC policy. Several European OECD countries (e.g. the Czech Republic, Hungary, Poland and the Slovak Republic) are still a long way from meeting the Barcelona target of 33% participation by children under age three that should have been met by 2010. Quality in ECEC is also an ongoing concern, as is affordability, particularly in many of the English-speaking OECD countries (OECD, 2018c). Related to this, there are also important ongoing concerns around equity, since in many OECD countries children from less advantaged backgrounds are much less likely to participate in ECEC than their better-off peers (OECD, 2016b; Pavolini & Van Lancker, 2018; Van Lancker & Ghysels, 2016; see also Chapter 8 by Vandenbroeck and Chapter 11 by Hook & Li in this volume). Gaps in participation across socio-economic groups are often widest in countries that rely largely on private service provision (e.g. Ireland, the Netherlands, and the United Kingdom), as well as in France, where affordable public services for the under-threes are in short supply (HCFEA, 2018; OECD, 2016b; Van Lancker & Ghysels, 2016).

#### **Statutory Fathers-Only Paid Leave**

Encouraging fathers to take paid leave has been a long-standing challenge for OECD countries (see also Chapter 15 by Bartova & Keizer in this volume). In many OECD countries, fathers have had the right to take paid leave for decades, usually through parental leave entitlements sharable among partners. However, in almost all cases, the use of sharable leave has been dominated by mothers (Moss, 2015). There are several possible reasons. Fathers often earn more than their partners (OECD, 2017a), so unless leave benefits (almost) fully replace previous earnings it usually makes economic sense for the mother to take the bulk of the leave. Societal attitudes toward the roles of mothers and fathers in caring for young children and concerns around potential career implications also contribute to a general reluctance among many fathers toward taking leave (Duvander, 2014; Rudman & Mescher, 2013).

To encourage fathers' use of paid leave, countries from across the OECD are turning to "fathers-only" paid leave entitlements that cannot be used by the mother and are lost if not taken by the father. In many cases, this means paid paternity leave—usually very short but often well-paid periods of leave that fathers can use within the first few months after childbirth. A few countries (e.g. Belgium, Luxembourg, and Spain) have actually provided paternity leave lasting at least a day or two since the 1960s, but it was not until the 1980s that OECD countries introduced anything longer than three days (OECD, 2020a). In 1980, Sweden became the first OECD country to provide fathers with a paid paternity leave lasting at least one week. In 1984, Denmark became the first to offer two weeks or longer (OECD, 2020a). In the years since, a further 18 OECD countries have introduced paid paternity leave in some form. This includes a diverse range of countries from across the OECD's regions, including Australia, Chile, Korea, Mexico, Latvia, Lithuania, Poland, and Turkey. In 2016, just under two-thirds (23) of OECD countries provided fathers with a statutory entitlement to at least one day of paid paternity leave (Fig. 9.5).

By 2016, 12 OECD countries (Fig. 9.5) provided two months or more of "fathers-only" paid parental leave—usually longer periods of paid leave that again can be used only by the father but that need not necessarily be taken directly around childbirth. These "fathers-only" parental leaves often take the form of "father quotas", or specific portions of an overall parental leave period that are reserved exclusively for the father, as in many of the Nordic countries. Other options include "father bonuses", where a couple may qualify for some extra weeks of paid leave if both parents use a certain amount of shareable leave, as in Germany, for instance—or more simply the provision of paid parental leave as an individual, non-transferable entitlement for each parent (e.g. Belgium, Japan, Korea, and Luxembourg).

Norway was the first OECD country to implement fathers-only paid parental leave, with the introduction of a four-week father quota in 1993 (OECD, 2020a). Sweden followed closely with a similar father quota in 1995. However, it was not until the 2000s that fathers-only parental leave became common cross-nationally. Nine OECD countries introduced or expanded fathers-only paid parental leave during the between 2000 and 2009, including Germany, Iceland, Portugal and, most notably, Japan and Korea, where fathers now hold an individual non-transferable entitlement to one year of paid parental leave (although in the latter two countries, take-up among fathers remains low).

**Fig. 9.5** The majority of OECD countries provide paid paternity leave, and onethird offer fathers-only paid parental leave. Paid paternity leave and paid fatherspecific parental and home care leave, in weeks, 2016 (*Note* Information refers to entitlements to paternity leave, "father quotas" or periods of parental leave that can be used only by the father and cannot be transferred to the mother, and any weeks of sharable leave that must be taken by the father in order for the family to qualify for "bonus" weeks of parental leave. Data refer to entitlements in place as of April 2016. Data reflect entitlements at the national or federal level only, and do not reflect regional variations or additional/alternative entitlements provided by states/provinces or local governments in some countries [e.g. Québec in Canada, or California in the United States]. *Source* OECD Family Database, http://www.oecd.org/ els/family/database.htm)

Because fathers-only leave remains relatively new in most OECD countries, it is still somewhat difficult to estimate precisely the effects on leave take-up and the downstream impact on engagement in childcare and unpaid work. Nonetheless, a growing literature from across the OECD points in many cases toward positive effects. Several quasi-experimental studies find that the introduction of a fathers-only leave increases leave take-up by men (see, for example, Cools, Fiva, & Kirkebøen, 2015; Kluve & Tamm, 2009; Patnaik, 2018). In Sweden, for instance, the introduction and subsequent expansion of a "father quota" led not only to increases in the number of fathers using any leave, but also to a steady and sustained increase in the share of days used by men (Duvander & Johansson, 2012; Ekberg, Eriksson, & Friebel, 2013). Moreover, there is also a growing body of quasi-experimental evidence to suggest that fathers-only parental leave produces lasting positive effects on fathers' engagement in various forms of unpaid work and childcare (Kotsadam & Finseraas, 2011; Patnaik, 2018; Tamm, 2018). A drawback of fathers-only leave is that it may discourage fathers from having additional children (Farré & González, 2019), while it may instead encourage fertility in a more gender equitable environment (Duvander et al., 2016).

#### **Support for Flexible Working**

Historically, compared to many other areas of family policy, the provision of flexible working supports has often been left to the employer and is subject to employee-employer negotiations (see also Chapter 21 by Chung in this volume). Flexible working arrangements are often informal, especially for workers in smaller firms. Modest and irregular arrangements, such as occasionally taking an hour or two off work for family reasons, are more common than more substantial arrangements such as working from home on a regular basis (OECD, 2016a).

However, there are strong equity concerns involved with leaving the provision of flexible work to the market alone. All else being equal, higher-skilled workers with greater bargaining power are more likely than others to be able to negotiate access to flexible working arrangements. Indeed, looking across European countries, professionals and managers enjoy much greater access to flexible working than low-skilled workers and those in clerical or service jobs (OECD, 2016a, 2017b). Workers in lower skill occupations are also much less likely to work from home, largely because work is organized on the basis of fixed schedules (OECD, 2016a, 2017b; see also Chapter 22 by Begall & Van der Lippe in this volume). And yet, in many cases, it is lowerskilled workers on lower wages who likely need flexible working arrangements the most. These workers are less likely to be able to afford to work parttime, and less likely to be able to buy in external services (such as additional childcare or cooking and cleaning services) to help balance work and family commitments.

In recent years, OECD governments have increasingly stepped-in to help promote flexible working for all workers regardless of status or occupation. Many OECD countries (e.g. Australia, Austria, the Netherlands, New Zealand, Norway, Portugal, Sweden, and the United Kingdom) now provide workers with a statutory right to at least request reduced working hours. Several of these countries also provide a complementary right to return to full-time work and/or automatically revert to previous hours after a certain specified period (Blum et al., 2018; OECD, 2017b). Some countries (e.g. Australia, Austria, Norway, Portugal, Slovenia, and Sweden) restrict these rights to parents with young children and/or workers with caring responsibilities, while others provide the right to all workers regardless of circumstances (e.g. the Netherlands, New Zealand, and the United Kingdom). In most countries, employers can refuse requests on business grounds, although in several (e.g. Portugal, Norway, and Slovenia) employers cannot refuse while children are under a certain age (Blum et al., 2018; OECD, 2017b).

The Netherlands and the United Kingdom provide two of the most comprehensive examples of the "right to request" policy (Blum et al., 2018; OECD, 2017b). While in most OECD countries employees are entitled to request changes to working hours only, in the Netherlands and the United Kingdom, the right extends to other areas of flexible working, such as the scheduling of working hours and the place of work. In both cases, these rights apply to all employees who have worked for their current employer for at least six months, although in the Netherlands, employers with less than ten employees are exempt. Also in both cases, employers can refuse if there is a clear business reason for doing so (Blum et al., 2018; OECD, 2017b).

An ongoing challenge for public policy is that, in many countries, flexible working arrangements are often regulated by collective bargaining or enterprise-level agreements (Cabrita, Boehmer, & Galli da Bino, 2016; Hegewisch, 2009). Nevertheless, governments can still help by providing information of flexible, facilitating the exchange of best practice, and encouraging collective bargaining on flexible workplace issues (OECD, 2016a).

## **Concluding Remarks**

This chapter has highlighted important and relatively recent developments in family policy aimed largely at helping parents to balance work and family commitments and reducing gender inequality. These developments have helped promote women's employment in particular, with positive effects on poverty risks and fertility. However, the most disadvantaged families often do not make use of ECEC and flexible working arrangements to the same extent as higher income groups. An ongoing challenge for all OECD countries is to improve family policies so that they better serve the most needy families and better counteract growing income inequalities.

A second challenge is to adapt family policies to changing family living arrangements. In particular, family dissolution and reconstitution are more common than in the past and patterns of family reconstitution today more heterogeneous than they were only few years ago. This creates new challenges for policies to support single parents (see also Chapter 13 by Maldonado & Nieuwenhuis in this volume) and child support policies (see also Chapter 12 by Skinner & Hakovirta in this volume). A growing number of parents cohabit informally, and a growing number of children move from one household to another on a regular basis. This poses obvious challenges to calculating benefit entitlements and service delivery, but it is vital that children have access to family supports regardless of the marriage status of their parents, and in this respect at least enjoy equal starting opportunities in life (Miho & Thévenon, 2019).

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **10**

# **Family Policies Across the Globe**

**Fernando Filgueira and Cecilia Rossel**

Family policies encompass actions aimed at supporting families and regulating family life. In a broad sense, policies that affect family life and that can support families can and do cover many areas of public policy.1 We focus here on three main areas of family policy: income support for families with children, early childhood education and care (ECEC) and maternity, paternity and parental leave (see Chapter 2 by Daly in this volume). We zero in on family policies that seek to support families especially during the first years after childbirth. In addition, and more recently, some of these policies also aim to modify private family relationships and, more specifically, how maternity and paternity are lived by societies (Lewis, 1992; Saraceno, 2018; Thévenon, 2011). Despite this general orientation, the way and extent to which different countries have followed these goals varies significantly across

F. Filgueira

© The Author(s) 2020 R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_10

<sup>1</sup>Many policies that are not meant to address explicitly family issues can a do have provisions that affect families. Policies in education (for example providing meal services and extending the length of the school day), health (i.e. suppressing co-payment in maternal and early childhood health checkups and interventions), housing (preference in provision of public housing or special interests rates in housing credit aimed at families with children), and even macroeconomic policies (joint or separate taxation schemes) affect family life.

Facultad de Ciencias Sociales, Universidad de la República, Montevideo, Uruguay

C. Rossel (B) Universidad Católica del Uruguay, Montevideo, Uruguay e-mail: cecilia.rossel@ucu.edu.uy

countries and regions in the world. In particular, issues such as coverage and quality, but also design and context of implementation vary significantly. Moreover, countries with very similar development levels might present very different patterns in terms of fiscal efforts regarding family allowances, work leaves, and child care services.

We analyze family policies across the globe, describing patterns in the development of family allowances, leave schemes, and ECEC services both in developed and developing regions. To guarantee fair comparisons between regions, whenever possible our analysis is based on the OECD family database and the ILO global social protection database. However, we also provide further insights using regional data sets and literature focusing on the analysis of family policies in different regions.

Both the literature stemming from the economics of families (Becker, 1981) as well as the literature inspired in the idea of a needs-based/culturally led second demographic transition (Lesthaeghe, 2010; van de Kaa, 2001) posit that as countries develop and enter the post-industrial economy, the sharp divide between the roles of men and women and the traditional nuclear married family should give way to a more fragmented and unstable family landscape and to dual earner households. In addition, such changes should push fertility rates later in calendar and below replacement levels.

Regarding developing countries, this literature would predict a movement toward replacement level fertility and an expansion of the nuclear male breadwinner family. This is due to the changing role, power and preferences of women and to the decreasing efficiency and required functions of extended households in industrial societies, for instance because the state extends polices to support the economic autonomy of the elderly population. If cultural trends continue to move toward more secular and individual self- realization, and if the educational level and labor market involvement of women grows and their control over reproductive choice is secured, these theories would predict a decline in "doing family" (i.e. later nuptiality, increased divorce rates, later and lower fertility, or even childishness). While moderately accurate in the past, these theories are increasingly confronted with empirical trends that make such claims problematic, since the new equilibrium seems to vary quite importantly by region and welfare regime type. While some European countries present close to replacement level fertility and have also seen a recent diminishing rate of divorce and moderate expansion of nuptiality, in the industrialized countries of East Asia fertility rates are at the low/low frontier or below and childless women have increased markedly. In sum, while some regions seem to move to the limit of "undoing" family others seem to remain rather robust at such practice.

Latin America high income countries show a sharp decrease in fertility but one that is slow to converge across social classes suggesting an increasing bipolar pattern in reproduction and one that does not fit easily with either theory. While the very low fertility of the upper middle classes could follow from Becker-type and second demographic transition arguments, the persistence of high and especially early fertility among low income families together with their highly unstable and fragmented nature is problematic. Also contrary to theory, the region shows highly informal family arrangements and increasing instability and fragmentation in the types of families across all social classes.

Partly what is missing in the theories predicting such outcomes is the importance of family policies and how such policies might lead to very different "equilibriums" regarding family types, nuptiality and divortiality, fertility and the role of men and women within. While this chapter makes no attempt to solve this puzzle, it does provide a wide comparative descriptive assessment of family policies as a first step tackle some of such empirical trends and puzzles.

In the next section we focus on the demographic changes in family structures and its relationship with different configurations in state policies toward families with small children. Then, the main features of the European experience in family policies are presented. The following sections describe the development of family allowances, work leaves, and ECEC services in other regions of the world—both developed and developing—comparing them to the European countries. The final section presents concluding remarks regarding the main challenges related to the consolidation of these policies across the globe.

# **Demographic Stages, Economic Development, and Family Policy**

Fertility rates have been dropping consistently around the world for the last 25 years. With the exception of the most developed countries (European Union) where fertility has slightly bounced back after hitting their lowest mark by the end of the twentieth century, the middle income and upper middle-income countries of the developing world are converging toward replacement level fertility (Fig. 10.1).

This, combined with an aging population (which implies less women in fertile ages and more elderly in the total population) will translate into a lower proportion of children and teenagers. Except for sub-Saharan Africa, where

**Fig. 10.1** Evolution of fertility rates by region, 1992–2015 (*Source* Prepared by the authors based on World Bank Open Data, updated 2016. Regions in the figure follow the definition of the World Bank Open Data Bank)

the population is still very young and aging has not advanced, the proportion of children has dropped quite dramatically and will continue do so in almost all regions.

Overall, it can be stated without much doubt that, as emerging and developing countries grow, fertility rates will continue to decline, the proportion of elderly will continue to increase and the proportion of children in the overall population will decrease (up until a certain point). Dependency ratios will also decrease in most regions of the world. This opens a window of demographic opportunity, since there is room for more social and economic investment in smaller child cohorts when fertility drops and the elderly are not yet a larger share of the population. Thus, similar levels of overall spending in children could imply quite different per-capita spending. Furthermore, since overall dependency ratios will go down, there should be economic and fiscal space to further increase spending per-child.

It is good news that the "window of demographic opportunity" will be opened for decades in countries that is behind in the aging process and in the reduction of fertility rates. However, the correlation between overall social spending and child social spending is markedly weaker. Child spending seems to be quite low in regions undergoing such stages of the demographic transition. Overall social protection spending does increase with higher GDP, but that is mainly due to elderly spending rather than spending on children.

**Fig. 10.2** Public expenditure on child benefits by region, and proportion of children aged 0–14 in total population, 2010/11 (percentage of GDP) (*Source* Prepared by the authors based on International Labour Organization [ILO], World Social Protection Report 2014/15 Building economic recovery, inclusive development and social justice, Geneva, ILO, p. 14)

In fact, while Western Europe does show a strong positive difference in terms of children's spending even when their proportion of children is low, in the rest of the world's regions spending seems to have either no relation with the proportion of children, or even be inversely-related (Fig. 10.2) (Filgueira & Rossel, 2017).

Such a pattern is highly problematic. Poorer countries systematically under-invest in children's welfare and in doing so they risk losing the possibility of harvesting the demographic bonus during the window of opportunity that will come when they lower their fertility rates, since their adult cohorts will be less productive than if investment had been robust during their childhood years. Furthermore, since poorer countries are also on average highly unequal it is quite likely that such under-investment in family policies will lead to weaker and less efficient forms of family arrangements in lower income groups since the material basis for stable partnerships will be lacking,2 thus, missing on the gains that could come from such cooperation among men and women or in same sex marriages and partnerships.

<sup>2</sup>Lower income families and in general lower income people show earlier entry into marriage or union, and childbearing and lower stability in family arrangements, thus increasing lone parent households.

## **Family Policy in Europe**

Although with significant variations between countries, Europe is the most demographically advanced region, with low fertility rates and an aging population, as well as with high female labor force participation rates (for more details see Chapter 9 by Adema, Clarke, & Thévenon in this volume).

The fiscal effort on family policy in Europe varies both in magnitude and composition. While the Nordic countries, France, Great Britain, and some Eastern European countries report levels of spending are above 3% of GDP, southern Mediterranean countries show levels below 2% of GDP. Composition also varies among high spenders. Nordic countries systematically show high spending in childcare services, while others such as Ireland and the UK focus more on cash transfers. In some countries—as in France and Germany—tax breaks are also quite important (Adema, Clarke, & Thévenon in this volume)

Family allowances, tax credits, and other forms of cash transfers to families with children constitute a major part of the European system of social protection for children and families. All countries in Europe have some form of family allowance anchored in national legislation (ILO, 2014). Eligibility for the most part is wide, either through social insurance mechanisms or through universal non-contributory systems. In many countries in addition to social security or even universal benefits additional targeted benefits for poor, needy, and single parent—usually female—headed families are in place.

Maternity leaves have been in place in Europe for a long time and they are usually linked to social insurance schemes (ILO, 2014). Leave designs vary in terms of duration, type of benefit, flexibility, entitlement and conditions for returning to work. In the Nordic countries (like Norway or Sweden) the trend has been toward relatively short maternity leaves combined with relatively long parental leave (Allewell & Pull, 2003; Bruning & Plantenga, 1999; OECD, 2011). The possibility for both parents to use leave and the establishment of "daddy quotas" which can only be used by fathers are important features in the Nordic design of leaves (Moss & O'Brien, 2006; Ray, Gornick, & Schmitt, 2008; Adema, Clarke, & Thévenon in this volume).

Early childhood care services are quite developed in large parts of Europe. The leaders in terms of coverage and quality, both historically and today, are the Northern European countries, especially the Nordic countries (around 50% or more, with the exception of Finland), though coverage rates for 0–2 years old are also very high in France (Fig. 10.3). Among the Southern Mediterranean countries, while Portugal and Spain were laggards in the

**Fig. 10.3** Europe: Participation rates in childcare and pre-school services for 0-to-2 year-olds, around 2006 and 2014 (*Source* Prepared by the authors based on OECD family database Chart PF3.2.A)

early 1990s they have caught up and show coverage rates above the European Union (EU) average in the 2000s. Greece, Italy, and many Eastern European countries show the lowest coverage rates. Denmark, Iceland, and Norway, together with Malta and Luxembourg also report smaller gaps in their coverage rates according to socioeconomic level (OECD, 2019b). In contrast, despite the high average rates of coverage in Belgium, the Netherlands, and France the differences in coverage between the lowest tercile and highest tercile are stark. This is also the case in some low coverage countries, though in others coverage is low across socioeconomic levels (OECD, 2019b). Regarding children at older ages—typically between 3 and 4 years to 5 years old—coverage has become almost universal in most Western and Eastern European countries, going from an average gross enrolment ratio of 75% to more than 90% (OECD, 2019b).

Still, the variations in Europe in the three different policies that have been depicted have clear implications in terms of child poverty and how such rates compare to those of the general population. The Nordic countries present generous and universal family allowances—in addition to other cash transfers for vulnerable families, by far the most developed systems of family (maternity, parental and paternity) leaves and widely available ECEC systems (Finland is the exception in use, though not in availability3). The result is the lowest rates of child poverty, and usually rates that are below to those of the general population (OECD, 2019a).

In contrast, Southern Mediterranean countries have weak family and child protection systems. Their weaknesses are not homogeneous. For example, Portugal has a rather robust leave system in design, yet informality makes coverage weaker. Spain has developed in the last years an extended ECEC system, but has one of the least generous family allowance systems in Europe. Greece, despite the recent expansion of leaves, remains low in most of the other policy arenas. Italy is a below average performer in the three policy arenas considered here. Child poverty is among the highest in Europe, and is always above the poverty rates of the total population.

# **Family Policy in Other Developed Regions**

## **North America (Canada and the US)**

The United States and Canada present rather different demographic profiles. The US remains a young nation among developed ones and has relatively high fertility rates (Total Fertility Rate [TFR] 1.76 in 2017, World Bank, 2020a) and medium labor force participation rates by women aged 15–64 (55.7% in 2019, World Bank, 2020b). Canada presents significantly lower fertility rates (TFR 1.49 in 2017, World Bank, 2020a) and higher female labor participation (60.65% in 2019, World Bank, 2020b).

These two countries have very different profiles in terms of fiscal efforts and policies on families and children. The US presents one of the lowest fiscal efforts compared to other OECD countries regarding public expenditure (1.2% of GDP) and a large part of that effort comes in the form of tax breaks, not direct cash (Fig. 10.4). Yet, the US presents wide variation in terms of family policy at the state level (see Chapter 18 by Parolin & Daiger von Gleichen, and Chapter 17 by Engeman in this volume). Canada's effort is not much higher, but a large part of such effort is directed toward cash transfers in the form of family allowances and maternal and parental leaves.

Most of the US's cash transfers to families with children come in the form of tax breaks or special cash transfers for needy or poor families. While the system in Canada is also means-tested on family income, its coverage is quite large and it is far more generous than in the United States. The US does

<sup>3</sup>This is partly explained by the extremely generous leave scheme in Finland where the combination of benefits maternal, paternal and parental) allows for almost two years of total leave after birth.

**Fig. 10.4** CANADA AND UNITED STATES: Public expenditure on family benefits by type of expenditure, in per cent of GDP, around 2013 (*Source* Prepared by the authors based on OECD Family Database, Chart PF1.1.A. Public spending on family benefits)

not have a family allowance system, while Canada has a well-developed one.4 Also, the US stands out as one of the few countries in the world with no national legislation defining a mandatory paid maternity leave or parental leaves, although the federal law guarantees job protection for 12 weeks of unpaid maternity leave. Yet even such a narrow entitlement is hollowed out by the exemption for employers with less than 50 employees to comply with the norm. Canada has had maternity leaves anchored in social security starting in the 1970s, and during the nineties has introduced and increased parental leave scheme (from paid maternity leave for 17 weeks in 1990 52 weeks of potential paid maternity leave including both maternity and parental leaves

<sup>4</sup>Created in 1944 with the family allowances act, the Canadian system continues to reflect its origin as a universal family allowance program, with subsequent modifications to provide more adequate and targeted benefits for low income families. These changes have resulted in an increase in benefits for families with moderate income and a reduction or ending of benefits to higher income families. The US program was passed quite later and enacted as the Earned Income Tax Credit (EITC), which was part of the Tax Reduction Act of 1975. The credit was designed to help the working poor-families with income below the poverty level despite having working family members. It was initially authorized for only one year. However, the Revenue Adjustment Act of 1975 extended the EITC through the 1976 tax year. This seemed to set a precedent and each year the credit became a part of tax provisions that extended its authorization. It wasn't until the Revenue Act of 1978 that the credit became permanent. In 1986, the Tax Reform Act indexed the credit amount and the phase out levels for inflation. The US program has remained true to its origin, mainly as a tax break or credit targeted at the poor population. One of the disadvantages of this reliance on yearly tax breaks rather than monthly payments is that families living close or under the poverty line enter and exit such a situation throughout the year not on a yearly basis. Thus a monthly payment constitutes an income floor better adapted to the fluctuations in the vulnerability of families that is better fitted to the real life of families.

in 2016). In real terms, replacement rates for maternity leave and parental leaves hover around 50% of past earnings on average.

The United States (0.6% of GDP) and Canada (0.2% of GDP) have relatively low public spending in ECEC. Private services and informal care dominate the childcare sector in these countries, and the early education sector is confined to preprimary schooling for children 5–6 years. The negative consequences of this weak investment by the public authorities can be seen clearly in the services on offer, in particular, high costs to parents leading to unequal access and the segregation of children according to income. Low investment defeats a major purpose of these services and leads inevitably to low quality of services, lack of sustainability, and child care shortages. Unavailability of services also raises barriers against women's full-time employment and channels women toward low-paid, part-time jobs (Immervoll & Barber, 2006).

Free of charge or subsidized ECEC is not widely available and show variation between states and provinces. Coverage is among the lowest across OECD countries, though not as low as could be expected given its low fiscal effort. Still they are below the EU average. According to UNESCO data, gross enrolment rates in the US for preprimary school children aged 3–5 barely reach 70% and in Canada are below 75%. The average for developed nations is close to 87%.

In sum, Canada and the US are different, but both share rather underdeveloped systems of ECEC. In terms of family allowances and leave policies, Canada outperforms the US clearly, even though the fiscal effort of Canada in terms of cash benefits remains one half percentage point below the EU average as a proportion of GDP. Poverty levels are thus different, but both countries present child poverty rates above the OECD average and higher child poverty than poverty in the general population.

## **Oceania (Australia and New Zealand)**

In demographic terms, Australia and New Zealand are advanced nations. Despite having mostly finished their first demographic transition, they have never reached the lowest-low fertility scenarios that can be seen in many European countries. The TFR in Australia hover around 1.8 for the latest years, while the TFR in New Zealand reaches lows of 1.8 and highs of 2.2, quite above replacement rates. The average TFR for the EU is 1.5. Likewise, while aging is a clear trend, they remain relatively "young" compared to other countries of similar GDP per-capita. Women's labor force participation rates are high, and in many cases higher than European countries.

**Fig. 10.5** OCEANIA (OECD): Public expenditure on family benefits by type of expenditure, percentage of GDP, around 2013 (*Source* Prepared by the authors based on OECD Family Database, Chart PF1.1.A. Public spending on family benefits)

Both Australia and New Zealand present relatively high levels of spending on child benefits concentrated on direct cash transfer and to a lesser extent services, while tax breaks for families with children are non-existent (Fig. 10.5).

Both countries have systems of family allowances that were inspired by principles of universality,5 but subsequently became means-tested to increase payments for lower income families. However, they remain high coverage systems that neither rely on narrow definitions of need, nor on contributory formal employment. In Australia, almost all families with children can access family allowance and additional payments for child support.

In the case of Australia, even though cash benefits are named as tax credits, they are given as direct monthly payment or as a yearly lump sum.6 In the case of New Zealand, targeting is somewhat more stringent and coverage seems to be a contested issue, as the failed attempt in 2014 to make benefits flat rate and universal demonstrates.

Consistent with their liberal tradition, Australia and New Zealand did not have systems of maternity leaves well into the twentieth century. As in the US, these countries had some form of non-paid maternity leave, but it was

<sup>5</sup>In the case of New Zealand, it was originally means-tested but as early as the 1940s it was reformed as a universal system.

<sup>6</sup>This form of providing benefits solves the problem of low income families losing eligibility, since tax records and tax abatement are not possible or more complicated to achieve among this population.

only in the late twentieth century and early twenty-first century that they developed paid maternity and parental leaves. Australia would only introduce paid maternity and parental leaves in the year 2011 with a total length of 18 weeks. New Zealand did it somewhat earlier, in 2003, allowing for 12 weeks of maternity leave. They would increase the length of paid maternity leave from that year onwards reaching a total of 18 weeks by 2016. No paid parental or paternity leaves are available in New Zealand while Australia introduced two weeks of paid paternity leave as an optional quota of parental leaves by 2013. Both countries have around 40 weeks of non-paid but job protected parental leaves.

The replacement rates vary by income level. For mothers—or fathers in the case of Australia—with earnings that are half of average earnings, replacement is slightly below full replacement, but as income goes up replacement level drop to less than 50% of past earnings.

There is no data on take up of leaves provided by the OECD, but the wide coverage in social security implies wide access to such benefits. According to a recent report (Growing Up in New Zealand, 2014) as many as 84% of mothers took a combination of paid maternity, non-paid maternity, and annual paid leave at the time of birth. Yet it is true that quite low replacement rates might make upper income women less likely to take such leaves.

ECEC have a relatively more recent development in Australia and New Zealand than most of Europe, but in contrast to the US and Canada, fiscal efforts are more robust. When looking at overall coverage at very early ages (0–2 years old) there is a clear upward trend since the 1990s, reaching and overall coverage of more than 40% in New Zealand and around 30% in Australia (Fig. 10.6).

Coverage in preprimary school in Oceania is high with almost universal coverage for 4 and 5 year-olds, but not for those aged three. For the latter group, coverage is relatively high in New Zealand, but rather low in Australia.

In sum, Oceania presents an intermediate position regarding child and family policies when compared with Europe and the North American countries. While laggards in ECEC and leave policies, and with relatively narrow coverage in ECEC and limited extension in leaves, they have a strong, almost universal family allowance system that is also quite generous (more so in Australia than New Zealand). The impact on child poverty and overall inequality is positive, making them countries with average levels of poverty when compared to other similarly developed countries, and levels that are either similar or slightly higher than those of the general population.

**Fig. 10.6** OCEANIA (OECD): Enrollment in child care services (children between 0 and 2 years old), 1995–2014 (*Source* Prepared by the authors based on OECD Family Database 2017, PF3.2: Enrollment in childcare and pre-school)

## **Family Policy in Developing Regions**

#### **Asia**

Asian countries are very heterogeneous regarding fertility, dependency rates, and life expectancy. East Asian countries, for example, perform similar to other developing regions, with a relatively low fertility (1.6 children per women in 2010–2015) and an important decline in dependency ratios in the last decades (from 71.7 in 1980 to 41.4 in 2011) (Cecchini, 2014). Some East Asian countries, such as Japan and South Korea have levels of fertility that approach or reach very low levels, with TFR of 1.4 and 1.3 respectively, while China presents a TFR of 1.6 (for the period 2010–2015) (United Nations Department of Economic and Social Affairs, 2015). By contrast, other Asian sub-regions (Southern and Central Asian countries) present fertility rates that are above 2 and in western Asia rates are almost around 3 children per women (United Nations, Department of Economic and Social Affairs, 2015).

Targeted cash benefits are used most often in Asia as instrument of social protection, established largely by central governments. The traditional child/family benefits are not extensively used, however. For example, only one-third of the countries (16 out of 48) provide any kind of child or family allowance, the smallest proportion of all the regions (ILO, 2014).

The two most developed Asian countries with information from the OECD (Korea and Japan) show moderate spending efforts with quite different profiles: strong on child care services in Korea, more concentrated on cash transfers in the case of Japan (Fig. 10.7). Still, both of these countries have notoriously increased child and family protection policies in the last years partly as a response to lowest fertility scenarios and limited women's labor force participation.

The existence of family allowances in Asian countries, with the exception of Korea and Japan, is rare. Family allowances are only available in Sri Lanka and payable to the contributing worker rather than to families (Mokomane, 2012), but several countries have adopted other forms of cash transfers targeting families with children. There are, however, important variations across sub-regions. While in Central Asian countries have cash transfers programs with very limited coverage (Gassmann, 2011), several East Asian countries have cash transfers programs targeting families with children, most of them conditioned to school attendance or children's health check-ups. This is the case of Cambodia, Indonesia, and the Philippines (Cecchini, 2014; Kohler, Cali, & Stirbu, 2009). In 2012, Indonesia's CCT, Program Keluarga Hatapan (PKH), was covering around 1.5 million households (around 2.5% of the country's total households), while the Philipinian CCT, Pantawid Pilipino Program, was covering 3 million households (around 15% of total households) (Cecchini, 2014). Evaluations show some of these programs are being effective, among other things, in increasing enrolment among younger children (3–11 years old) and increasing attendance among 6–17 years old (Chaudhury, Friedman, & Onishi, 2013).

**Fig. 10.7** ASIA (OECD): Public expenditure on family benefits by type of expenditure, in per cent of GDP, 2013 (*Source* Authors' elaboration based on OECD Family Database, Chart PF1.1.A. Public spending on family benefits)

Although still far from the European parameters, a few Asian countries offer relatively generous maternity leave policies and several countries have maternity leave with a duration according to or near ILO's recommendation. After reforming its system in 2013, Vietnam stands out as the most generous country regarding maternity leave, with 6 months (and the possibility of an extra month if the mother has more than one child) of leave for mother. By contrast, in Bhutan and Maldives there is no maternity leave anchored to national legislation.

Some Asian countries (for example, Korea, Japan, Taiwan) have also introduced parental leave policies for relatively long periods and with some kind of replacement of salaries (Chin, 2012). Considering only Korea and Japan, Korea provides a more generous scheme in salary replacement through maternity leaves, while Japan offers more paid coverage in the parental leave scheme than Korea.

Several Asian countries are developing policies to increase accessibility to childcare services by increasing the number of childcare facilities, as well as increasing the number of children who are eligible for childcare subsidies (Chin, 2012). In Korea, between 2001 and 2015 the proportion of children in those ages that were enrolled in childcare services increased from 3 to 34%, while in Japan it increased from 22% in 2007 to 30% in 2014 (OECD, 2019c).

#### **Africa**

African countries have the highest fertility rates in the world, reaching an average of 4.7 children per women, and in some sub-regions (middle Africa, for example) reaching almost 6 children per woman (United Nations, Department of Economic and Social Affairs, 2015).

In the last decades, however, North African countries show a declining trend in fertility, combined with a cultural transformation in marriage and childbearing practices (Roudi-Fahimi & Mederios Kent, 2007). Also in the last decades, there has been an increase in women's participation in non-agricultural employment (Mokomane, 2012). Despite these trends, in comparative terms African countries are regions where the family—along strong traditional patriarchal lines—is a long-established institution and also the main provider of welfare for individuals in times of crisis or while facing risks such as unemployment, sickness, or old age (Mokomane, 2012).

As in other developing regions, family allowances in Africa are an exception and cash benefits for families have adopted the form of cash transfers to reduce poverty (Adato & Hoddinott, 2007). There are several cash transfer programs in different countries. In some cases, they are conditioned to certain behaviors (such as sending children to school or vaccination). Some examples are Burkina Faso's CCT for Orphans and Vulnerable Children, Ghana's Livelihood Empowerment against Poverty (LEAP), Nigeria's In Care of the People (COPE), Tanzania's Community-Based CCT (CB-CCT), and Senegal's Conditional Cash Transfer for Orphans and Vulnerable Children. The maintenance grant in Namibia, for example, is oriented to guarantee maintenance of children with disabilities and establishes a USD 26 of monthly cash transfer for the first child plus USD 13 per month for every additional child, up to 6 children (Mokomane, 2012). Although some CCTs in Africa are implemented nationally and funded mainly by governments, some CCTs are the result of partnerships between governments and international donors like the World Bank, UNICEF, the Department for International Development (DFID) in the UK, and other governmental and nongovernmental organizations. Also, African CCTs are frequently implemented in regions and relying on community participation (Southern African Social Protection Experts Network, 2016).

Unconditional transfers have a long history in the region and have their roots in the South African pension system introduced in the 1920s to protect basically the minority white population (Mokomane, 2011). Indeed, the earliest unconditional cash programs in Sub-Saharan Africa were old age pensions established in South Africa (1928), Namibia (1949), and Mauritius (1958) (Mokomane, 2012). However, with the exception of Mauritius and South Africa, they do not specifically target families with children (Niño-Zarazua, Barrientos, Hulme, & Hickey, 2010). The most relevant policy here is the South African *Child support grant*, which has proven to have significant impacts on children's nutrition and height (Agüero, Carter, & Woolard, 2006, 2007).

Maternity leaves in Africa are mainly of the contributory type (this is, part of the basic social insurance scheme), and apply to a minority of salaried workers (Mokomane, 2012). Therefore, although maternity leave is comprehensively available in most African countries, this benefit applies only to a minority covered by social security as salaried workers. Also, only a few countries adopted ILO's recommended of 14 weeks or more (Mokomane, 2011) (Fig. 10.8).

Only three countries (Mauritius, Uganda, and Tanzania) recognize paternity leave formally. In the rest of the countries, there is either a special multi-purpose leave provision which could potentially be used by fathers as paternity leave (Mokomane, 2011) or no legislation at all.

**Fig. 10.8** AFRICA (selected countries): Length of maternity leave (in weeks), circa 2014 (*Source* Prepared by the authors based on International Labour Organization (ILO), World Social Protection Report 2014/15 Building economic recovery, inclusive development and social justice, Geneva, ILO)

#### **Middle East**

In Middle Eastern countries, women's participation in the labor force participation is low and the development of family policy is limited (O'Brien, 2012). In fact, although there is variation across countries, social policy for families has "tended to reflect and reinforce the patriarchal gender contract" (Moghadam, 2005, p. 38).

While in these countries families remain as the key provider for care and economic support, there have been some developments regarding family policies. In the last few decades, several countries have established noncontributory benefits. These are mainly cash transfer programs, some of them targeted to families that have lost their breadwinner or do not have one (e.g. orphans, widows, divorced women, or even women who are single after a certain age). However, there is still relatively low coverage of poor and vulnerable working families with children (Machado, Bilo, Veras Soares, & Guerreiro Osorio, 2018).

In contrast to what happens in other regions such as Latin America, most cash transfer programs are unconditional and frequently followed by in-kind transfers, such as food distribution and fuel subsidies (Machado et al., 2018)

Compared to African countries, Middle Eastern countries seem to be more advanced in the development of preprimary education. As shown in Fig. 10.9, enrolment in pre-primary education in some of these countries shows a growth trend.

Although there are no available statistics of enrolment in children under 2 years old, the coverage is expected to be much lower. Also, in several countries, employers are obliged to provide childcare facilities dependent on the number of female employees (O'Brien, 2012).

The development of work-related leaves in the region is still incipient and mostly relying on the benefits offered through public sector employment (Kabeer, Ashwini, & Ragui, 2019). However, some countries (Egypt and Dubai, for example) have recently introduced changes to their legislation to either reach the minimum international standard or give women working in the private sector the same rights of those enjoyed by public sector employees.

**Fig. 10.9** MIDDLE EAST (selected countries): Enrollment in pre-primary education (3–5 years old), 1999 and 2015 (*Source* Prepared by the authors based on data from UNESCO/UIS)

#### **Latin America**

Most Latin American countries have experienced a decrease in mortality and fertility rates, as well as an increase in life expectancy. As a result, the region's population is aging and households are becoming smaller. At the same time, changes in divorce and nuptial patterns (with an increase of people living together without being married and a decrease in formal family arrangements) translate directly into changes in the shape of families (Arriagada, 2004; Rico & Maldonado, 2011).

Also, the region has gone through the massive incorporation of women into the labor market. Increases in women's participation in the labor market contribute to a "care crisis", due to pressures that are being experienced by families and particularly by women because of the combination of paid work outside their homes and demands for care of children and other dependents (Rico, 2011). This crisis is due to the pattern of the unequal distribution of unpaid work between men and women, which are very stratified by socioeconomic variables (CEPAL, 2010; Espejo, Filgueira, & Rico, 2010).

The only two countries for which comparable OECD data on overall spending on children and family was available, Chile and Mexico present low spending levels in the international landscape. Compared to Mexico, Chile's effort is larger—similar to the cases of low investment in other OECD countries—and balanced between cash transfers and care services. Mexico's effort is lower than almost all other OECD countries. Even spending on cash transfers to children and families with children is quite low, despite the fact that it has one of the largest CCT programs in Latin America.

A recent study carried out at the Economic Commission for Latin America and the Caribbean (ECLAC) (Tromben & Podestá, 2019) attempts for the first time to apply the OECD methodology to estimate family spending in nine countries in the region, including Chile and Mexico. The results show that the region falls clearly behind OECD levels of spending regarding monetary transfers with an average of 0.4% of GDP against the OECD average of 1.2 and European Union average of 1.5 (this includes contributory and non-contributory direct transfers, estimated pay on maternity, paternity, and parental leaves and fiscal credits to families with children). Only Argentina reaches levels that are slightly below the OECD average with 1.1% of GDP. Honduras, Guatemala, Perú, and Costa Rica spend 0.2% or less of their GDP in these transfers while Chile, México, and the Dominican Republic spend between 0.4 and 0.7% of their GDP (Tromben & Podestá, 2019).

Most of Latin America did not have a system of family allowances such as the ones that in the post war period swept through most of Europe. Exceptions in this sense are Argentina, Uruguay, and Chile, which developed early systems of non-targeted yet contributory family allowances, similar to the systems anchored in social security in Europe. The lower rates of formality implied of course lower coverage rates. Other countries developed some form of contributory family allowance in the second half of the twentieth century, but this was usually restricted to small groups of workers (miners in Bolivia, state employees in Venezuela, and other categories around the region). Colombia presents a particular case where family allowances exist in the formal sector but targeted to low income formal workers with children. Even when some of these systems increased contributory coverage with time, the lack of a strong formal workforce hindered coverage, especially of the poorer sectors of society.

The region would have to wait for the wave of mean-tested conditional cash transfers programs for families with children to really reach part or most of the poor and those working in vulnerable sectors. Some programs are quite stringent and aim at covering the poorest of the poor. But many have moved beyond such narrow coverage and cover today an important proportion of families with children (Fig. 10.10). Most of them have some conditionality attached to eligibility, usually regarding children educational attendance and health check-ups and vaccination.

The value of conditional cash transfers also varies significantly across countries. Looking at the basic benefit for families with children such values go from close to USD 100 per-children to less than USD 20 (monthly transfers) (Cecchini & Atuesta, 2017). In some countries, such as Ecuador, the value of the benefit is for the whole family and does not increase with number of children. In other countries, such as Uruguay, the values increase at a slower rate by number of children. Most countries have a maximum limit in terms of eligibility for additional child transfers no matter if they have more children. Once they reach the ceiling no additional benefit is granted.

International organizations have played a significant role in promoting the expansion of CCTs in the region (Borges, 2018; Osorio Gonnet, 2019). Although investment in CCTs has grown since 1996 (Cecchini & Atuesta, 2017), the fact that many CCTs lack support in national laws, or when they have, lack criteria for funding and indexation, makes such systems more volatile than other social protection policies. Even if frequently they have gained legitimacy and have survived government changes, the recent economic downturn seems to be having a clear detrimental effect on the

**Fig. 10.10** Coverage of individuals in households targeted by CCTs, around 2015. (percentage of total population) (*Source* Simone Cecchini and Bernardo Atuesta (2017) 'Conditional Cash Transfer Programmes in Latin America and the Caribbean. Coverage and Investment Trends'. Social Policy Series 224. ECLAC)

fiscal priority of these programs in a number of countries. In sum, a question remains around the extent to which these programs are to become a stable feature of a rights-based welfare architecture.

All countries in the region have statutory maternity leave policies. However, there is variation in terms of length and most countries offer less than the 14 weeks established by the ILO. Honduras is probably the most emblematic case, but the Dominican Republic, Bolivia, El Salvador, Guatemala, Haiti, Mexico, and Nicaragua all provide 12 weeks or less. Only nine countries (Panama, Costa Rica, Brazil, Peru, Ecuador, Venezuela, Colombia, Chile and Uruguay) provide 14 weeks or more. Chile is a noteworthy case in the region, with a 24-week maternity leave approved in 2011.

Data on payment rates of maternity leaves is a bit more encouraging: most Latin American countries offer 100% replacement rates, although in some cases this doesn't necessarily apply to the entire period. But the most important deficit in Latin America's maternity and parental leaves is still basic coverage. In Latin America, maternity leaves are still limited or non-existent for particular sectors, even if formal, like domestic workers, subcontracted, and temporal workers. As a result, the proportion of employed women that actually use the maternity leave benefit is relatively low. Furthermore, given the high level of informality and the fact that almost no system includes informal workers most systems of maternity and parental leaves are restricted to a small proportion of working mothers.

Flexibility is another dimension in which Latin America presents limited progress, with scarce options for women on when the leave can be taken or how it can be combined with part-time jobs. An exception to this is Chile where the new post-natal leave allows women to use until 18 weeks and go back to work in a part-time scheme, until the child is 30 weeks old.

Parental leaves are really an exception in Latin America (OIT, ONU-Mujeres, & PNUD, 2012; Pautassi & Rico, 2011). To date, only three countries have parental schemes, included within the maternity leaves. Cuba offers a 9 months' unpaid maternity/paternity leave both to mothers or fathers after the statutory maternity leave. Chile offers mothers with the possibility to transfer the benefit to fathers after the 7th week after childbirth and for a maximum period of 3 months. In 2013, Uruguay instituted a shareable full paid parental leave that allows either parent to work half-days until the child is six months old (Rossel, Filgueira, & Rico, 2015).

According to ECLAC, in 2009 Latin America's net enrollment in care services for children between 0 to 3 years old was around 5% (in Guatemala, Honduras, Dominican Republic, Paraguay) and 20% (in Cuba and Mexico). The net enrollment rate for children between 3–6 years old was much higher, but only in Cuba and Mexico reaches levels that are near universal. Brazil, Uruguay, Argentina, Peru, Colombia, and Panama present enrollment levels of around 60 and 70% (CEPAL, 2011; Rossel et al., 2015).

A recent study published by ECLAC shows that enrollment in ages 3–5 has grown significantly in the last decade. In 2014, the average coverage for eight countries was 64.6%. The national coverage of ECEC services for 3–5 years reached 86% in Uruguay, 79% in Brazil, 77% in Chile, 72% in Mexico, 64% in Peru, and 61% in Colombia. Even countries with a very limited coverage in 2009, like Honduras, presented a modest growth in 2014 (37.6%). Also, enrollment in child care services is highly stratified by socioeconomic level (the high-income population is the one with higher access to services, while lower income sectors present significantly lower attendance rates) and by the urban/rural cleavage (with less coverage in rural areas) (CEPAL, 2017).

The design of child care services in Latin America is quite varied, combining public facilities with private services or services provided by civil society organizations, usually subsidized with state funding (Martínez Bordón & Soto de la Rosa, 2013; Vegas & Santibáñez, 2010). In the last decade, countries like Mexico, Chile, and Uruguay have expanded child care infrastructure, improving the availability of services for children between 0 and 3 years old (Staab, 2010).

In addition to varying coverage rates, the quality of ECEC services in the region is highly varied. The scarce evidence on the subject suggests a high heterogeneity and persistent precariousness in many countries, concentrated mostly in rural areas and in low socioeconomic contexts (Araujo, López Boo, & Puyana, 2013; Vegas & Santibáñez, 2010).

# **Conclusion**

Compared to pension systems, health care and health insurance, and formal education, family policies are relatively more recent, have a wider variety of institutional settings and usually carry less weight in the fiscal effort of states. Yet it is also true that these policies have grown significantly both in developed as in developing countries.

There are at least two reasons why this trend takes place. First, there is increasing scientific evidence regarding the importance and efficiency of early investment in children promoting their capabilities and protecting them from risks and vulnerability. While children have no vote, knowledge has helped making their case to a larger extent than before. Development will happen through human capabilities and human innovation. And increasingly the scientific community and the policy paradigm are agreeing that the foundation of such capabilities takes place between gestation and the first 4–5 years of life. Still, despite such increasingly positive and strong consensus, the efforts fall short in most cases leaving an important part of small children unprotected. Second, the increasing incorporation of women into the labor market and the relatively frozen landscape regarding male incorporation into domestic chores and care implies that at some point the state would need to enter into the equation organizing services and creating policies that allow for the balancing act of paid and unpaid work that women face day in and day out. Leaves and ECEC are very much led by family changes and demands. In some cases, this takes place without state involvement. Enterprises and markets provide the basis for leaves (through firm-related leave arrangements for families with a newborn) and early childhood care (through private childcare facilities). But such a solution is usually fragmented, insufficient and does not reach the large group of women who cannot access such market solutions, or who are not employed by enterprises granting time of work for family and child care. As shown by Adema, Clarke, and Thévenon in this volume (Chapter 9), in the case of OECD countries many of these policies were meant to avoid a very low fertility scenario. In many countries in Eastern Europe as well as in Japan and Korea major hikes to transfers and leaves have taken place with precisely this goal in mind. Despite a major economic malaise affecting the Southern Mediterranean countries, it is possible to see in these countries increasing efforts to strengthen family policies and childcare protection.

The recent experience from the leaders in family policies offers an opportunity for developing regions. Regarding family allowances and cash benefits, European countries have moved toward expanding the coverage but also the generosity of transfers. They have also created specific benefits for loneparents (but see Chapter 13 by Nieuwenhuis & Madonado in this volume). Regarding work leaves, European countries have expanded both length (way further than the 14 weeks recommended by ILO) and the generosity of paid benefits, sometimes merging maternity and paternity/parental leave schemes. Also, some countries have increased flexibility in leave policies, allowing for more workers to take them and use them for a longer period. "Daddy quotas" have become an increasingly common innovation in European leave systems. Regarding child care services, European countries have significantly expanded coverage both by increasing the number of hours' services are available and the population they are open to (for example, expanding the eligibility criteria to children under 2 years old). They have also installed the importance of encompassing coverage expansion with improvements in service quality, to guarantee that attendance to child care services translate into the expected beneficial outcomes in child development.

Laggards in leaves such as the liberal countries have advanced in creating a modest leave system with the exception of the United States, while they have also moved ahead in coverage and quality in ECEC. Regarding family allowances, the picture is more mixed. Some universal systems have become more targeted (not narrowly, but targeted nonetheless), but in general they have also protected or even increased value. More clear laggards were the Asian developed countries of Korea and Japan. These countries have made major strides in catching up with their western counterparts in ECEC and leaves. The rest of Asia remains underdeveloped in all three policy areas, though some leave systems have advanced and CCT programs do exist in a limited number of countries.

In developing regions, the last 15 years showed major advances in all the areas that we considered throughout this chapter. Monetary transfers to families with children, extended leaves for mothers and larger coverage in many cases, and albeit more modestly, early childhood care and education are part of the agenda and in some cases a policy reality. Yet what seems to be lacking in these regions are three critical aspects that are required for sustainability and effectiveness: (1) rarely are these policies based on the idea of a universal set of interrelated transfers and services; (2) overall, the fiscal effort in most of these policies remains modest and the economic downturn of the last years might have a very negative impact on them; (3) the impact of these policies on inequality and poverty so far has remained limited.

Family policy can play—if established generously and of reasonable quality—a major role in increasing social investment for society at large and in decreasing inequality and vulnerability for children and young families. They operate in a critical period where families are most vulnerable and when children are developing their basic cognitive functions. Of course, not all spending in children and families will get the job done. If family allowances and leaves remain limited to formal workers and if ECEC are of bad quality the promise will not be fulfilled. Universal good quality transfers and services are needed, or at least services and transfers that reach both the lower income groups and the middle classes so as to guarantee access to those most in need and reasonable quality.

Developing countries have a long road ahead if they want to transform what today is an embryo of a family and childcare protections system into a full-blown pillar of their social states. Fiscal costs will be important, but the long run benefits in human capabilities, productivity and equality far outweigh the fiscal costs of setting such system up.

**Acknowledgment** The main contents of the chapter are based on our previous publication Filgueira and Rossel (2017) *Social protection for families and early childhood through monetary transfers and care worldwide*, prepared for and funded by the Economic Commission for Latin America and the Caribbean Project "Social Protection System and Inequality".

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **11**

# **Gendered Tradeoffs**

**Jennifer L. Hook and Meiying Li**

The literature linking family policy to women's labor market outcomes is a rapidly expanding field, spanning many disciplines and analytic approaches. Several recent reviews provide a solid overview of the field and we refer readers to them (Ferragina, 2019; Hegewisch & Gornick, 2011; Olivetti & Petrongolo, 2017; Rubery & Figueiredo, 2018; Steiber & Haas, 2012). This chapter has an explicit focus on two central questions: (1) Do generous family policies perpetuate gender inequality? That is, do family policies have unintended consequences or promote women's inclusion into the labor force at the expense of gender equality in labor market outcomes, such as in working hours, occupational integration, or wages? and, (2) do generous family policies promote gender equality for certain groups of women at the expense of other groups of women? We also consider evidence regarding a weaker form of this question—do certain groups of women benefit more from family policies than others? In answering this question, we focus on how policies differentially affect women by social class as measured by educational attainment.

We begin by briefly reviewing the types of evidence brought to bear on these questions. We then turn to the primary questions that animate this review, discussing the literature on gender inequality-promoting effects of family policy and heterogeneity in effects of family policy by social class. We review the uneasy consensus that has emerged around how two of the most

J. L. Hook (B) · M. Li

© The Author(s) 2020

University of Southern California, Los Angeles, CA, USA e-mail: hook@usc.edu

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_11

studied family policies—childcare services and parental leave—affect women's employment outcomes in general and differentially by social class. Our review is heavily weighted toward questions that have emerged in the comparative study of women's employment over the last 15 years. For an excellent review of foundational research in the 1990s we refer readers to van der Lippe and van Dijk (2002).

# **How Do We Know?**

Researchers are generally interested in how variation in family policy is associated with women's labor force participation, employment, work hours, or wages. The recent literature linking family policy to women's labor market outcomes generally draws on individual-level data from anywhere from two to dozens of countries and seeks to explain variation across contexts (for pioneering works in the approach see Gornick, Meyers, & Ross, 1998; Rosenfeld, Van Buren, & Kalleberg, 1998). They tend to use cross-sectional data, although recent work exploits variation over time (Blau & Kahn, 2013; Christiansen, 2016; Cipollone, Patacchini, & Vallanti, 2014; Nieuwenhuis, 2014; Nieuwenhuis, Need, & Van der Kolk, 2017; Thévenon, 2009, 2016).

Two primary methodological approaches are used. In the first, researchers compare exemplars of welfare state regimes and assess whether differences across countries are congruent with expectations generated from a comparison of family policy contexts. Work in this vein often draws on Esping-Andersen's (1990, 1999) "three worlds" typology and his later conceptualization of de-familialization or feminist critiques thereof. Although typologies differ, as detailed in Chapter 6 by Zagel and Lohmann in this volume, the central focus is how the state supports (or does not support) the reconciliation of work and care. A leading conceptualization focuses on four strategies of welfare state support (Misra, Budig, & Moller, 2007) or familialization (Saraceno & Keck, 2008). Countries are grouped into categories: (1) the primary caregiver strategy or supported familialism, wherein the state subsidizes women's care work; (2) the primary earner strategy or familialism by default, wherein the state encourages women's employment but provides little support for care; (3) the choice strategy or optional familialism, wherein state policy supports women's choice to work or care; and (4) the earner-career strategy or de-familialist, wherein policy supports the combination of work and care for women as well as men.

In the second approach, researchers combine individual-level data with contextual-level family policy data in multilevel models. Most studies use hierarchical models with individuals nested within countries. Although these models have been used extensively in the literature, they are criticized in several ways: they are sensitive to the selection and number of countries included, which are typically less than thirty and often much less; data tend to be cross-sectional, thus causal claims are questionable; and many key findings in the literature are based on data from the 1990s, which are now dated (Brady, Blome, & Kmec, 2020; Steiber & Haas, 2012).

This second approach requires quantifying indicators of specific policies (e.g., weeks of paid parental leave) or closely linked outcomes of specific policies (e.g., enrollment rates in publicly funded child care programs). There is substantial controversy, however, over how these indicators should be used. One option is to measure family policy as a singular index representing a "broad phenomenon" (Mandel & Semyonov, 2005, p. 955). For example, Mandel and Shalev (2009) argue that family policies should be treated as an integrated bundle, writing that "while it appears that individual components of the family policy package may contribute to women's attainments, they also have shared effects that cannot be detached from one another, either theoretically or empirically" (p. 1879). Other scholars argue, however, that family policies should be delineated by whether they provide time to care or services to replace care (Kamerman & Kahn, 1994), that is whether policies are "work-reducing" or "work-facilitating" (Jacobs & Gerson, 2004). Pettit and Hook (2009) draw on this literature to argue that family policy "conflates a variety of specific policies that may have countervailing effects on women's employment and other economic outcomes" (p. 11). Between the two poles of conceptualizing family policy as an inseparable package versus as individual policies, lies an approach that groups family policies by the type of effect they are likely to have on families. For example, Korpi, Ferrarini, and Englund (2013) create indices to measure three dimensions of family policy: policies that support gendered family arrangements (i.e., encourage women to provide care at home), dual-earner arrangements (i.e., encourage women to work), and dual-carer arrangements (i.e., encourage men to care at home). This last approach most closely resembles the regime or typology approach.

An emerging third main approach is a hybrid, wherein the effects of singular policies are tested for policy or institutional "complementarities" (Thévenon, 2016). Arguing that the impact of family policy on women's employment is likely to be context-dependent, Thévenon examines complementarities between policy pairs (paired interaction terms), institutions (sum of the direct effects of all policies), as well as the effects of policies within regimes (policies interacted with regime dummies). He concludes that "female labour force participation reacts differently to different policy measures, depending on the institutional environment in which they play out" (p. 494).

Given the difficulty of establishing causality in macro-level studies, microlevel policy evaluation adds a considerable amount to the evidence base. Studies tend to exploit spatial and temporal variation in implementation of family policy reforms with one country. A review of the methodological issues is beyond scope of this chapter; we refer readers to an excellent discussion in Müller and Wrohlich (2018). Some of the evidence about social class differences in policy effects reviewed below is drawn from this literature. Although superior for establishing causality, these studies are country-specific, raising the question of whether the intervention effects would "travel" to other contexts. Conflicting findings from single country studies and evidence from macro-level analyses (e.g., Thévenon, 2016) suggest researchers use caution extrapolating findings to other countries.

# **Unintended Consequences, Welfare State Paradoxes, and Gendered Tradeoffs**

## **Do Generous Family Policies Perpetuate Gender Inequality?**

Mandel and Semyonov (2005, 2006) argue that although family policies bring women into the labor market, they create a paradox wherein they have adverse consequences for women's occupational integration and earnings. Family policies incorporate lower-skilled women who would otherwise not be in labor market into employment, which increases the gender gap in wages and occupational sex segregation among those in the labor market, including those who would be in the labor market irrespective of family policies. Specifically, they argue that generous family policy packages lower women's accumulated work experience, promote discrimination against women, and concentrate women in female-typed jobs in the public sector. They test this idea with cross-sectional data from the 1990s for 20 countries housed in the Luxembourg Income Study (LIS) database. They create a singular "welfare state index" composed of parental leave, childcare, and public sector employment. They find that generosity on the index predicts a larger gender pay gap (Mandel & Semyonov, 2005) and women's lower representation in managerial positions (Mandel & Semyonov, 2006) although their models suggest that the findings are largely driven by parental leave.

Rather than seeing gender inegalitarian consequences of all family policy, Pettit and Hook (2009) argue that each policy should be considered along two dimensions—whether it promotes labor market inclusion or exclusion and whether it discourages or promotes equality among those in the labor force. They examine exemplars of each combination: childcare (inclusion and equality), part-time work (inclusion and inequality), unionization (exclusion and equality), and parental leave (exclusion and inequality). This specification differs from Mandel and Semyonov (2005, 2006) in that policies that promote inclusion do not necessarily engender inequality; the key is how women are incorporated. Pettit and Hook (2009) posit that childcare includes women in the labor market and puts them on more equal footing with men, whereas lengthy parental leaves remove women from the workplace for several years, put them on less equal footing with men, and encourage employer discrimination. Thus, the effects of leave and childcare on women's labor market outcomes are not expected to reflect "broad phenomenon," but are potentially countervailing. They explore this idea using up to 63 crosssectional surveys, spanning the years 1969 to 2000, from 21 countries housed in the LIS database. They find support for the proposition that family policies can support both inclusion and equality, finding evidence that childcare supports both employment and occupational attainment, particularly among higher educated women.

Brady et al. (2020) methodologically critique these analyses and provide two analyses refuting the conclusions described above. They analyze LIS data both cross-sectionally (21 countries in the mid-2000s) as well as with a country panel using fixed effects models (12 countries, two time periods, mid-1980s/1990s and mid-2000s). Contrary to the existing literature they not only find no evidence of tradeoffs between labor market outcomes, but that "work-family policies are not associated with labor market outcomes for women or mothers. The vast majority of coefficients for work-family policies fail to reject the null hypothesis of no effects" (Brady et al., 2020, p. 142). They do not explicitly consider, however, differences by class or contextual effects, which are discussed below.

#### **Uneasy Consensus on Policy Effects**

The above discussion reveals that the literature is rife with differing conceptualizations of policy, approaches to analyses, and not surprisingly, empirical findings. That said, an uneasy consensus has emerged around publicly funded childcare and to a lesser extent, parental leave.

Childcare provision is associated with higher levels of employment (Christiansen, 2016; Erhel & Guergoat-Larivière, 2013; Jaumotte, 2003; Thévenon, 2016), particularly among mothers and mothers with young children (Boeckmann, Misra, & Budig, 2015; Nieuwenhuis, Need, & Van Der Kolk, 2012; Pettit & Hook, 2005, 2009; Steiber & Haas, 2009; Uunk, Kalmijn, & Muffels, 2005), including positive effects on mothers' working hours (Andringa, Nieuwenhuis, & Van Gerven, 2015; Boeckmann et al., 2015), preferences for longer working hours (Pollmann-Schult, 2016), and lower motherhood wage penalties (Misra, Budig, & Boeckmann, 2011). Childcare has also been associated with women's increased representation in professional occupations (Pettit & Hook, 2009). We describe the consensus as uneasy, however, as there is debate as to whether the employment-promoting effects of childcare provision are causal (Steiber & Haas, 2012), particularly when effects are assessed with enrollment data (for more on data challenges, see Chapter 24 by Sirén, Doctrinal, Van Lancker and Nieuwenhuis in this volume). Findings from Norway suggest that additional childcare places crowd out informal arrangements and benefit mothers who are already employed (Havnes & Mogstad, 2011). Positive effects of policy expansion, however, are found in other contexts such as Germany (Bauernschuster & Schlotter, 2015). Reviewing the literature on family policy, Olivetti and Petrongolo (2017) conclude that evidence supports, "overall positive effects of subsidized child care on female employment" but "no obvious consensus emerges from the literature that has studied the labor market impact of parental leave rights and benefits" (p. 227).

We argue that although the debate about parental leave is more contentious, an uneasy consensus does exist. Parental leave is generally considered employment supportive, especially if it is well paid (Boeckmann et al., 2015). There is evidence, however, that parental leave that is "too long" is associated with lower rates of employment among mothers. That is, there is a curvilinear inverted U-shape for mothers with young children (Pettit & Hook, 2005). How long is too long? Researchers generally find that leaves over two years depress rather than facilitate employment (Nieuwenhuis et al., 2017; Thévenon & Solaz, 2013), although some estimates approach three years (Christiansen, 2016; Pettit & Hook, 2009). Nieuwenhuis and colleagues (2017) conclude that "while short periods of leave can be useful, or even necessary, to maintain women's attachment to the labour market after becoming a mother, very long interruptions of employment indeed seem to be a 'mechanism of exclusion'" (p. 13). Some studies fail to find evidence that leave can be too long (Keck & Saraceno, 2013). Long leave, however, has also been linked to negative consequences in other labor market outcomes, including larger motherhood wage penalties (Misra et al., 2011), particularly for highly educated women (Pettit & Hook, 2009). Childcare and paid leave of modest duration are also supportive of single mothers' employment (Maldonado & Nieuwenhuis, 2015; Van Lancker, 2018).

Evidence from single country studies supports the above findings. Extensions of leave delay return to work (Lalive & Zweimüller, 2009 in Austria; Ziefle & Gangl, 2014 in Germany) and long leaves can promote discrimination against mothers (Glass & Fodor, 2011 in Hungary). Glass and Fodor (2011) contend that state policies that make motherhood salient for employers, in the absence of anti-discrimination enforcement, contribute to discrimination. Evidence from the Swedish context finds that controlling for selection, taking 16 months of leave or more has a negative effect on women's upward occupational mobility upon return to work (Evertsson & Duvander, 2011). We conclude that there is an uneasy consensus on the effects of family policy *in general*, but the consensus fractures when we consider differences by social class.

## **Gender-Class Tradeoffs**

#### **Conceptualizing Social Class**

Social class is a complex concept, encompassing both material and cultural accounts. The vast majority of quantitative analyses take a materialist perspective focusing on class divides in opportunities, such as the pay and quality of jobs, and constraints, such as the cost and quality of care (Crompton, 2006; McRae, 2003). Although sociologists tend to favor conceptualizations of class derived from occupational attainment, the vast majority of studies that speak to "class" utilize level of education attainment (Cooke, 2011; Grönlund & Magnusson, 2016; Korpi et al., 2013; Nieuwenhuis, 2014; Pettit & Hook, 2009; Steiber, Berghammer, & Haas, 2016). This arises from both theoretical and practical considerations. Theoretically, the study of women's employment is informed by micro-economic theories emphasizing the importance of human capital, of which education is a primary indicator. Practically, education circumvents the difficulties inherent in assigning non-employed women their partners' social class (see Sorensen, 1994). Educational attainment also tends to be the most readily available and comparable measure in commonly used data sources. Alternate measures used in the literature include "income classes" measured by earnings quintiles (Mandel & Shalev, 2009), and simplified occupational schemes (e.g., managers and professionals, white-collar, and blue-collar and elementary occupations) (Pavolini & Van Lancker, 2018).

Little work explicitly tests both constraint and cultural explanations for class gaps in women's employment or family policy use. In a recent contribution, Pavolini and Van Lancker (2018) use cross-sectional data from 27 EU countries to examine class gaps in formal childcare usage. They find greater support for materialist explanations; parents' perceptions of availability and affordability of care better explained the persistent class differential in usage than cultural norms about mothers' employment. The authors, however, did not test whether class-specific norms impacted class differentials in usage. Research finds class gaps in gender egalitarianism across welfare states (Shalev, 2008; Svallfors, 2006).

### **Do Generous Family Policies Promote Gender Equality for Certain Groups of Women at the Expense of Other Groups of Women, Particularly by Social Class?**

Shalev (2008) argues that women's normative *and* material interests diverge by class with respect to family policies. He argues that generous family policy limits high achieving women's ability to achieve gender equality in the labor market. In a series of studies, Mandel and colleagues find that "generous family policy" promotes gender equality for women at lower socioeconomic levels, whereas ungenerous policy promotes gender equality for more advantaged women (Mandel & Shalev, 2009; Mandel, 2011, 2012).

As noted earlier, however, effects depend on the type of family policy and these, too, vary by class. Pettit and Hook (2009) find that publicly supported childcare supports higher educated women's employment and occupational attainment, more so than for other women, with no negative effects on wages. This is consistent with research showing that childcare services are utilized more by high-income families than low-income families (Van Lancker, 2013) and childcare expansion benefits the already employed, crowding out informal care (Havnes & Mogstad, 2011). Hook and Pettit (2016) argue that replacing informal with formal care is likely to benefit employed women's labor market outcomes because formal care is less likely to break down than informal arrangements, which puts mothers on a more equal footing with their colleagues. When looking instead at high rates of part-time employment and lengthy parental leave, Pettit and Hook's (2009) conclusions are consistent with Mandel and colleagues (ibid.), as these policies dampen the wage premiums for highly educated women.

Korpi et al. (2013) approach the question of tradeoffs among groups of women by examining policy dimensions. They analyze LIS data from around 2000 for 15 countries and categorize family policy on three dimensions: *traditional family* (child allowances, part-time daycare services, home care allowances, marriage subsidies), *dual*-*earner* (daycare services for 0–2, full-time daycare services for over-threes, earnings-related parental leave), and *dual*-*carer* (paid leave, paid leave reserved for fathers). They combine dual-earner and dual-carer indicators because they are highly correlated into an earner-carer dimension. They find that earner-carer policies promote the employment of lower educated women "without resulting in significant diminution in women's access to top wages and positions of power" (Korpi et al., 2013, p. 28). They conclude that publicly provided childcare and "judiciously designed parental leave" benefit mothers without creating disadvantages for tertiary educated mothers, and that earner-carer policies can "promote gender egalitarianism and class equality simultaneously" (p. 30).

#### **No Consensus on Class Differences**

In contrast to the *relative* consensus on family policy impacts on maternal employment, there are conflicting theoretical mechanisms and empirical findings about how the effects of family policies on women's employment vary by women's social class, primarily measured by educational attainment. On the one hand, some argue that highly educated women tend to be in the labor force regardless of motherhood and policy context; it is women with lower levels of education who are more sensitive to variations in context (Del Boca, Pasqua, & Pronzato, 2009; Korpi et al., 2013). Essentially, workfacilitating contexts sway only women who experience less monetary and intrinsic rewards to employment (Damaske, 2011; Steiber et al., 2016). This is consistent with the mechanism of selection as hypothesized in the welfare state paradox (Mandel, 2011; Mandel & Semyonov, 2006).

On the other hand, some researchers argue that lower educated women suffer from such low job quality that reconciliation policies fail to be incentivizing, thus more highly educated women are more responsive to family policy conditions (Cipollone et al., 2014). This is consistent with findings that higher levels of publicly provided childcare promote employment more among highly educated than other women, incentivizing women with the most to gain from employment (Pettit & Hook, 2009). Similarly, Nieuwenhuis (2014) found that paid leave reduces the motherhood employment gap more so among highly educated women than among the less educated. Evidence from Germany found that a legal claim to subsidized childcare and the expansion of childcare had positive effects on women's employment, but the effect was weakest for women with lower levels of educational attainment (Boll & Lagemann, 2019; Müller & Wrohlich, 2018).

Finally, other researchers find little evidence of differential effects by educational attainment. Keck and Saraceno (2013) hypothesize that lower educated mothers will be more disadvantaged by very short or very long leaves and by high childcare costs. They find, however, that employment penalties linked to education occur independently from (or before) motherhood and family policies (Keck & Saraceno, 2013).

One potential explanation for discordant findings is that the importance of educational cleavages varies across policy regimes. Hook and Pettit (2016) develop expectations about how family policy regimes are connected to occupational segregation of mothers from childless women and how this varies by women's educational attainment. They find that where conditions enable mothers to devote substantial time to caregiving, through extended parental leaves, for example, occupational inequalities cluster around motherhood. In countries with meager provision for reconciling work and family, occupational inequalities diverge acutely by educational attainment. Finally, in countries where conditions enable mothers to compete in the paid labor force, by providing childcare services, for example, there is less differentiation by motherhood and by educational attainment.

# **Recent Possibilities**

The gendered tradeoffs literature is an ongoing debate. Promising directions consider how the impact of family policies on women's employment outcomes varies across contexts. For example, cultural norms may moderate the impact of family policies on women's employment outcomes (Kremer, 2007; Pfau-Effinger, 2005). Evidence generally supports that egalitarian gender norms are positively associated with women's employment (Boeckmann et al., 2015; Uunk, 2015; although see Uunk & Lersch, 2019), although gender norms do not account for family policy effects on women's employment (Uunk et al., 2005). Egalitarian gender norms, however, may moderate the relationship between policy and women's employment outcomes. Budig, Misra, and Boeckmann (2012) examine how gender culture moderates the relationship between family policy and mothers' earnings. They hypothesize that family policies that narrow the motherhood earnings penalty will have a stronger effect in cultural contexts that are supportive of maternal employment. They find support for this argument. They conclude, that "cultural attitudes amplify, and even change the nature of, associations between parental leave, publicly funded childcare, and maternal earnings" (Budig et al., 2012, p. 186).

Another potential source of contextual effects comes from the labor market context. Estévez-Abe (2006, 2009) argues that in coordinated market economies (CMEs), characterized by long term employment contracts, strong employment protections, and investments in firm-specific training, women are at greater risk of becoming "labor market outsiders" due to breaks in employment. Family policies that encourage breaks in employment may have greater negative effects in CMEs than in liberal market economies (LMEs). Grönlund and Magnusson (2016) connect this work explicitly to the welfare state paradox debate by examining the gender wage gap by skill level in a three country comparison of two CMEs that differ in family policy (Sweden and Germany) and one LME (the UK). They hypothesize that the gender wage gap will be greatest among high-skilled women in Sweden (a CME with dualearner family policies) and that sex segregation and on-the-job training would explain more of the gender wage gap for this group than for others. They do not find support for this proposition. The unadjusted gender wage gap is lowest in Sweden and high-skilled women do not face a larger gender wage gap than women in the other two countries. Furthermore, gender wage gaps among the high skilled in CMEs are not well explained by the hypothesized mechanisms.

Another avenue for contextual consideration of tradeoffs by social class is income inequality. Hook (2015) argues that it is difficult to derive hypotheses about women's employment decisions in different policy contexts without considering both the overall level of income inequality in a country and a woman's social location within it. She contends that in the absence of the state de-familializing care, "there are two central questions: (1) what are the alternatives to family care? and (2) how do these alternatives vary by a family's social location? Available alternatives will depend, in part, on the level of inequality in each country" (Hook, 2015, p. 17). She develops expectations about differences by social class under different regimes of familialization (Saraceno & Keck, 2008). She finds that in countries with high-income inequality and low de-familialization or optional familialism, families polarize between dual full-time and male breadwinner families, and do so largely by women's educational attainment. She concludes that while class is an important correlate of work-family arrangements in nearly all contexts, it is of utmost importance in countries with higher income inequality, greater reliance on the market, and optional familialism. These arguments have yet to be examined outside of a regime approach.

# **Conclusion**

In sum, there is an uneasy consensus around the relationship between family policy and women's labor market attachment and outcomes, and no consensus around class differences in policy effects. Regarding the former, publicly funded childcare promotes mothers' employment with no discernable negative consequences for gender equality in other labor market outcomes, such as the gender wage gap. There are no observed deleterious effects of childcare on any group of women, but there is an open debate on who benefits more or less from childcare provision or expansion. Parental leave is also a valuable tool for promoting mothers' labor market attachment. Leaves that are "too long," however, can depress women's labor market attachment resulting in inegalitarian consequences for other labor market outcomes, such as sex segregation and the gender wage gap. There is no consensus around how these relationships vary by women's social class, although there is some evidence that long leaves harm highly educated women's labor market outcomes the most (consistent with welfare state paradox or tradeoff arguments).

To move this debate forward we need more research that considers multiple labor market outcomes, which is at the crux of the welfare state paradox or tradeoff arguments. To resolve this debate, studies should engage both employment and other labor market outcomes, including intensity, segregation, and wages. We also need better data on policy and innovative ways to include it in our models (see Chapter 24 by Doctrinal et al. in this volume). Common measures of paid parental leave, for example, obscure nuances in replacement rates, caps on replacement rates, eligibility criteria, and flexibility in use. These aspects of policy design should affect the labor market consequences of these policies as well as heterogeneity within countries in the benefits different groups of women receive (Bártová & Emery, 2018; Ray, Gornick, & Schmitt, 2010). Our current conclusions in this area are based on less than ideal policy data.

Two key themes emerge from the current research. First, the effects of family policy on women's labor market outcomes are likely to depend on the context in which they are implemented. Recent research provides guidance on how to move forward exploring these complex interactions (e.g., Thévenon, 2016). Second, the effects of policy change are likely to depend on woman's social location. Given that there is no consensus yet on how there are ample opportunities for researchers to move the field forward. We note two important opportunities here.

First, we need more expansive thinking about context dependence of family policy effects and *in relation to* heterogeneity of effects within countries. The sparse literature on context dependence has produced important insights but remains under-developed. Furthermore, the strands of the literature on context dependence and heterogeneity in effects within countries have been largely separate, but could produce considerable insight if aligned.

Second, another key opportunity for future research is to consider other aspects of mothers' social location. Although educational attainment dominates cross-national studies of women's employment, single country studies document differences in maternal employment and family policy uptake by race, ethnicity, and nativity (Kil, Neels, Wood, & de Valk, 2018; Kil, Wood, & Neels, 2018; Vidal-Coso, 2019). Given the challenges inherent in incorporating country-specific patterns and meanings of race, ethnicity, and nativity in a cross-national design, most work in this area is in the form of single country studies or in two-country comparisons (e.g., Pettit & Hook, 2009). Expanding cross-national work in this area is important for understanding heterogeneity in policy effects as well as inequality more broadly.

**Acknowledgements** We gratefully acknowledge Lynn Prince Cooke, Laure Doctrinal, Sebastian Sirén, and the editors for providing helpful comments. We acknowledge funding from the Alexander von Humboldt Foundation for the first author's Humboldt Research Fellowship for Experienced Researchers. The first author also gratefully acknowledges support from the WZB Berlin Social Sciences Center, where she was a Visiting Researcher (2018–2019) in the research groups Demography and Inequality and Work and Care, as well as her hosts Anette Fasang and Lena Hipp.

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **12**

# **Separated Families and Child Support Policies in Times of Social Change: A Comparative Analysis**

**Christine Skinner and Mia Hakovirta**

Child support policies are designed to ensure that following family breakdown, parents in separated families continue to pay for the upkeep of their children until they reach adulthood. This obligation is enshrined within the United Nations Convention on the Rights of the Child (UNCRC). Article 27 (4) stipulates that:

State Parties shall take all appropriate measures to secure the recovery of maintenance for the child from the parents or other persons having financial responsibility for that child.

Article 27 therefore, sets out a moral and legal expectation for parents to ensure they cannot abrogate their financial responsibility to children on separation, but also for states to ensure separated parents continue to fulfill their responsibilities. This is a laudable aim and in practice, child support policies (or child maintenance policies) also try to reduce poverty, especially among single parent families following relationship breakdown. Child support refers to a cash sum to be paid regularly for the upkeep of children. Normally it is

M. Hakovirta (B) University of Turku, Turku, Finland e-mail: miahak@utu.fi

C. Skinner

University of York, York, UK e-mail: christine.skinner@york.ac.uk

paid by one parent to the other and is separate from any spousal or matrimonial support. In some countries with guaranteed child support schemes, the state may pay an advanced amount of child support claiming it back later from the other parent if appropriate (Skinner, Bradshaw, & Davidson, 2007).

Previous comparative research on child support policies endeavored to: map out the different institutional arrangements and ways policy operates (Corden, 1999; Skinner et al., 2007; Skinner, Hakovirta, & Davidson, 2012); how policies have developed historically in legal and moral terms (Wikeley, 2006); and more recently, how they have handled complex families and shared care arrangements (Claessens & Mortelmans, 2018; Hakovirta & Skinner, 2021). We argue in this chapter that it is important to examine social change and explore whether child support systems are adapting to changing family arrangements. Child support systems sit at the very fulcrum of change at a time when obligations to kin are questioned, adapted, rejected, or renewed by separated parents during the process of family breakdown. Yet social policy often struggles to keep up with changing family patterns (Meyer & Carlson, 2014) and in some countries (UK for example) policy makers have failed to recognize how changes in social norms affect the rates of compliance with child support payments (Andrews, Armstrong, McLernon, Megaw, & Skinner, 2011).

In this chapter, we first provide an overview of child support systems in 15 European countries drawing on a novel set of data. Second, we explore changes in child support systems with respect to gender equality and children's rights; whether societal changes in these domains are reflected in child support systems. This chapter draws from an international comparative study on child support systems from fifteen countries (two of which are regions within counties), Australia, Belgium, Denmark, Estonia, Finland, France, Germany, Iceland, New Zealand, Norway, Poland, Spain (Catalonia region), Sweden, United Kingdom (UK), and the United States (US, state of Wisconsin). We analyze the responses of national informants to a standardized questionnaire, which among other things, provides information on the policy objectives underpinning child support systems; the key operational features; how child support amounts are calculated and any changes in policy approach since 2006. We focus our analysis using a mixture of informants' accounts about the policy principles and whether informants explicitly mention gender equality as being an important component, alongside their reports on whether paternal/shared care and mothers' earnings were important aspects of their systems and featured in policy changes over time.

The structure of the chapter begins with a review of evidence outlining the key social changes which reflect a more gender equal division of paid work and care responsibilities in families. This sets the context to the study using the extant literature. We then describe the international study and explain the national informant method before moving onto our findings. First, the findings provide an up-to-date overview of the different types of child support systems and how they determine child support obligations from new research data collected in 2017. Second, we present our analysis of the national informants' reports on the main policy principles in their child support systems and consider any changes since 2006. We use that time frame in order to update and extend the earlier comparative study of 14 countries provided by Skinner and colleagues (2007). It used the same informant method and covered broadly similar aspects of child support systems as we do here in 2017.

# **Social Change Reflecting Gender Equality**

Research studies assessing gender equality commonly use a number of quantitative measures including: female and male employment rates and earnings; engagement in different occupation sectors; time spent caring (for children and other relatives); and time spent doing domestic labor in the home. We make no attempt to discuss gender equality in these broad terms, other than to note that despite increasing rates of female employment, women are still subject to a gender pay gap (OECD, 2019). Rather, we aim to set the context to this chapter by focusing on key features of gender equality specifically in regard to parental responsibility: that is mothers' participation in employment and the division of labor in intact families.

#### **Maternal Employment**

It is well known that maternal employment rates have increased consistently in most developed countries. Looking at the latest data from the OECD, around 66% of mothers with children aged 0–14 were in employment in 2014, but rates differ considerably across countries (OECD, 2019). Looking at the 15 countries we focus on in this chapter, for five of them (Belgium, Denmark, Finland, France, and Sweden) more than 70% of mothers with children aged 0–14 were in work, with rates being particularly high at around 82–83%—in Denmark and Sweden. By contrast, in Australia, Estonia, Germany, New Zealand, Poland, Spain, UK, and U.S. employment rates for mothers with children aged 0–14 were between 60 and 65%. For Norway there was no data.

Employment rates also differ by partnership status. A comparison of employment rates by different partnership status shows that in 2014 single mothers were employed about ten or more percentage points lower than partnered mothers in Australia, Belgium, Denmark, New Zealand, Sweden, and the UK. Whereas in Estonia and the US the employment rates for single mothers were higher than for partnered mothers. In the remainder of our comparison countries, employment rates for single mothers were only slightly lower than partnered mothers, ranging from 65% employment rate for single mothers in Finland to 58% in Spain (OECD, 2019).

Overall, data from the OECD shows that in eight of our countries single mothers had higher employment rates than partnered mothers and in the remaining seven there were only slight differences (also see Chapter 9 by Adema, Clarke, & Thévenon in this volume). On average therefore, single mothers' employment rates are on a par with partnered mothers or are slightly higher. However, maternal employment rates are only part of the story of social change in families. Understanding how change is manifest in terms of the division of labor between parents gives a different picture.

### **Division of Labor**

In examining the division of labor, O'Connor (2013) conducted a comparative analysis of work and care responsibilities between parents showing these as "work-care models." These models provide information on the numbers of single earner families; dual earner-carer families and within that group, dual full-time earners and dual one-and-a-half earner families (where the latter is commonly a full-time male earner and part-time female earner). We have updated that analysis here using recent OECD data (2019) and have presented the results across our 15 countries organized by type of child support system shown in Table 12.1. This sets the context of the division of labor between parents in intact families comparatively for the 15 countries in our study.

Table 12.1 shows the percentage of children (aged 0–14) living in couple households by the employment status of adults in their household. Data are therefore presented from the child's perspective; they look at the proportion of children in a given household type with a given employment status. In most cases at least one member of the couple is one of the child's parents, but the adults may also be other relatives (an older sibling or grandparent) or non-relative guardians. The table presents a more nuanced picture than the simple OECD (2019) average of 50% of children living in dual earner


**Table 12.1** Employment patterns among couple households with children aged 0–14, 2014 by type of child support system across 15 countries

\*Captures all other types of working patterns, including where both adults are in part-time employment and single-earner households where one adult works parttime and the other does not work, plus any single or two earner households where information on one of the adults working hours is missing

\*\*Data in the U.S. refers to children aged 0–17 and no distinction is made between full-time and part-time work

*Source* OECD (2019). Family policy database. http://www.oecd.org/els/family/database. htm#labour\_market

households. It shows Denmark and Sweden are most characterized by fulltime dual earners followed by the US. In Poland and Finland about 50% of children live in households where both parents work full-time. Percentages are slightly lower in France, Estonia, and Belgium. Australia, the UK, and Germany are characterized by one adult working full-time and the other adult working part-time, a one-and-a-half earner model. Yet, there is still a substantial minority of households where over a third of children live in more traditional male breadwinner households in Australia, Estonia, Poland, Spain, and the US. Only Sweden and Denmark have a very low percentage of children living in a single earner household (16 and 14%, respectively). This, alongside their higher percentage of dual full-time earners, implies they have the greatest gender equality in employment between parents.

Despite the importance of the labor market however, the household is the central site in determining how the division of labor is organized in families. Such that changes in work patterns (i.e., increases in female employment rates) are closely interlinked with the sharing of housework and caring tasks between parents. We might expect therefore, that women spending more time in paid work would result in an equivalent rise in men's participation in domestic labor. On average however, while men and women's contributions have become more equal over time, this is because women are doing much less housework than hitherto, not because men are doing much more (Bianchi, 2011). On average men's contributions have increased only slightly.

Regarding childcare tasks a similar picture emerges: while fathers are spending more time on childcare activities, it is still not gender equal (OECD, 2019). We also know that historically for separated families, mothers have been the main resident parent and fathers have had a minor caring role which has been actualized by contact arrangements rather than resident arrangements. More recently however, evidence shows a growing number of separated parents do jointly care for their child(ren) either equally, or at least for 30% of the time (Fehlberg, Smyth, Maclean, & Roberts, 2011; Hakovirta & Skinner, 2021; Smyth, 2017; Trinder, 2010). The practice of "joint physical custody" where a child spends equal time living with both parents and both parents have responsibility to physically care has also become a more popular arrangement in some countries. In most of the Nordic countries (except Finland), for example, 22–35% of children have this kind of joint custody arrangement post-separation (Hakovirta & Eydal, 2020). In contrast, in the UK, reports from a number of different surveys show that shared care prevalence ranges from 3 to 17%. Notably however, some accounts from resident parents (commonly mothers) suggest that equal 50–50 time arrangements could be as low as 1% (Haux, McKay, & Cain, 2017). Currently the evidence presents a very mixed picture, partly because there is no common definition of what shared care or shared physical custody means making it difficult to measure comparatively (Hakovirta & Skinner, 2021).

Measuring the amount of time spent by men and women doing various tasks has proved useful in understanding the division of labor in families, but there are many other explanations for these patterns which lie outside such measurement. This includes, among other things, the way heterosexual couples "do gender" and how gender borders become established in families (Lyonette & Crompton, 2015). Research by Craig (2006a, 2006b) also shows that parenting as a mother is not the same as parenting as a father, even for women who work full-time in the paid labor force. The conditions of childcare appear to be harder for mothers: mothering involves more double activity, more physical labor, a more rigid timetable, and more overall responsibility than fathering.

A strong body of evidence therefore continues to grow. We highlight that here to show the structural and social changes in family practices in terms of the division of labor between parents but also to show the small but concomitant move toward greater gender equality (as measured by time spent doing different activities). These trends toward a more equal sharing of parenting responsibilities demonstrate a shift in social norms which vary by country and are influenced by many factors, including family policies. It is not the purpose of this chapter however, to offer any explanation of these influencing factors; rather the central concern here is to consider whether these changes are recognized in child support systems and to raise questions about the possible implications. Certainly, the traditional breadwinning family is no longer common in many countries, making it harder for child support systems to decide how best to allocate economic responsibility between separated parents. This analysis will explore how child support systems might recognize gender equality in the sharing of work and care in separated families using new data on child support systems across 15 countries (which includes two regions: the state of Wisconsin in the US and Catalonia in Spain).

## **Methods**

We use a national informant method, in which policy experts with knowledge of their own country's child support systems are recruited to complete a detailed standardized questionnaire. This is a method that has been used successfully in the past in this policy area (Corden, 1999; Skinner et al., 2007, 2012). It differs in important respects from comparative documentary analysis or comparative analysis based on data generated by administrative institutions or government departments as it generates new primary data provided by informants. This means it is qualitative data and the elements that we examine here are based on informant's assessments of their systems regarding policy objectives and operations.

Our analysis, therefore, does not attempt to systematically capture all objectives, functions, rules, and regulations that might operate in the different child support systems. That is because such an approach would not only require the collection of detailed (and expensive) data from a range of different institutions within and across countries (courts, enforcement agencies, child support agencies, local welfare intuitions, and social security systems) but even then still might not give an indication of the relative importance of policy principles within each country.

We report the findings on the basis of the emergent policy principles arising from our comparative analysis of informants' accounts of their systems. It is informant's perspectives that are being analyzed here, valuable in their own right as they provide an insider view. However, that is both the strength and weakness of this method. We gain deeper insights from an insider interpretation of the policy framework, but this is highly dependent upon the informant's own perspective, their level of knowledge, and the quality of their assessment of their systems. It should therefore be regarded as a complementary method to other forms of comparative analysis of child support systems rather than a substitute for them.

Mostly we recruited one national informant for each country, many of whom were academics that had earlier experience in similar research studies either acting as informants, or were involved in collecting or analyzing data in previous comparative studies on child support, or were recruited based on their earlier contribution to the field. As each informant was an expert in the field of enquiry in their own country, it eased the task of data collection and validation. Data was collected at the end of 2017 and involved national informants from 15 countries and therefore a key strength of the study is the diversity and range of child support systems there-in, covering Europe, the US (the state of Wisconsin), Australia, and New Zealand. Among other things in the questionnaire, informants provided information on child support policy, the main operational aspects of their systems, a description of the main policy principles and how these changed over time since 2006.

We begin presenting our findings with an up-to-date overview of current child support systems followed by an analysis of the policy objectives as reported from the national informant insider perspective. We use their answers to the first part of the questionnaire. First, we describe the policy principles based on our analysis of the emergent themes arising out of the informant reports. We identified four themes which we have grouped under the broad heading of "child's rights and enforcement." Second, we provide an analysis of informants' accounts where they spontaneously mention gender equality as being important, or where the recognition of paternal care and maternal earnings/incomes in calculating child support amounts is considered of importance. This focus helps us address the question of how and whether child support systems are regarded as responding to social changes such as greater gender equality in the parental sharing of work and care responsibilities. This part of the analysis however has limitations as there was no explicit question asking informants to comment on whether systems had become more gender equal. We therefore rely on informants raising this as an important issue in either of the two ways we have just described.

# **Overview of Child Support Systems**

In an earlier comparative study on child support systems, Skinner et al. (2007) analyzed some key aspects of the systems across 14 countries and grouped them according to where the locus of responsibility lay for determining the child support amounts to be paid by non-resident parents. Three main types emerged: agency-based systems, court-based, and hybrid systems (in which the courts or another agency/institution could also get involved in calculating amounts). We have followed the same approach here and in Table 12.2 present an overview of the current systems grouped into these three types. It is important to note that parents can also make private agreements in all countries and while this represents a common basic principle, parental agreements are not always officially ratified or may have specific conditions attached if a parent is in receipt of social assistance benefits. Because private agreements are a common option in all countries, we consider the typology as being based on agreements relating to the formal system rather than on purely private agreements. Such formal arrangements may often mean they occur where parents cannot agree private arrangements between themselves.

Table 12.2 shows that five countries operate a primarily agency-based national system (Australia, Denmark, New Zealand, Norway, and the UK). Of the five agency systems, only Denmark and Norway offer a guaranteed amount of child support which is paid in advance by the state and is usually claimed back later by the state from the non-resident/liable parent. Such guaranteed schemes are common in Nordic and European countries (nine of our countries offer this), whereas none of the four English speaking countries offer such schemes.

Five countries and one region mainly use court-based systems to determine support amounts (Belgium, Estonia, France, Germany, Poland, and Spain [Catalonia]). These systems are characterized as discretionary and in most



**12 Separated Families and Child Support Policies … 277**



cases, they are less likely to apply standard rules and formulae when working out support liabilities. In some of these countries, various agencies may also be involved in enforcement where there is non-payment (*Dienst voor Alimentatievordering* [DAVO] in Belgium; Social Security Agency in France), but generally these agencies are not involved in calculating the level of the original liability. Hybrid systems on the other hand operate where there is more than one locus of responsibility in setting support amounts in formal agreements; three countries and one state (Finland, Iceland, Sweden, US [Wisconsin]) involve courts and other agencies.

This overview provides the most up-to-date comparative analysis of the key institutional characteristics of child support systems from an insider perspective. This is important new data and we use it here to set the context for our analysis on emergent policy principles underpinning these systems, to which we now turn.

# **Policy Principles**

In this section we describe the informants' accounts of the original policy principles (the explicit or implicit policy goals) that operated in child support systems before 2006<sup>1</sup> and the current objectives they have identified as being important in 2017. A lot of policy activity has taken place since 2006 and the data is presented in Table 12.3. Somewhat uniquely, Iceland reported no changes since 2006 and was therefore described as being policy inactive. The first column is a summary of the data as reported by informants. The second includes emergent policy principles as identified by the authors in their interpretation of the informants' reports of policy objectives operating prior to 2006. The last two columns report changes since 2006, firstly presenting informant's views followed by the authors' interpretations of emergent policy principles arising from that.<sup>2</sup> For ease, we keep the countries organized in Table 12.3 by the three types of child support systems (agency, court, and

<sup>1</sup>This is the date Skinner and colleagues (2007) collected data on change in child support systems, so this new study here in 2017 provides some chronological continuity.

<sup>2</sup>For the purposes of completeness, Table 12.3 also identifies some of the key administrative changes as reported by informants. This serves to describe some of the policy context and is not part of the analysis per se, suffice to say that in 2017 many countries were attempting to make their systems simpler, more transparent and efficient (UK, Denmark, Belgium, Estonia, France, and Spain [Catalonia]). Finland, France and Denmark also reported administrative changes that involved more centralization or standardization occurring over time and the UK moved toward greater privatization whereby all parents were now free to make their own arrangements outside the formal system (although note Norway had also privatized their child support system in 2003). Sweden and France seemed to be working on the complex balance and interaction between publicly funded social security benefits and private child support payments.


**Table12.3**Childsupportkeypolicyprinciplesasidentifiedbynationalinformantsacross15countries(includestworegionsinCatalonia

in

(continued)


**Table 12.3**

(continued)


(continued)



(continued)



(continued)


**Table12.3**(continued)


• UNCRC = UN Convention on the Rights of the Child

• Advance CS and Guaranteed CS are similar approaches in which the state provides an amount for child support up front PWC = Parent with care(usuallymothers).NRP=Non-residentparent(usuallyfathers) hybrid). We show these two perspectives (the informants' accounts and the authors' analysis of emergent themes arising from those accounts) to enhance the rigor of the analysis and to be as transparent as possible. In the text however, we only discuss the findings of the emergent policy principles in detail.

The set of emergent policy principles are grouped under the broader theme of "child's rights and enforcement" and these are: parental duty, child's rights, poverty alleviation, and enforcement. We discuss these first before moving to our analysis in the next section which explores whether systems are reacting to social change and perceived trends in greater gender equality in parenting.

# **Child's Rights and Enforcement**

# **Parental Duty**

Parental duty is the most common policy principle identified across all counties as occurring within the original manifestation of child support systems, such as where a "legal duty or obligation" is set for separated parents to continue to support their children post-separation. Some country informants specify the exact nature of parental duties as being split into health, education, and finances while others are less specific. Two informants (Norway and New Zealand) noted that the principles stressed the permanence of the parental duty; that it could not be revoked by parents' remarrying or repartnering. Germany noted that since 2008 all children were now treated equally, regardless of their parent's marital status (previously children were treated differently depending on whether their parents were legally married or cohabiting). Germany was exceptional among the countries studied here, but is now aligned in treating the parental duties of all types of parents equally.

# **Child's Right**

In regard to child support being a child's right this was explicitly identified in seven countries (Australia, Belgium, Estonia, Iceland, Norway, Sweden, and the US [Wisconsin]), but generally only in the original manifestation of the systems. It was not mentioned as an important policy change or principle post 2006. Even so a child's right was captured in a number of different ways including: through direct reference to upholding the UN Convention of the Rights of the Child (Estonia); through the use of assessment tools that based child support calculations on the cost of a child standard (Australia); or through the principle that children's "standard of living" should not be affected by separation U.S. (Wisconsin), that their living standard should be maintained (Belgium) or that the children had a right to share in their parents' standard of living post-separation even if that standard increased (Sweden and Australia).

There are subtle differences across the principles with regard to a child's standard of living: one that aims to tie a child's right to a standardized "cost of the child" assessment; one that aims to protect children from a fall in their living standards post-separation; and another that enables them to share in any future prosperity of their parents. There was also a small set of changes that did occur post 2006 that could be regarded as relating to maintaining a child's standard of living; index linking child support payments. This was mentioned in Australia, where the minimal amount of child support was index linked to the Consumer Price Index and in Finland where all child support was index linked to living costs.

Another aspect of policy that might relate to the principle of ensuring a child's right, is setting a universal minimum amount of child support. This was mentioned by Iceland as an original principle pre-2006 and also that child support belonged to children; the parents were expected to spend it only on children. In relation to a child's right to have the child support spent on them, Australia also discussed how changes post-2006 increased the proportion of child support children could receive directly from the paying parent (up to nearly a third of the child support). This principle of direct pay to children could conceivably relate to theories of the "new sociology of childhood" in which children are seen as independent actors and capable of taking charge of their own support money. Alternatively, it could be argued it is a means by which fathers can bypass mothers' control over the spending of child support. This is something that fathers in Anglo-Saxon countries have raised concerns about (Cook & Natalier, 2013; Skinner, 2013). In any event, this analysis demonstrates a range of ways that a principle of a "child's right" could be manifest and regarded as important. This is different to another policy principle that emerged from the data, that child support should alleviate poverty.

#### **Poverty Alleviation**

Poverty alleviation emerged as an important early policy principle and was described in different ways by eight informants as being part of their child support systems before 2006 (Belgium, Denmark, Finland, Germany, Poland, Norway, Sweden, and the US). For some it was described as an explicit policy goal and for others it was more implicit. A key aspect of child support systems was the operation of advanced or guaranteed support. Among these eight countries, all but the US and Belgium operated such schemes. All the Nordic countries mentioned it in relation to poverty alleviation (bar Iceland, which did not, although it too has a guaranteed support scheme). The guaranteed/advanced schemes operated in a range of different ways: it could be universally available to all parents with care who claimed it, or available only to single parents with care; it could be for a flat rate amount or means tested. There is not enough scope here to describe the specifics of different schemes; suffice to say, they were mentioned by some informants in relation to a principle of alleviating poverty pre-2006 and were most common in Nordic countries. Finally, of interest is New Zealand, in which poverty alleviation was said never to be a goal of the child support system, but rather the goal of welfare benefits and tax credits. So here is one country at least, where the principle of poverty alleviation was reported as not belonging to child support policy.

For policy changes post-2006, three informants mentioned that poverty alleviation was an important principle (the UK, France, and Poland). This does not mean to say it was unimportant in other countries, but from the insider informant perspective it was not mentioned as salient currently. For Poland however, poverty alleviation was highly important both pre- and post-2006. As the informant explained, this was because Poland tried to abolish their guaranteed support scheme in 2004 replacing it with a new social security benefit for single parent families. However, after "mass social protests from a new social movement of single mothers" it was reinstated in 2007, although in a much more limited way than previously.

Clearly for Poland, poverty among single parent families was highly politicized. For the UK and France however, poverty alleviation seemed to emerge mostly as a principle post-2006. In France, it emerged implicitly as a result of complex changes to the ways in which single parents' entitlements to social security benefits were established and in turn how this interacted with "fixing the amounts" of child support. The informant notes there is not enough information about the payment of child support in France, and since 2016, the authorities have become more interested in improving effectiveness of policies to alleviate poverty in single parent families. In contrast in the UK, poverty alleviation was not an explicit policy objective when the Child Support Agency (CSA) was set up in 1993 (though it was widely regarded as being implicit). It was not until 2008 did it become one of four explicit objectives following major reforms which replaced the CSA and returned

child support obligations back into the hands of parents. Ironically therefore, the state made an explicit claim of poverty reduction at a time when it withdrew from taking responsibility for ensuring child support was paid. This highlights how policy principles are differentiated from policy operations and how they can be inconsistent with one another. We note, however, that some country's child support systems are better than others at helping alleviate child poverty. We briefly discuss that evidence here.

Evidence from other comparative studies shows that child support is associated with a decrease in poverty among single mother families within a variety of countries; that is when it is paid (Cuesta, Hakovirta, & Jokela, 2018; Hakovirta, 2011; Hakovirta, Meyer, & Skinner, 2019). For example, in the UK and Australia child support payments reduced poverty among single mother families, respectively, by 14 percentage points and 21 percentage points (Skinner, Cook, & Sinclair, 2017). In attempting to reduce child poverty however, one key problem for child support systems is dealing with non-compliance because failure to do so can reduce the effectiveness of child support. Certainly, not all eligible single mothers receive payments from the other parent. The highest rate of receipt is in countries where the state guarantees child support when the non-resident parent does not pay. Comparing Finland's system (which has a guaranteed scheme) with the UK and the US helps demonstrate this. In Finland 77% of single mothers received child support in 2013, but in the UK and the US only one third did (Hakovirta et al., 2019). Yet, even if compliance rates are high, another hidden problem has recently come to light which shows how the anti-poverty effectiveness of child support is disrupted by the interactions between child support payments and social security benefits (Hakovirta et al., 2019; Skinner et al., 2017).

The effects of the interactions mean that in some of the 15 countries discussed here, we know that child support payments are treated as a substitute for social security benefits leaving single parents no better off financially. This is because the state recovers the child support money through various mechanisms, effectively capping single mothers' incomes at the level of the social security benefits they receive. This capping effect appears to operate in New Zealand when assessed using model families (Skinner et al., 2017) and for Finland and Germany when assessed using Luxembourg Income Study (LIS) data on actual recorded child support payments in real families (Hakovirta et al., 2019). In contrast, the UK stands out as it treats all child support payments as a complement to social assistance benefits, meaning it can act as a top up to single parent incomes above social security benefit levels, but note only if it is paid (Hakovirta et al., 2019). Similarly, in Australia single parents can keep nearly all of the child support, again if it is paid (Hakovirta et al., 2019). Therefore, while poverty alleviation is an important policy principle and operationally some countries child support systems appear to be more effective than others, we can see that enforcement could be an integral part of poverty alleviation and an important part of maintaining child rights.

## **Enforcement**

Only three informants however, emphasized enforcement as being particularly important (France, Poland, and Estonia). For France looking at the changes over time pre- and post-2006, enforcement was described as being originally set within the civil legislation to uphold parents' liabilities. Whereas in 2016 the emphasis changed; enforcement also became part of the social security benefit reforms mentioned earlier under poverty alleviation. The social security system was now able to enforce "simple" child support agreements made between parents. So again, for France we see how child support obligations interacted in complex ways with social security entitlements, especially for poorer single parent families. It seemed however, that despite the complexity, they were trying to find better ways of enforcing payments.

Similarly in Poland, the child support system emphasized enforcement as an explicit policy principle, though perhaps in stronger terms than reported in France. From its inception, the child support system reportedly set out to make non-payment "socially unacceptable," making it a criminal offense to not pay. Later, a new act in 2015 required local authorities to add the names of non-payers into a "National Debt Registry" to better enforce prosecution. However, the wording of the act was amended in 2017 to ensure that non-payers were given the chance to pay off their debts, rather than be fined, or imprisoned as the first course of action. For Estonia, debates regarding improving enforcement measures arose more recently in 2008 and were reemphasized by the new government in 2015. New policy measures ensued and in 2016 penalties for non-payment (such as revoking certain licenses for hunting, driving, and gun ownership) were increased. Additionally, bailiffs were given greater powers to remain in contact with non-payers so assets could be seized should they persistently not pay. Overall, France, Poland, and Estonia appeared to share similarities: all three were court-based systems and the policy emphasis shifted toward encouraging payments, rather than simply punishing non-payment (especially evident in Poland).

In sum, the analysis highlighting emergent policy principles across countries are all acting to protect children's rights, in one way or another. But we found other sets of principles which related more to how systems dealt with parental responsibilities in terms of recognizing paternal care/shared care arrangements or maternal employment or incomes. To which we now turn.

### **Gender Equality**

As discussed above, one of the key aims of the analysis of this comparative data was to ascertain whether key social changes in families, such as greater equality in parenting responsibilities, were being reflected in child support systems. Certainly, gender equality is manifest in calls from separated parents and others for more equal "joint physical custody" arrangements post-separation (Hakovirta & Skinner, 2021). However, informants were not directly asked about gender equality. We have therefore used any explicit mention of gender equality (where it occurred spontaneously in responses in the first part of the questionnaire) alongside our analysis of informants' reports on the importance of systems taking account of shared care/paternal care and mothers' employment/incomes in determining child support liabilities.

#### **Recognition of Paternal Care**

When informants were asked to describe the main policy principles of their systems, a few explicitly mentioned gender equality as being particularly important. As shown in columns three and four in Table 12.3 these countries were Australia, New Zealand, and Norway. Overall six informants mentioned recognition of paternal care/shared care as being an important policy principle. This recognition occurred more often after 2006. For example, only New Zealand and Norway mentioned paternal/shared care as important both pre-2006 and post-2006, and four other countries mentioned it only after 2006 (Australia, Estonia, Finland, Spain [Catalonia]).

New Zealand reported having an original policy goal of equity between parents in which paternal care was recognized and this was later strengthened following public consultation and the development of a new act in 2013. This act altered the child support formula to reflect the apparent increase in care contributions made by separated fathers, but it also recognized the rising rates of maternal employment (discussed further below). Similarly for Norway, they strengthened their position of treating mothers and fathers equally post-2006 (see discussion on maternal employment below). In Australia, major reforms in 2008 reportedly emphasized fathers' contribution to care more strongly and indeed waived child support liabilities for low income fathers who remained in contact with their children post-separation. Reforms also recognized the increased participation rates of mothers in the labor market (discussed below). In Estonia's court-based system, paternal care recognition was said to occur within a legal framing, with the introduction of a new family law act in 2010 stipulating more clearly the definition of custody and the obligations of both parents to care for their children. In Finland suggestions were made in 2016 that shared care arrangements be added into child custody laws. Whereas for Spain (Catalonia), assumptions around shared care were introduced for the first time in divorce legislation in 2005 and while it was thought this would have little effect, a legal presumption of shared care spread across a number of northern regions in Spain with new laws enacted in 2010.

What emerges from these accounts is differences in the way informants in the six countries talked about recognition of paternal care, with those in Court or hybrid systems (Estonia, Finland, and Spain [Catalonia]) tending to refer to debates around "custody" and shared care within family law. Whereas in New Zealand, Australia, and Norway these are agency-based systems, and these informants tended to report changes in the child support formula made to better recognize paternal care. Either way, recognition of paternal or shared care has been reported as an important part of the policy principles either as a source of debate, or of policy change in child support systems for six of the fifteen countries. This indicates at least some recognition of changing social norms and associated trends in gender equality. Closely related to recognition of paternal care, is recognition of mothers' increased capacity to have independent incomes through paid employment.

## **Recognition of Maternal Income and Employment**

As mentioned above, gender equality was a strong principle mentioned explicitly only in informants' accounts from Australia, New Zealand, and Norway, though there were differences. In New Zealand, it seemed they adopted a principle of gender equality early on, whereas in Australia it was embedded in debates that took place around the 2008 policy reforms. Policy debates in Australia acknowledged mothers' increased participation in the labor market, and the possibility at least, of mothers gaining independent incomes from earnings. This resulted in mothers' incomes being included in a new "incomes share" formula for calculating child support. It is important to note at this point that other countries may also have counted mothers' incomes in their calculations, but what we are reporting here are the informants' perspectives of their main policy objectives and not providing a detailed administrative account of all operational features.

In Norway too, gender equality was mentioned as a strong principle in regard to the acknowledgment of trends in maternal employment; policy debates there focused on separated parents both having equal care and equal financial responsibilities. Norway has taken this principle further by using child support policy to manipulate parental behaviors in the labor market. For example, the informant reported that when calculating child support, if the mother is unemployed, an "imputed" earnings figure is applied to the child support calculation and this is done to encourage her participation in employment. But it is also done to protect fathers from having "undue" support costs being passed onto them as a result of mothers choosing to "opt out" of the labor market. This implies that the movement in and out of the labor market for separated mothers in Norway is seen as a free choice they can make. Whether or not this policy assumption reflects the reality of Norwegian mothers' capacity and freedom to enter the labor market, it is important to note that Norway has a largely privatized child support system and parents can agree to have no child support arrangement if they so wish. So, gender equality based principles (and related operational tools) are only applicable when separated parents choose to use the formal system (for which they pay a fee) or where the parent with primary care seeks guaranteed child support from the state. But what does all this mean? What can we say about gender equality and child support systems? We now discuss that in our conclusion.

## **Conclusion**

Overall, we found that broader social changes for greater gender equality in the division of labor in families are not reflected very much in child support systems over the last 10 years (2006/7–2017). Among the fifteen countries studied here, only in Australia, New Zealand, and Norway, did gender equality emerge in informants' accounts as an important policy principle, or as an important part of debates in child support policy reforms. All three countries have agency-based systems and this may be an important factor, possibly because they tend to operate more explicit rules and formulae in calculating child support liabilities. Thereby decision-making is made more transparent potentially rendering these systems more susceptible to greater scrutiny regarding the gender equal treatment of both parents' incomes (Cook & Skinner, 2019; Skinner, 2013). However, it is interesting to note from other research that the remaining two agency-based systems (UK and Denmark) do not count mothers' incomes in calculating child support amounts (Hakovirta & Skinner, 2021). Arguably therefore, they are more attached to a traditional male breadwinner model of separated families, highlighting there is little recognition of changing social norms as of yet, at least in terms of recognizing parental earning capacities.

Using the new data in the study reported here, we cannot be sure however whether assumptions of a male breadwinner model operate in the court and hybrid child support systems. Partly because these systems tend to operate more discretionary decision-making procedures when setting child support liabilities. In that regard, three informants from Estonia, Finland, and Spain (Catalonia), did note that while policy discussions had taken place in their countries around giving greater recognition to paternal care, this mostly occurred within a broader family law frame than in the operation of child support systems per se (for example, in divorce and custody arrangements post-separation). So, equality in care time at least, may become a more important policy objective in the future.

Certainly, our analysis is exploratory and has limitations regarding the type of data collected using a national informant method and questionnaire format. Even so, it seems that from the informants' reports on the policy principles, there is little recognition of the broader social changes denoting shifts toward greater gender equality in parental responsibilities (i.e., more maternal employment and more paternal care). There might be very good reasons for this—such as policy focusing more on other key objectives of child support policy—that is to protect children's rights and alleviate child poverty. Gender equality could therefore be seen as a competing policy objective and related more to parents' interests than to protecting children interests.

Certainly, in the three countries reported here where gender equality was an explicitly recognized feature, not all their child support systems were good at poverty reduction; at least as measured in other studies focusing on single parent families. We know, for example, that New Zealand treats child support payments as a substitute for state social assistance benefits paid to poor single parent families. This means that child poverty is not reduced by child support payments in New Zealand as incomes are capped at the rate of social assistance benefits (though poverty may be reduced via social security benefits) (Skinner et al., 2017). In contrast, Australia moved to treating child support as a complement to social security benefits, thereby it topped up incomes. In theory at least, the anti-poverty effectiveness of child support payments is enhanced in Australia on this basis (Hakovirta et al., 2019). Norway is fairly unusual, parents are supported to enter into a private agreement on child support, but if they cannot agree, the parents can apply for child support to be established by the national social insurance institution. Even so, gender equality featured quite strongly in Norway and appeared to be based on assumptions of an adult worker model.

The findings from this new comparative study demonstrate that we have a long way to go before we can fully understand whether child support systems are adapting to perceived social changes in gender equality in the division of labor in families. Certainly, it is clear that more extensive research would be needed to address the question of whether the pursuit of gender equality in child support systems would be beneficial and to whom it might benefit, or to whom it might cause possible harm. To do so would require a full examination of child support systems, social security systems, and the interactions between both systems in terms of the poverty effects produced. But whether child support systems are, or should be, a key policy in the fight against poverty is itself a politically sensitive topic.

As we have highlighted, child support systems could face potentially competing objectives, the pursuit of gender equality (at least in terms of acknowledging an equal division of earning and caring responsibilities) is likely to be at odds with the principle that child support policy should tackle child poverty. Yet, while the traditional breadwinner model of families is in decline and families may achieve greater gender equality, the question of whether a principle of gender equality as applied to child support policies is a good thing to pursue, is one of the most difficult challenges for future family policies. Our comparative analysis makes a unique contribution to that debate.

**Acknowledgment** This work was supported by the Academy of Finland grant number 294648, 2016.

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **13**

# **Dual-Earner Family Policies at Work for Single-Parent Families**

**Laurie C. Maldonado and Rense Nieuwenhuis**

Family dynamics are changing and single-parent families are becoming more common across countries. In their flagship report "Progress of the World's Women, 2019–2020," UN Women (2019) demonstrated that, contrary to popular belief, couples with children do not constitute a majority of all families, but rather there are many different types of families (also see Chapter 5 by Razavi and Chapter 9 by Adema, Clarke, and Thévenon, both in this volume). Single parenthood is considered a "new social risk" in poverty and inequality (Bonoli, 2013). Therefore, policy makers and legislators have designed *targeted* policy specifically for single parents, such as targeted child benefits to single parents. In addition, legislation and social policy have been designed and implemented specifically for single parents, such as child support (as analyzed in Chapter 12 by Skinner and Hakovirta in this volume) and family law such as child custody and shared residence. This chapter takes a different approach, based on the *universalist* argument that without adequate social protection that benefits all families, those families that are more vulnerable are often hit the hardest. We focus on family policies, and

R. Nieuwenhuis Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden e-mail: Rense.nieuwenhuis@sofi.su.se

L. C. Maldonado (B) Molloy College, Rockville Centre, NY, USA e-mail: lmaldonado@molloy.edu

specifically we examine whether and to what extent single parents benefit from the same family policies that are available to all families with children.

Based on the extant literature on single parents and family policies, there seems to be a strong case that family policy does reduce poverty among all families with children. In fact, some literature even suggests that single parents benefit more from universal family policies than couples with children. This position can also be found in our own work (Maldonado & Nieuwenhuis, 2015; Nieuwenhuis & Maldonado, 2018b), but the claim is far reaching and requires further review. This chapter sets out to examine how family policies differently affect the poverty rate of single-parent families versus couples with children and also probes whether or not there is a premium—or penalty—for single parents. Following Daly (Chapter 2 in this volume), we focus on child income support (with an emphasis on family benefits), early childhood education and care (ECEC), and parental leave. We review the literature on the question of whether these policies benefit single-parent families more. In addition, we examine a number of comparative family policy data infrastructures to empirically examine the question of whether family policies provide additional benefits to single-parent families.

The remainder of this chapter is organized in three parts. First, we introduce the rise of single parenthood and aspects of their well-being such as their employment and their elevated poverty risk. It provides a theoretical framework to explain challenges associated with single parenthood. The next section provides a literature review on family policy outcomes for single parents and parents in couples. This is followed by an empirical illustration of what can be learned regarding the question to what extent single parents and two-parent families benefit from family policies. To conclude, we examine how to improve the data infrastructures and research to deepen our understanding of the role of family policy in improving the lives of single parents and their children.

# **Single Parents in a Triple Bind**

Single parenthood has become increasingly common, but varies substantially across OECD countries. The share of households with children that are headed by a single parent (as a percentage of all households with dependent children) ranges from around 10% (e.g., in Italy, Poland, or Slovenia) to around 25% (e.g., in the United States, Sweden, and Ireland) (Nieuwenhuis & Maldonado, 2018b). Gender inequality is a major concern, as the majority of single-parent families are headed by women. At the same time, there are shifting demographics and a small and slightly growing percentage of singleparent families that are headed by fathers. In fact, fathers are becoming more involved in the care for their children (even separated fathers), especially in the Nordic countries (Nieuwenhuis, 2020). Single parents face higher risks of poverty compared to couples with children, and this is more pronounced in some countries (Nieuwenhuis & Maldonado, 2018b). The challenges faced by single parents can be explained by the triple bind framework.

Single-parent families risk being caught in a *triple bind* of inadequate resources, employment, and policy (Nieuwenhuis & Maldonado, 2018b). With respect to their resources, single parents are likely to have a low level of education compared to parents in a couple (McLanahan, 2004). This pattern of an "educational gradient" was found across a number of social-democratic, liberal, and continental welfare regimes (but not in Southern Europe), yet was found to hardly be the "smoking gun" (Härkönen, 2018, p. 43) in explaining single parents' elevated poverty risks. Instead, with only one earner and caregiver in the household, single parents are without a potential second earner to fall back on during difficult economic times and without a second caregiver to help combine work and family responsibilities. The majority of single parents work; however, many are employed in low-quality jobs. Jobs that are often low wage, without protection, and without adequate workfamily supports. Working single parents are more likely to experience in-work poverty compared to parents in couples (Nieuwenhuis & Maldonado, 2018a) and experience worse work–life balance (Esser & Olsen, 2018). The third bind pertains to inadequate or missing policy. For instance, many countries have policies that support employment as a means to reduce poverty, but if such demands are not met by the support to combine work and family responsibilities, it may be particularly difficult for single parents to meet such demands. Taken together, the triple bind makes it difficult for single parents to work and care for their families. Inadequate resources and employment, combined with no social safety net and an ineffective family policy, can leave far too many single-parent families in poverty.

The triple bind of single-parent families provides some insights for family policy. On one hand, most aspects of the triple bind are not unique to single parents. Couples with children, too, can experience in-work poverty, particularly if they have not obtained a high level of education. In addition, they can experience difficulties in combining work and family, and of course, caregiving responsibilities are not equally shared among couples (even in comparatively gender-equal societies as Sweden; Evertsson & Nermo, 2007). As such, it may be expected that single parents and couples with children alike benefit from income supports to children, as well as ECEC and parental leave to combine work and family. On the other hand, single parents are more likely to face many aspects of the triple bind compared to couples with children. Although parents in couples face similar work–family conflicts, single parents have far greater challenges with less income and less ability to share care work. This increased need for child income support and work– family reconciliation policies may also increase the benefit they gain from such family policies.

# **Literature Review**

## **Child Income Support: Family Benefits**

Child income supports, such as family benefits, have proven to be a key instrument against poverty among families with children. Family benefits, a type of cash transfer that offsets the cost of raising children, are particularly effective in reducing poverty for single-parent and two-parent families (Bradshaw & Finch, 2002; Bradshaw, Keung, & Czhen, 2018; Chzhen & Bradshaw, 2012; Gornick & Jäntti, 2012; Maldonado & Nieuwenhuis, 2015; Morissens, 2018). Bradshaw and Finch (2002) studied child benefit packages across countries using model families to assess the generosity of child benefit levels. They found that countries with more generous child benefit packages have lower child poverty rates. Chzhen and Bradshaw (2012) later examined both poverty and material deprivation among children in singleparent families. They found that poverty is lower in countries with higher transfers, whereas material deprivation is more related to a country's standard of living. Gornick and Jäntti (2012) report similar findings: that redistribution significantly reduced child poverty. Bradshaw et al. (2018) showed that the amount of received, as well as other benefits, is a significant portion of single-parent families' household disposable income.

Some have raised concerns about moral hazard regarding income benefits and family formation. For instance, income benefits may encourage single parenthood. However, this argument was strongly countered by Brady, Finnigan, and Hübgen (2017). In their framework that distinguishes between the prevalence and penalties associated with risk factors for poverty, they found that higher benefit transfers reduce the poverty penalty associated with single motherhood but that this did not increase their prevalence. More generally, the literature shows a limited impact of child benefits on fertility as detailed in Chapter 9 by Adema, Clarke, and Thévenon in this volume. In other words, benefit transfers do not entice people to become a single parent but do reduce their poverty.

In addition, it is important to examine which benefits are targeted toward low-income families. Research has investigated the effectiveness of benefits targeted specifically to those who need it the most vis-à-vis benefits that are designed for all families (Kenworthy, 2011; Korpi & Palme, 1998; Marchal & Van Lancker, 2019; Marx, Salanauskaite, & Verbist, 2016; Skocpol, 1991; Van Lancker, Ghysels, & Cantillon, 2015; Van Lancker & Van Mechelen, 2015; Van Oorschot & Roosma, 2017). In their seminal study, Korpi and Palme (1998) showed that universal transfers were more effective to reduce poverty overall than transfers targeted specifically at the poor. They addressed this "paradox of redistribution" by arguing that heavily targeted benefits lose the partisan support of middle-class and higher income voters, and therefore, such benefits are at risk of being cut. Brady and Burroway (2012) also found that universal transfers have been more effective in reducing poverty than transfers targeting single-parent families.

Other researchers have shown that the "paradox of redistribution" is less of a concern. Marx et al. (2016) found that current minimum income protection schemes alone are insufficient to reduce poverty. Instead, poverty reduction may benefit from a combination of both universal and targeted benefits. Van Lancker et al. (2015) found that targeting child benefits for single-parent families is an effective way to reduce poverty, as long as adequate overall levels of redistribution are maintained. Morissens (2018) found that targeted benefits and universal benefits are both indispensable in reducing poverty among both working single parents and single parents who are out of work.

#### **Early Childhood Education and Care (ECEC)**

Public services for ECEC are arguably among the most important to support working parents (Olivetti & Petrongolo, 2017) as well as beneficial to children's well-being and development (Gambaro, Stewart, & Waldfogel, 2015). ECEC is most effective when it is available, affordable, and of adequate quality (ibid., also see Chapter 8 by Vandenbroeck and Chapter 24 Sirén, Doctrinal, Van Lancker, and Nieuwenhuis, both in this volume). Family care needs can form barriers to paid work and be nearly impossible to reconcile with work requirements. One alternative is to rely on informal care arrangements, but those do not necessarily offer the stability necessary to maintain employment (Van Lancker & Horemans, 2017).

Public childcare is essential to single mothers' employment. Bainbridge, Meyers, and Waldfogel's (2003) US-based study found that childcare subsidies for working single mothers stimulated employment more so than for single mothers on welfare. They also found that family tax benefits to support childcare explained a larger share of single mothers' employment. They underlined the importance of examining specific aspects of childcare policy. Berger and Black (1992) also discussed childcare policy provisions and the need for high-quality childcare. In their study of Finland, Netherlands, and the UK, Moilanen, May, and Räikkönen (2016) found that working single mothers experience more childcare-related challenges, particularly when working nonstandard hours. In a large comparative study, Van Lancker (2018) found across European countries that single parents who had used ECEC services when their children were young were indeed more likely to be employed later in life.

The benefits of ECEC extend beyond increasing the likelihood that single parents are employed; they also improve employment conditions or improve work–life balance among those who are employed. Nieuwenhuis, Tøge, and Palme (2018) showed that ECEC services were associated with more single parents in employment but also that this employment was associated with a larger health benefit in societies with more extensive ECEC provisions. At the same time, single parents who were not employed experienced poorer health (but less so in countries with higher levels of child income support). More generally, Millar and Rowlingson (2001) reported that single parents do well in jobs that are not only well paid but also well supported (also see Rowlingson & McKay, 2002). Esser and Olsen (2018) found that working single parents experience less work–life balance but also that their work–life balance is supported in countries with dual-earner, dual-carer family policies (such as ECEC, but also paid parental leave).

### **Parental Leave**

Parental leave policies ensure that parents can take leave from their employment around the time of the birth of their child(ren) and during the early years of their childhood, while guaranteeing that parents can return to their job after the leave and still receive some level of continued wage payment or wage replacement (Gornick & Meyers, 2003). As such, it is a policy that inherently promotes and protects parents' attachment to the labor market and can be an effective strategy to reduce poverty. In a comparative study of 18 OECD countries, leave was found to reduce poverty by facilitating employment—more effectively so among single parents than two-parent families (Maldonado & Nieuwenhuis, 2015)—so that a single parent may take time off to care for a sick child and return to work without fear of losing the job. However, parental leave had to be matched with some level of wage replacement, as unpaid parental leave did not reduce poverty. Moreover, paid parental leave was found not only to elevate single parents out of poverty but also to increase the likelihood of achieving a middle-class income (Byun, 2018).

As is the case with many family policies, the design of the parental leave policy matters to determine its effectiveness and possible trade-offs. These tradeoffs are discussed in great detail in Chapter 11 by Hook and Li in this volume, so here it suffices to focus on one such tradeoff particular to single parents. It has often been documented that overly long periods of parental leave are a mechanism of exclusion of women from the labor market (Nieuwenhuis, Need, & Van der Kolk, 2017; Pettit & Hook, 2009). Misra, Moller, and Budig (2007) showed that a number of different family policies reduced poverty among single-parent families, although they cautioned that overly lengthy parental leave may have unintended consequences for single mothers and the labor market. Van Lancker (2018) showed that leave was associated with a larger probability of single parents being employed in countries with moderate levels of leave. Yet, in countries with very long durations of leave, having used parental leave was associated with a lower probability of being employed. Van Lancker thus confirmed that many of the tradeoffs associated with overly long parental leave also apply to single mothers.

A life-course perspective helps understand how paid leave policies can be effective for single parents later in life, even when they are separated from their partner (becoming a single parent) years after they last qualified for taking parental leave. Bernardi and Mortelmans (2018) described the trajectories of single parenthood and their poverty, work, health, migration, and how these vary across the life course and are diverse across families and countries. Zagel (2014) detailed the various employment trajectories of single mothers. Single mothers' employment state includes full time, employmentoriented, part time, part time return, gradual return, causally employed, and inactive. Zagel (2014) explained that single parents enter and leave single parenthood at very different points in time. There are multiple pathways to being a single parent, such as divorce, separation, death of a partner, and choice such as birth/adoption of a child by a single parent. In sum, the parental leave used by parents in couples, thus promoting gender equality in terms of labor force attachment and the accumulation of work experiences, resonates later in the life course—including among those who have become a single parent.

Recently, there has been greater attention directed to fathers' involvement in care work. Paid paternity leave is associated with higher participation of fathers in childcare and better school performance among children (Nepomnyaschy & Waldfogel, 2007). The positive effect of paternal leave is also associated with greater participation in care for their child(ren) later in life (Duvander & Jans, 2009). In Sweden, parental leave rights are individualized and are mostly non-transferable, which encourages fathers to take substantial amounts of leave. As parents can use their 8 months of (paid) parental leave rights until their child reaches 8 years of age, even fathers who are separated from the mother of their child(ren) can continue to take leave (Duvander & Korsell, 2018). Fransson, Brolin Lafman, Ostberg, Bergstrom, and Olsen (2018) showed that in Sweden, it is increasingly common that children from separated parents live about equal amounts of time with both parents. Such forms of shared residence currently are a hot topic of ongoing investigation, but studies across a variety of contexts seem to indicate that shared residence is beneficial to the well-being of children (Baude, Pearson, & Drapeau, 2016; Nielsen, 2014).

# **Empirical Analysis**

Following the literature review above, this section utilizes various data infrastructures to illustrate the extent to which family policies benefit single-parent families more, less, or to same extent as two-parent families. Again, we focus on child income support (with a focus on family benefits), ECEC, and parental leave policies.

## **Child Income Support: Family Benefits**

When it comes to analyzing how family policies differently benefit single parents versus couples with children, family benefits are probably the best covered in common data infrastructures. We highlight the Luxembourg Income Study (LIS) database to illustrate an empirical example of, and data from the Social Policy Indicator Database (SPIN) on child benefits and outof-work benefits from a social rights perspective. It should be noted that particularly with respect to family benefits, a wide range of other data infrastructures are available, such as EUROMOD (Sutherland & Figari, 2013)—a tax-benefit microsimulation model for the European Union, that comes with the Hypothetical Household Tool (HHoT) for model household type based analyses of social rights and the CSB Minimum Income Protection Indicators Database (Van Mechelen, Marchal, Goedemé, Marx, & Cantillon, 2011) that includes data on minimum income protection provisions in Europe and the United States over two decades.

The LIS database harmonizes pre-existing micro-data to a common template in order to maximize comparability across countries, providing a wide range of variables on persons and households. These data focus on various sources of income (and taxes paid), including different components pertaining to redistributive social policies such as family benefits. Other sets of variables pertain to demographics, employment, and household composition; the latter makes it possible to use LIS to examine the impacts of family benefits on single-parent and couple-parent families.

In a common type of policy analysis with LIS, researchers aim to quantify the redistributive impacts of welfare state redistribution (whole or part of a specific policy) on a measure of income, inequality, or poverty (Chzhen & Bradshaw, 2012; Gornick & Jäntti, 2012; Gornick & Smeeding, 2018). Here, we examine the impacts of family benefits on single parents and couples with children who are at risk of poverty. To this end, poverty rates are calculated based on two income concepts: first, based on disposable household income (after all benefits are received and taxes are paid); and second, based on disposable household income, minus the amount of family benefits that the family has received. This latter poverty rate represents how high poverty would have been without family benefits. The difference between the two poverty rates can be attributed to family benefits. This type of decomposition is not without limitations, as it is assumed that people do not change their behavior in the absence of certain benefits, and when the analyses are performed for multiple benefits, it remains unclear which of the benefits lifted the household out of poverty (Nelson, 2004). Nonetheless, the type of decomposition as described above indicates the capacity of family benefits to lift different families out of poverty.

An example of this type of decomposition analysis is presented in Fig. 13.1, for single-parent families and couples with children separately. The green bars indicate the at-risk-of-poverty rates based on disposable household income, and the orange bars indicate how high poverty would have been without family benefits. The length of the orange bar can be interpreted as the redistributive impact of family benefits on poverty.

A few findings stand out in Fig. 13.1. First, single parents face poverty risks that are substantially higher compared to couples with children, with considerable differences among countries. For instance, single parents' poverty is lower in the Nordic countries than in Canada, Luxembourg, or the United

#### **312 L. C. Maldonado and R. Nieuwenhuis**

**Fig. 13.1** Family benefits are associated with lower poverty for single-parent and coupled-parent families (*Source* LIS data)

States. Second, without family benefits, poverty among single parents would be substantially higher in most countries. This can be seen by the length of the orange sections of bars. Third, again in most countries, family benefits are associated with larger poverty reductions among single parents than among couples with children (indicated by the larger orange bars among the former).

At first glance, Fig. 13.1 suggests that family benefits reduce poverty risks for all families and more strongly among single parents. However, these analyses are in fact not informative about why this is the case. It can, for instance, not be distinguished whether single parents benefit more because they had incomes closer to the poverty line (and therefore require smaller family benefit amounts to be lifted out of poverty) or because they receive higher amounts. Therefore, Fig. 13.2 explores this poverty threshold inquiry further.

Figure 13.2 shows for poor families just how far their income falls short of the poverty line. For the calculation of poverty gaps, we used the poverty line and household income before family benefits were included. The poverty gaps are represented as a percentage of national poverty. The results suggest that there is no consistent pattern among countries that show poor single parents are closer to the poverty line than poor parents in couples. In fact, in more than half the countries, poor single parents are further behind the poverty line compared to poor couples with children.

**Fig. 13.2** Poverty gaps among poor single parents are not systematically larger than poverty gaps among poor couples with children (*Source* LIS Data)

For instance, in Denmark, the income of single parents in poverty is, on average, 20% below the Danish poverty line, whereas among couples with children, it is about 40% below the poverty line. It is important to point out that single parents in Denmark had a greater likelihood to be poor than couples with children (Fig. 13.1), but this indicator (Fig. 13.2) shows that among the poor, couples with children in Denmark tend to be worse off. Therefore, this evidence shows it is not the case that single parents are closer to the poverty line but instead shows why family benefits are associated with larger poverty reductions among single parents.

So far, Figs. 12.1 and 12.2 build a strong case that higher family benefits reduce poverty. However, these illustrations do not fully capture whether or not higher benefits reduce poverty more among single-parent families. Therefore, in Fig. 13.3, we present the association between how family benefits are and their poverty reduction about single parents and couples with children.

In Fig. 13.3, countries are positioned on the x-axis according to the amount of family benefits that households have received (equivalized for household size). For instance, in Australia and the United Kingdom, single parents (18–22% of the poverty line) received substantially more than couples with children (around 10%), whereas in Estonia and Luxembourg, couples with children received higher family benefits than single parents. The y-axis

**Fig. 13.3** Higher family benefits reduce poverty more, in particular, among singleparent families (*Source* LIS Data)

shows the poverty reduction associated with family benefits, which corresponds to the length of the orange bars in Fig. 13.1. Higher family benefits tend to reduce poverty to a larger extent, and the association between the levels of family benefits and the degree of poverty reduction is stronger among single parents than among couples with children. In other words, a given level of family benefits lifts more single-parent families out of poverty than couples with children. There are several explanations for this finding. First and foremost, single parents have higher poverty rates, which mean that a larger proportion of this group can be lifted out of poverty. A second plausible explanation is the degree of targeting. High degrees of targeting may be associated with lower levels of benefits overall (Korpi & Palme, 1998), as a given level of family benefits targeting may be associated with greater poverty reductions (Van Lancker et al., 2015). In an analysis of the institutional design of family benefits (Marchal & Van Lancker, 2019), it was found that many countries (e.g., the United Kingdom and the United States) target family benefits heavily toward lower incomes, although there was no clear pattern that this degree of targeting is stronger among single parents than among couples with children. However, as single parents are more likely to have a low income and be in poverty (Fig. 13.1), the degree of low-income targeting contributes to the explanation of why more single parents than couples with children are lifted out of poverty at a given level of family benefits.

Next, we examine family benefits and unemployment insurance using model family types (for a discussion of the use of model family types as compared to observational data see Chapter 24 by Sirén, Doctrinal, Nieuwenhuis, and Van Lancker, in this volume). In many countries, couples with children receive higher benefits than single parents (cf. Fig. 13.3). However, it cannot be inferred that the family benefit policies were designed to award couples with children higher family benefits. Take for instance, the levels of family benefits that are dependent on the number of children in families. So even if the policy is designed to award a universal amount of family benefits per child—irrespective of income and family composition—couples might receive more benefits if they have more children. Such differences in socioeconomic and demographic makeup of families (often referred to as "policy demand") obfuscate analyses of micro-level data that seek to infer what the policies intend to provide to families (often referred to as "policy supply").1 A commonly applied technique to isolate the design to better understand intentions of policy is to analyze model families (Bradshaw, Ditch, Holmes, & Whiteford, 1993), which is a set of narrowly defined family compositions, for the social policy rights are calculated across contexts. As the families are kept identical across contexts, differences in which family benefits are entitled to can only be attributed to differences in how the policies are implemented, not to differences in the socio-economic composition (the "policy demand") of these families.

The SPIN database is a long-standing data infrastructure on policy indicators (Ferrarini, Nelson, Korpi, & Palme, 2013). With a focus on social security, SPIN is comprised of several databases, including the Out-of-Work Benefits Dataset, the Child Benefit Dataset, the Parental Leave Benefit Dataset, the Social Assistance and Minimum Income Protection Interim Dataset, the Social Citizenship Indicator Program, and the Social Insurance Entitlements Dataset. These databases are based on the social rights perspective (cf. Marshall, 1950), using the method of model families. As described above, such indicators capture what policies intend to do (e.g., what rights citizens can expect) rather than what citizens do with these policies (e.g., take up). As such, these indicators can be used to analyze changing welfare states as well as different welfare state outcomes across countries or over time. Many of the databases in SPIN have model families representing both couples with children as well as single-parent families, with the exception of the Child

<sup>1</sup>For more on this issue, see Chapter 24 by Doctrinal, Sirén, Nieuwenhuis, and Van Lancker, in this volume, on childcare policy indicators.

Benefit database (for which the source data did not provide information on single parents). Nonetheless, indicators on income support such as child and family benefits are included in several other databases.

In Fig. 13.4, we present results from the Out of Work Benefits database (Doctrinal, Nelson, & Sirén, 2015). The model families are assumed to have worked but then became unemployed and therefore qualify for insurancebased unemployment benefits. In insurance-based benefits no differences in benefit levels are to be expected between family forms—in contrast to means-tested, social assistance-based benefits. Yet, the income situation of different family types may differ in times of unemployment, for instance, due to additional transfer such as from family benefits. Both the single parents and the couples are assumed to have two dependent children. The benefits these families receive are calculated at different wage levels (referring to their employment before becoming involuntarily unemployed), ranging from 50 to 200% of average wage. The indicators in Fig. 13.4 represent the average of these calculations at different wage levels. The green bars represent the level of

**Fig. 13.4** Family benefits remain an important source of income, even after accounting for unemployment benefits, more so for single-parent families (*Source* Out of Work Benefits Database)

unemployment benefits these families receive, and the orange bars represent the unemployment benefits topped with family benefits.<sup>2</sup>

In comparing single parents to couples with children, it is observed that in most countries, their income position is similar to that of couples with children but that their income relies to a larger extent on family benefits. This is particularly the case in Australia, Ireland, and the United Kingdom. It should also be noted that for these model families who have been working full-time but lost their jobs, insurance-based unemployment benefits are a more important part of their income. Nonetheless, family benefits represent an important addition to their income to provide financial support for their families. It should also be noted that in contemporary dual-earner societies, many couples (with and without children) have two incomes. An income replacement of substantially less than 100% during unemployment is a harder burden for a single parent without a second earner. This has been identified as a major risk factor for single-parent poverty (Alm, Nelson, & Nieuwenhuis, 2020).

Figures 13.1 through 13.3 support that higher amounts of family benefits reduce poverty for families with children. Figure 13.3 further shows that single parents have a steeper slope and perhaps benefit more. Figure 13.4 accounts for socio-demographic characteristics through the use of model family types and examines unemployment benefits topped up with family benefits. The findings suggest that family benefits are an important source of income, but that in particular for single parents, family benefits represent a substantial top-up in a large number of countries.

#### **Early Childhood Education and Care (ECEC)**

ECEC is an important policy measure for the work–family reconciliation of all families and is crucially important for single-parent families. Here, we focus on an indicator of ECEC affordability, for both single parents and couples with children. These data, presented in Fig. 13.5, are from the OECD Family Database,<sup>3</sup> based on calculations using the OECD Tax and Benefit Models 2015,<sup>4</sup> by model household types. This ensures comparability across contexts, and this particular example allows for an important comparison of

<sup>2</sup>It is not possible to disentangle these amounts in all countries, as the level of net benefits received from either unemployment or family benefits can be adjusted to what other income (from benefits or otherwise) these families receive. Nonetheless, the indicators represent the share of family benefits that make up their income.

<sup>3</sup>www.oecd.org/els/family/database.htm.

<sup>4</sup>http://www.oecd.org/social/benefits-and-wages/.

**Fig. 13.5** Single parents do less in ECEC than couples with children; however, single parents pay a larger share of their household income (*Source* OECD Family database)

single-parent families and couples with children. The single parent is assumed to work full-time at the national average wage, whereas the couple consists of a full-time worker and one working at least 67% (both at the national average wage). Both families have two children (aged 2 and 3), who use childcare full-time in a typical day-care center.

Panel A in Fig. 13.5 shows the out-of-pocket costs for ECEC, separately for single parents and couples with children. These are nominal amounts, relative to the average wage in a country. This shows that, generally, the single-parent family pays less for childcare (in absolute terms) compared to the couple. This might be the case because the single-parent household (as defined above) has a lower disposable household income. In Sweden, the childcare fees are a percentage of household income. In Luxembourg, single parents receive a higher benefit/compensation for the childcare costs. Panel B shows the costs of families' childcare expressed as a percentage of their disposable household income; thus, a dual-earner family can more easily carry these costs. Indeed, here it is shown that in the majority of countries, single parents pay a larger (and sometimes substantially larger) share of their disposable household income for the same amount of childcare as couples with children. In Ireland, single parents pay more than 40% of their disposable household income, whereas in the United States, single parents pay nearly 60%.

Together, these graphs suggest that single parents tend to pay less for childcare compared to couples with children, in part because of lower childcare fees and because of higher benefits/compensation. Yet, as single parents lack a second earner in the household, these costs impose a larger burden on their disposable household income—in some countries, substantially more so.

#### **Parental Leave**

There are a number of data infrastructures on parental leave legislation and benefits. However, there are no indicators that quantify separately the social rights of single parents versus couples with children. Only in one case such indicators are currently under development.

The International Network on Leave Policies and Research presents an annual review of leave policies, and the 2019 edition (Koslowski, Blum, Dobroti´c, Macht, & Moss, 2019) covered 45 countries, including countryspecific reports and country-comparative indicators on maternity, paternity, parental leave, and other leave measures. Durations of leave, paid leave, and well-paid leave are reported, and conditions for flexibility are systematically indicated. However, it is not systematically indicated whether or not single parents have specific rights, other than in the notes of countryspecific situations. For instance, it is reported that unpaid childcare leave has an extended duration for single parents in Bulgaria, leave to care for sick children in Chile is transferable to the father in case of the mother's death, and single mothers in Finland have the right to paternity benefit days. Although this is valuable information, uncoded data do not lend themselves well for quantitative, comparative analysis. Similarly, the Mutual Information System of Social Protection<sup>5</sup> provides qualitative descriptions of leave policies that sometimes highlights specific rights for single parents but does not systematically quantify these rights. The OECD Family Database does provide quantified indicators on leave policies, including the duration and pay levels, information on users of leave, additional leave entitlements for working parents, and replacement rates that are based on the model family types, but the information is not yet differentiated to single parents and couples with children.

The Parental Leave Benefits database, part of the SPIN infrastructure introduced above, is currently developing indicators that explicitly focus on

<sup>5</sup>https://www.missoc.org.

#### **320 L. C. Maldonado and R. Nieuwenhuis**

**Fig. 13.6** Duration of parental leave varies more between countries than between family types (*Source* Parental Leave Benefits database [preliminary data])

single-parent families.<sup>6</sup> Using model families, a number of aspects of these leave policies are coded in a way that is comparable to couples with children. This includes the duration and wage replacement rates during maternity, paternity, and childcare leave, as well as other income support benefits directly related to early parenthood. For some countries, preliminary results are presented below. It should be noted that these results should be cautiously interpreted. Nonetheless, some interesting patterns emerge.

Figure 13.6 shows the durations of paid leave in 13 countries for three model family types: a single-parent family, a single-earner couple, and a dualearner couple. The durations include wage replacement and are for maternity leave, parental leave, paternity, and/or childcare leave. Unpaid periods of leave are not included. In the single-earner couple it is the person who was employed who takes the leave, and the person who was not working does not take any leave (even if it would be available to them). The dual-earner couple is assumed to take their leave in a way that maximizes the duration of their leave, therefore benefitting from additional durations of leave awarded

<sup>6</sup>We in particular thank Sofie Burman for her hard work on making these indicators available to us, as well as Pär Dalén, and take any responsibility ourselves for the fact that these are preliminary results subject to change.

when the leave is taken equally, and taking their leave sequentially (rather than together at the same time) when allowed.

The results show that by and large, each of these family types receives the same or very similar durations of leave rights in most countries, and that differences between countries vastly exceed differences between family types. In Finland and France, single parents receive shorter leave compared to the dual-earner couple, because in in these countries parental leave for fathers is not transferable to single parents. Slovakia provides three weeks of extended leave for single mothers with newborns. In a few other countries, including Norway, Germany, and Sweden, the single-earner couple actually receives shorter leaves than the single parents or dual-earner couple: in these countries the parental leave scheme is set up to encourage gender-equal taking up of parental leave, but do not penalize single parents in case they are not able to share the parental leave equally.

At first glance, the duration of parental leave does not necessarily benefit single-parent families more or less. Therefore, Fig. 13.7 takes a different approach. It does not show the duration of paid leave but the total income replacement (related to having a young child) that different families would

**Fig. 13.7** Income replacement of full-year parental leave, single parents receive slightly more; however, leave varies more between countries than between family types (*Source* Parental Leave Benefits database [preliminary data])

receive. The replacement rate is here defined as the total household income while a parent is taking parental leave as a percentage of the income before taking leave—assuming that everything else remained the same. It is furthermore assumed that each worker has an average wage, and that the families choose to be on leave, or otherwise not on the job, for a full year—even if they are entitled to a shorter period of leave. The calculation of income replacement for a full year in each country provides a comparable indication of how flexible families can be in choosing their own duration of leave. It is calculated how much their total income differs from their average wage prior to the year on leave. In these calculations, wage replacement during parental leave is accounted for, as well as benefits related directly to early parenthood. This includes child benefits, tax credits, and to the extent that the government guarantees child support payment to the single parent if the "other parent" does not pay. For an extensive review on child support, see Chapter 12 by Skinner and Hakovirta in this volume.

The duration of parental leave captures the degree to which a family can stay home to provide full-time care for their newborn without having to rely on additional benefits such as social assistance. Similar to the out-of-work benefits shown in Fig. 13.4, the couple with children can have a second earner in the household, whereas this is typically not the case for single parents. As such, single parents would have to rely solely on these levels of income to avoid claiming other benefits such as social assistance, whereas couples with young children may have other sources of income to compensate for the income loss during parental leave. The impact of this is represented by the distinction among the orange bars (couple with children, single earner) that show the rate of wage replacement only for the person on leave. The purple bars (couple with children, dual earner) show the effective level of income a dual-earner family has when one parent goes on full-time leave for a year while the other parent remains working full-time at the average wage.

Figure 13.7 confirms the earlier results that the differences among countries seem to be more substantial than the difference between single-parent families (green bars) and single earner couples (orange bars). In most countries, single parents receive slightly more than the single-earner couple—in part resonating the finding in Fig. 13.6 that many countries incentivize gender-equal sharing of the parental leave, and in part due to the state providing guaranteed payment of child support. The United Kingdom, for instance, had a fairly average duration of parental leave matched with low levels of income replacement and targeted child benefits for low-income parents. Thus, parents with an average wage end up with a low level of income replacement if parents go on leave for the full year.

The small differences between the green and orange bars show that most countries do not provide very different entitlements to people taking leave based on their family composition. However, in terms of household income during leave, it matters greatly whether there is a second earner in the household—something single parents often do not have access to. This is demonstrated by the purple bars, which show the effective income of a couple where one partner goes on full-time leave and the other continues full-time employment. Naturally, their income is substantially higher than that of a single earner going on leave. In countries with low replacement rates during parental leave, the income position of single parents falls substantially behind that of dual-earner couples, not because they receive a substantially lower replacement of their own wage but because they do not have a second earner in the household. In other words, although the policy provisions are fairly similar across family types, how these policies play out results in very different situations for families.

## **Conclusion**

In societies with increasingly diverse family structures, it is important to address uniqueness of families while also providing universal family policies to all. The evidence is clear that all families with children benefit from family policies that include child income support, ECEC, and paid parental leave, and that these benefits are extended to single parents. These policies help to level the playing field for all families, and in numerous cases, it was shown that single parents actually benefit more than couples with children.

For child income support—and in particular, family benefits—the results are quite clear. The evidence is strong that family benefits do in fact lower poverty for all families and especially for single-parent families. Family benefits are more effective for single parents, not because they are closer to the poverty line but because they receive higher amounts of family benefits. In addition, because of economies of scale, even receiving the same amount as two-parent families would benefit single-parent families more. The differences among countries suggest that family benefits significantly reduce poverty for couple-parent families more so than for single-parent families. The only two exceptions, Luxembourg and Estonia whose family policies benefit those families with the most resources as opposed to families that are resource poor.

ECEC policies are perhaps the most important in facilitating mothers' employment (Olivetti & Petrongolo, 2017). The literature review clearly showed how childcare provisions facilitate single parents with young children to be employed, improve work–life balance, have health benefits, and even reach into the middle class. The empirical results, however, showed that the ECEC costs in many countries represent a larger share of their household budget compared to dual-earner couples.

The literature review further indicated that paid parental leave—if not overly long—can be beneficial to all families with children. Although there is some evidence that parental leave benefits single parents slightly more, the mechanism does not seem to be based on single parents receiving longer periods of leave or substantially higher wage replacements. Parental leave entitlements clearly differed more among countries than among family types. Thus, the findings in the literature that single parents still benefit more from parental leave than people in two-parent households might not be due to differentiated leave entitlements but rather to the absence of a second earner and care giver in the household. Yet, the empirical results suggested that particularly in countries with low replacement rates during parental leave, the income position of single parents on leave is substantially worse than among dual-earner parents.

It is highly conventional to conclude a chapter like this by calling for better family policy indicators to facilitate future research. We leave most of that work to the chapters in the final section of this Handbook—in particular, Chapter 24 by Sirén, Doctrinal, Nieuwenhuis, and Van Lancker in this volume. There is, however, one important finding that requires more attention here. In both the data on ECEC and on parental leave, we found that although the provisions to single parents and couples with children were highly similar, single parents were comparatively worse off. In dual-earner societies, single parents find it challenging to pay even a slightly reduced childcare fee or be on a single-wage replacement during leave. Paradoxically, we also found support in the literature that despite the discrepancies between their entitlements regarding ECEC and leave policies, single parents use these policies to the same extent as couples (e.g., Van Lancker, 2018), and in some cases, to greater effects in terms of employment and poverty reduction. To resolve this paradox requires more family policy indicators that examine in detail the entitlements provided to different family types.

However, this finding not only represents a challenge in data collection but also a true challenge to policy makers, related to selectivity within universalism (cf. Van Lancker & Van Mechelen, 2015). If equal provisions of the policy produce unequal living conditions, this raises questions of horizontal and vertical equity, as well as the distinction between the equality of policy provision and the equality of policy outcomes. Addressing these challenges is ever more pertinent in societies where large shares of dual-earner couples influence what are commonly accepted living standards. In such societies, the challenge is to design equitable family policies with provisions to single parents that are adequate in relation to the provisions and other income sources of couples with children.

This review of the literature and analysis of selected data infrastructures has demonstrated that family policies are indispensable for single parents and their families. Even when the policy provisions do not seem specifically tailored to single parents, they benefit. Even when single parents have to pay a larger share of their household income to ECEC services, it seems they still benefit from the provided care. In a time that family forms have become more diverse, and family relations change more over time and within one's life-course, it is an important lesson that the family policies envisioned by and the principles underlying—the dual earner, dual carer model also benefit the single earner, single carer, single parent.

**Acknowledgements** Nieuwenhuis was financially supported by FORTE: the Swedish Research Council for Health, Working Life and Welfare (grant number 2018-00988).

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **14**

# **Policies for Later-Life Families in a Comparative European Perspective**

**Pearl A. Dykstra and Maja Djundeva**

Discussions of "family friendly" policies tend to focus on young children and their parents, as exemplified in Berger and Carlson's (2020) decade review of family policy in the *Journal of Marriage and Family*. The authors explicitly "retain the focus on families with children" despite acknowledging that "families come in a variety of forms that may or may not involve legal, biological, or intergenerational ties" (p. 479). The focus on young children and their parents is grounded in the principle that families produce and shape citizens of the future. Unfortunately, it disregards the fact that those parents are also children, even grandchildren, in a multigenerational family structure. Writing on public support to the oldest members of society comes under headings such as "pension policy", "health policy", or "long-term care policy"—as if people in later-life phases have no families. The neglect in family policy circles for persons who are helping older relatives is all the more remarkable given that growing numbers of frail older adults rather than growing numbers of children are putting pressure on families' ability to

P. A. Dykstra (B)

Erasmus University Rotterdam, Rotterdam, The Netherlands e-mail: dykstra@essb.eur.nl

M. Djundeva Netherlands Institute for Social Research (SCP), The Hague, The Netherlands e-mail: m.djundeva@scp.nl

<sup>1</sup>https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32019L1158&from=EN.

provide care (Kröger & Yeandle, 2013). That families provide care to *both* young and old is acknowledged in the 2019 European Directive on work-life balance,1 which requires Member States to implement measures aimed at "allowing parents, and other family members with caring responsibilities to enter, remain in, or return to the labour market".

Across OECD countries, around 13% of people aged 50 provide care with everyday tasks to older family members, friends, and neighbors at least weekly (OECD, 2019a). Almost two-thirds of those providing daily care are women, especially spouses or adult daughters (in-law), but more men become carers at older ages. As improvements in survivorship continue, more people will find themselves living longer with *multiple family generations* (Murphy, Martikainen, & Pennec, 2006). Whereas early studies on the "sandwich generation" focused on how women combined caring for frail parents and dependent children (Brody, 1981), later scholarship pointed out that the typical experience of "being caught in the middle" pertained to helping both aging parents and young grandchildren (Soldo, 1996). Furthermore, because of gender differences in age at marriage and life expectancy, women are more likely than men to simultaneously occupy the role of adult child and grandparent (Hagestad & Dykstra, 2016). Research on those who are "sandwiched" between multiple generations, by having at least one parent/parent-in-law and one grandchild alive, goes against the assumption that helping the one generation comes at the expense of helping the other generation: more intense support for one generation is associated with a higher likelihood of supporting the other generation (Herlofson & Brandt, 2019; Vlachantoni, Evandrou, Falkingham, & Gomez-Leon, 2019). These findings support a "solidarity hypothesis" rather than a "competing demands hypothesis" (Grundy & Henretta, 2006): those with strong commitments to solidarity tend to assist both generations rather than prioritize recipients, whereas those low in solidarity are less likely to help both generations.

In this chapter, we consider cross-national differences in policies for laterlife families in Europe, focusing on state support *freeing* family members from caring responsibilities or *enabling* them to care for older generations. These policies come under the umbrella of long-term care (LTC), help required by persons with a reduced degree of functional capacity (whether physical or cognitive) for an extended period of time (European Commission, 2018a). The policies are not only targeted at frail older adults but pertain to adults of all ages with a dependency on others. Publicly funded LTC includes "inkind" services, where care is provided by professionals at home or in an institution, and "cash benefits" which can be used to purchase professional care or which can be paid to informal caregivers as income support. Apart from long-term care for dependents themselves, there are policies supporting family members in their caregiving tasks: "cash for care" (financial compensation for helping those with physical of cognitive impairments), "care leaves" (the right to be absent from work in order to care), and "care credits" (time spent on caring that is credited toward a basic pension). Note that the issue of financial responsibilities, providing money to older family members who have insufficient retirement income, is beyond the scope of this chapter.

# **Models of Long-Term Care**

We start with a general overview of LTC models and adaptations to these models over the past 25 years. These adaptations have taken place in response to concerns about an expected decrease in the pool of family carers and a concomitant growth in demand for LTC, and in response to cultural shifts toward more self-determination for care users.

#### **General Overview**

In Europe, LTC provision is characterized by significant cross-national differences regarding the division of responsibilities between families, the state, for-profit organizations, and the volunteer sector. The differences in LTC models stem largely from cultural and political traditions regarding the role of families in society (Pavolini & Ranci, 2013). Is care primarily a private obligation with the state stepping in only when absolutely necessary? Or, is care a social right, a basic need of citizens? Models of LTC define relations of *generational interdependence* (Dykstra & Hagestad, 2016; Hagestad & Dykstra, 2016): the extent to which public policy arrangements impose reliance on older and younger family members or enable individual autonomy between family generations (Frericks, Jensen, & Pfau-Effinger, 2014; Leitner, 2003; Lohmann & Zagel, 2016; Saraceno & Keck, 2010).

Several LTC models have been identified, ranging from a *residual* model, also termed informal care-led model (Pavolini & Ranci, 2008) or family care model (Anttonen & Sipilä, 1996; Bettio & Plantenga, 2004), to a *universalistic* model (Ranci & Pavolini, 2015), also termed services-led model (Pavolini & Ranci, 2008) or Scandinavian model (Anttonen & Sipilä, 1996), with various *intermediate* models in between. In countries with a residual LTC model (most of the Mediterranean and central-eastern European countries), care for the frail is mainly provided by families, volunteer organizations, and religious associations. In countries with a universalist LTC model (the Nordic countries and the Netherlands), there is generous state funding for nursing care, personal care, and medical help via recognized institutions. Countries with an intermediate LTC model (Austria, France, Germany and the United Kingdom) have made significant financial investments into LTC in the past 25 years, at different times and following diverse institutional and organizational models (Carrera, Pavolini, Ranci, & Sabbatini, 2013).

### **Adaptations to Long-Term Care Models Over the Past 25 Years**

Recent decades have shown a blurring of differences in LTC models as countries have reacted to what scholars identified as "problem pressures" (Ferrera, 2005) and "new social risks" (Taylor-Gooby, 2004). One of the identified *pressures* concerns the supply of family care. International organizations like the OECD repeatedly report possible shortages of available kin to support older adults in the future (e.g. Haberkern, Schmid, Neuberger, & Grignon, 2012). Due to reduced fertility rates and increased divorce rates there may be fewer adult children and spouses to take care of older adults in need. Note however, that having multiple children does not mean that all of them are providing care (Fontaine, Gramain, & Wittwer, 2009). Moreover, increases in longevity imply that higher proportions of future older adults are likely to have a surviving child than any generation ever born (Murphy et al., 2006). If the family's capacity to provide support diminishes in the future, it is less likely the result of changes in fertility and mortality patterns, and more likely connected to changes in family structure (e.g. the increase of singleparent families), the unequal but steady rise in the labor force participation of women, and the changing nature of work which results in less free time, longer commuting, and greater residential distances between family members (Limmer & Schneider, 2008). To support family carers, a limited number of countries have introduced cash benefits that are granted directly to the carer (e.g. Finland, Hungary, Ireland and the United Kingdom) and many countries have leave schemes that allow caring relatives to take some time off from gainful employment or to reduce their working hours (Spasova, Baeten, Coster, Ghailani, Peña-Casas, & Vanhercke, 2018).

Numerous reports have pointed to financial pressures linked with the expansion of an older population in need of care. Notwithstanding a potential compression of morbidity, the numbers of older people with cancer, hip fractures, strokes, and dementia will grow, and many older people will have multi-morbidities (Rechel et al., 2013). Public spending on LTC is projected to increase from 1.6 to 2.9% of GDP in the EU between 2016 and 2070 (European Commission, 2018a). Although older people account for a substantial proportion of long-term care, other factors, especially progress in health sciences and the development and use of new technologies have a much larger effect on aggregate costs (De la Maisonneuve and Oliveira Martins, 2013). It has also been suggested that new generation of older people, who might be wealthier or more educated than were previous generations, will have greater demands for care services (Rechel et al., 2013). Cost containment measures that have been adopted in recent decades in countries with more generous care provisions include the freezing of service levels, shifts from institutional to home-based care, targeting care services to those with the most severe needs, and increased co-payments (Van den Broek, Dykstra, & Van der Veen, 2019).

Changes in the organization of LTC provision can also be traced to cultural shifts emphasizing self-determination and autonomy with regards to care (Genet et al., 2011; Ranci & Pavolini, 2013). The expansion of cash for care measures is at least partially a response to demands by disability groups for freedom of choice in care receipt (Da Roit & Le Bihan, 2010, 2019; Le Bihan, Da Roit, & Sopadzhiyan, 2019). Moreover, the growth of home-based care and the contraction of residential care in countries with universalistic LTC models fits older European's preferences to live in a familiar environment, traditionally the family home, as long as possible and to avoid moving to a form of institutional care (European Commission, 2007).

# **Cross-National Comparisons of Long-Term Care Provisions**

In this section, we describe differences in LTC provisions across European countries, including changes over time. Cross-national comparisons are rather challenging, given differences in definitions of disability and dependency, divisions between government departments and state agencies in the delivery of care, and different methods of financing LTC (European Commission, 2018a). Differences in definitions provide additional complexity. Sometimes publicly funded LTC is used synonymously with "formal care", a broader category that also includes privately paid professional care. Sometimes "informal care" also includes care provided by family members that is partially paid by public funds in the form of cash for care benefits. Thus, merely distinguishing between formal and informal care does not capture the complex policy arrangements that vary greatly across European countries. Further challenges in the comparative investigation of LTC arise from country-specific definitions of long-term care services.

#### **Availability of Beds**

Most countries provide information about beds in residential long-term facilities as a ratio between recipients and the older adult population (recipients per 1000 adults aged 65 and over). Unfortunately, definitions of what constitutes a "residential bed" have been subject to change. For example, in 2017, Austria reclassified large parts of alternative living facilities in residential long-term care facilities as inpatient services (OECD, 2019b). In 2012, the Netherlands expanded the definition of beds in residential long-term care facilities to also include places in care residences for disabled persons and not only nursing and residential care homes for older adults (OECD, 2019c).

Figure 14.1, based on harmonized data compiled by the OECD, reveals changes between 2005 and 2017 for 25 European countries in the availability of beds in residential long-term care facilities per 1000 persons of the population aged 65 and over. In a wide range of countries (Czech Republic, Finland, France, Iceland, the Netherlands, Norway, Poland, Sweden, Switzerland, United Kingdom), the number of beds decreased over time. Some of these reductions have occurred due to countries implementing policies to move LTC out of residential facilities and into the community (Colombo, Llena-Nozal, Mercier, & Tjadens, 2011). Deinstitutionalization is not a problem per se but becomes one when it is not matched with a sufficient and affordable increase in home care and community-care provision (Spasova et al., 2018). Germany, Italy, Lithuania, and Slovakia, and more so Estonia and Spain show modest increases in the availability of residential long-term care.

**Fig. 14.1** Beds in residential long-term care per thousand of the total population aged 65 and over, selected European countries, 2005–2107

#### **Long-Term Care at Home**

The OECD also has harmonized data on developments between 2005 and 2017 in the receipt of long-term care at home (see Fig. 14.2). Data for some countries refer only to people receiving publicly funded care, while other countries include people who are paying for their own care (OECD, 2017). In 2017, the proportion of over-65s receiving long-term care at home varied from 1% in Portugal to 16% in Switzerland. The proportion of LTC recipients living at home declined over the past decade in Denmark, Finland, the Netherlands, and Norway—countries with a universalistic LTC model. The decrease of home-based care in these countries is linked to fiscal measures aimed at cost containment and greater means-testing of services (Spasova et al., 2018). The decrease in Estonia is attributable to a reduction of the number of "curators" appointed by local government to care for people at home (OECD, 2017). An expansion of home-based care is evident in Italy, Portugal, and Spain—countries where home care services were relatively underdeveloped. In Germany, Hungary and Sweden, the expansion resulted from a deliberate policy to strengthen community care (Spasova et al., 2018). The proportion of older adults receiving long-term care at home also increased in Switzerland, a country where LTC costs are predominantly funded from private sources (Colombo et al., 2011). The proportion of LTC recipients living at home showed no change in France, Luxembourg, and Slovenia, which belong to the group of countries where home care has priority over residential care (Spasova et al., 2018).

**Fig. 14.2** Recipients of long-term care at home as percentage of the total population aged 65 and over, selected European countries, 2005–2107

#### **Cash for Care**

A novelty since the 1990s in long-term care schemes in Europe has been the introduction of cash for care policies (see the left-hand column of Table 14.1 for an overview), but it served different purposes and was elaborated in different ways (Da Roit & Le Bihan, 2019). In some countries (e.g. the United Kingdom), cash benefits were primarily framed as compensation for the costs of disability. In other countries (e.g. the Netherlands), the rationale for introducing cash benefits was to increase users' choice and control, in addition to cost containment. In yet other countries (e.g. Germany), the cash benefit was designed as support for family caregivers. Finally, there is the model of fee for professional service along with creating and regularizing care employment (e.g. Spain). Not surprisingly, given the different rationales underlying their introduction, cash for care schemes differ widely across Europe (Spasova et al., 2018). One difference pertains to eligibility: it can depend on the degree of care dependency, income and assets, and the age of the dependent person. Countries also differ widely regarding the requirements on the use of and accountability for the cash benefit. At one end of the spectrum, the benefit serves as an income supplement for the household without any requirements on how it is spent. At the other end, the benefit is to be used only to pay for professional services and home assistants. Some countries require proof of a formal employment contract. There is also considerable variation in payment levels, which is a function of the roles cash benefits play in each country's LTC program (Nadash, Doty, Mahoney, & Von Schwanenflugel, 2010). As noted earlier, a limited number of countries grant cash benefits directly to the carer (Spasova et al., 2018). Such a program can act to replace lost income, linked to social protection coverage, but can also serve as recognition (albeit often symbolic) of the labor of caring.

#### **Care Leaves**

Leave policies focus on the well-being, labor force attachment, and work-life balance of the carer rather than the person being cared for. All European Union countries, with the exception of Cyprus and Latvia have introduced leave schemes in recent decades (see the middle column of Table 14.1 for details). The leaves are not necessarily only for workers caring for frail older adults but also for workers caring for ill and handicapped adults more broadly. Most countries have both short- and long-term leaves, and they generally allow the carer to continue building up social security rights (Bouget, Spasova, & Vanhercke, 2016). Moreover, job protection is guaranteed during










**14 Policies for Later-Life Families in a Comparative … 347**





the carer's leave. Remuneration also varies: some countries apply a flat rate (e.g. Belgium, Denmark and Hungary), others pay a proportion of previous earnings subject to various ceiling conditions. In Denmark, Ireland, the Netherlands, Norway, Sweden, and the United Kingdom, workers are entitled to leave to look after dependents outside the family circle (Bouget et al., 2016). Unfortunately, information on the take-up or non-take-up of care leaves is often unavailable (Heymann, McNeill, & Earle, 2013). Hence, it is unclear to what extent factors such as non-payment, lack of flexibility or perceived barriers restrain workers from using the leaves to which they are legally entitled.

### **Care Credits**

In late life, the risk of poverty is generally higher for women than for men (European Commission, 2018b). Reasons are women's over-representation in less paid occupations, a lower statutory pension age for women in a number of countries, and women's greater likelihood to engage in part-time work or to have career breaks due to caring activities (D'Addio, 2013). To mitigate pension inequalities, a number of countries provide care credits that count toward a basic state pension, but such credits are more often given as compensation for childcare rather than care for handicapped adults or eldercare (Vlachantoni, 2011). Care credits reflect an amount of time in months/years that is "credited" to the carer's working record as if the carer were employed in the labor market. They do not, however, compensate for wage penalties associated with being outside the labor market (D'Addio, 2013). Care credits are also available to men, but the large majority of recipients are women (European Commission, 2018b). The right-hand column of Table 14.1 shows that fewer than half of EU-28 countries offer pension credits for periods providing unpaid care to adult dependents. It is important to note that that countries like Denmark, Finland, and the Netherlands do not offer credits for family or eldercare because their basic old-age pension is based on years of residence, and hence automatically covers periods spent outside the labor force providing unpaid care (D'Addio, 2013). Care credits are a topic of debate in the policy literature (Foster, Chau, & Yu, 2017). The issue concerns the extent to which they promote women's emancipation or perpetuate existing structures of gender inequality (Ray, Gornick & Schmitt, 2010). One view (espoused by "care feminists") is that care credits are a justified reward for invaluable unpaid activities that generally fall on women. An opposing view (espoused by "employment feminists") is that care credits create disincentives to engage in gainful employment and reinforce traditional assumptions about gender roles.

# **Long-Term Care Policies and Caregiving in Families**

The last part of this chapter focuses on the ways in which LTC arrangements shape caregiving in European families. We describe the division of responsibilities between families and the state, and how this division changes in response to policy changes. We also consider to what extent public arrangements lighten the task of providing care to frail relatives, acknowledging that providing care to an aging family member may lead to costs, for instance, related to lost working days and foregone career opportunities, and to health problems (European Commission, 2018a). We end with a specific focus on unfavorable consequences of cash for care policies. Throughout we consider ways in which the policy context shapes inequality with respect to both class and gender, not only among frail older adults but also among those who care for them.

#### **Specialization Between Families and the State**

Early research on the intersection of families and the welfare state was guided by policy concerns that public provisions would weaken family members' propensity to care for their dependents. By now, cross-sectional findings have repeatedly shown that generous long-term care services complement rather than "crowd out" family care (Motel-Klingebiel, Tesch-Roemer, & Von Kondratowitz, 2005). The availability of social service professionals in a given country shapes the types of supportive tasks that adult children perform for their aging parents (Dykstra, 2018). It is crucial to distinguish practical help (e.g., assistance with household tasks, paperwork) and physical care (e.g., assistance with bathing, dressing, eating) given to parents. The proportion of adult children providing practical help to parents is higher, but the proportion providing physical care is lower in countries with a larger social service sector (e.g. Attias-Donfut, Ogg, & Wolff, 2005; Bonsang, 2007; Brandt, Haberkern, & Szydlik, 2009). There is a "crowding in" of practical help, but a "crowding out" of physical care. When professionals take on the complex, demanding, and routinizable physical care tasks, family members have greater opportunities to provide spontaneous and non-technical forms of help. Hence, professionals and family members specialize in performing caregiving tasks for which they are best equipped (Balia & Brau, 2014; Brandt, 2013; Igel, Brandt, Haberkern, & Szydlik, 2009).

### **Family Caregiving in Response to Changes in Long-Term Care Provision**

The expansion of repeated cross-sectional data sets has enabled research into the impact of changes in LTC provision on exchanges in families. Pickard's (2012) study is rather unique because it considers both the expansion and subsequent retrenchment of institutional care in the United Kingdom. She shows that the increase in residential long-term stay for older people during the late 1980s and early 1990s led to a decline in the most intense types of intergenerational care, but when numbers in nursing homes/hospitals began to fall in the late 1990s, very intense co-resident care by adult children began to rise. The majority of studies have solely focused on the effects of decreases in access to publicly funded long-term care services. In the United Kingdom (Patsios, 2008) and Sweden (Johansson, Sundström, & Hassing, 2003) cutbacks in the 1980s and 1990s in care provided to older adults in the community were accompanied by increases in the provision of care by relatives and in the purchase of private help. Apparently, when the coverage of public services declined, older people turned to their families and to the market. A similar pattern has been observed in Finland, where declining eldercare services since the 1990s have been followed by an increase in family care (Kröger & Leinonen, 2012) and a marketization of social care (Anttonen & Häikiö, 2011). In the Netherlands, stricter eligibility criteria for LTC services introduced in the 2000s have also been accompanied by a rise in care provided by adult children (Van den Broek et al., 2019). We have not found any studies from Southern, Central and Eastern Europe on family care over time.

A number of investigations have revealed that the decrease in public provisions in Sweden affected older people in *different social groups* in different ways: those with more economic resources increasingly bought services on the market, whereas older people with fewer economic resources increasingly received help from family members (Jegermalm & Grassman, 2012; Szebehely & Trydegård, 2012). Another issue has been whether cutbacks in public provisions have differentially affected the help-giving roles of adult sons and daughters (Van den Broek, 2013). Findings are mixed. Focusing on the period 1994–2000 in Sweden, Johansson et al. (2003) found an increase in help by adult daughters but not by adult sons. Ulmanen and Szebehely (2015), whose study covers the period 2002–2010 in Sweden, found an increase in help by adult daughters mainly among older adults with lower education and an increase in help by adult sons mainly among older adults with higher education. Thus, the assistance given by children became more gender equal among older people fewer resources, and less gender equal among those with more resources. In the Netherlands, daughters more often provided household support to parents than did sons between 2002 and 2014, but there was no increase in the gender gap over time (Van den Broek et al., 2019).

## **Policies Mitigating the Risks of Caregiving**

The act of giving is rewarding in the sense of being valued by and being important to others (Batson, 1998). Nevertheless, the provision of unpaid care to dependent family members or friends can be costly—to one's health and to one's financial status. A wide body of research has demonstrated a negative relationship between informal caregiving and well-being outcomes such as depression, stress, self-efficacy, general subjective well-being, and physical health (Pinquart & Sørensen, 2003). Assessing the causal impact of caregiving on mental and physical *health* in a recent review of studies, Bom, Bakx, Schut, and Van Doorslaer (2019) reported that especially female and married caregivers and those providing intensive care experience negative health effects Bom et al. (2019). Studies investigating whether the magnitude of costs to well-being depends on the policy context are starting to emerge. Verbakel (2014) shows that the negative relationship between caregiving and happiness was smaller in European countries that provide more generous public LTC resources, and greater in those with few LTC provisions. Interestingly, the gap in happiness between caregivers and non-caregivers did not vary by level of services offered to informal caregivers, such as leaves, cash benefits, flexible work hours, counseling, and respite care. Rather crude measures of support services might be the reason why no effect was found. Verbakel suggests that future work should measure services more precisely, and determine which types help, under which conditions, for which groups of caregivers. Using data collected between 2004 and 2015, Van den Broek and Grundy (2018) examined the influence of declines in LTC coverage on caregiver quality of life in Denmark and Sweden. Both countries traditionally had generous LTC coverage, but cutbacks were implemented in the 1990s in Sweden and after 2005 in Denmark. Over time, the difference between Denmark and Sweden in the magnitude of the negative impact of caregiving on quality of life lessened. Presumably, caregiving was more strongly perceived as a matter of choice in Denmark at the start of the period under examination, and less strongly so at the end.

A large part of the *financial* costs of caring for frail family members, a role that is more often adopted by women than men (Eurofound, 2016), derive from temporary or permanent detachment from the labor force. Women's greater responsibilities for caregiving influence their labor supply decisions in ways that reduce earnings and make them less attractive to employers (Folbre, 2018). Care credits partially help to compensate the loss of pension benefits as the result of interruptions of the employment career to provide care to family members (European Commission, 2018b). Most research has focused on how LTC policy arrangements might influence the labor force participation of family carers. Consistent with the hypothesis that women are more likely to give up work if there is no viable alternative to family care, Kotsadam (2011) found that the effects of caregiving on women's labor force participation were more negative in Southern Europe (Portugal, Spain, Greece, Italy), less negative in Nordic countries (Denmark, Finland), and in between these extremes in Continental Europe (Netherlands, Germany, Belgium, France). Contrary to Kotsadam, who did not include actual measures of policies in his analysis, Naldini, Pavolini, and Solera (2016) incorporated indicators of home care and residential facilities in addition to total public spending on LTC in their comparison of women's labor force participation in 21 European countries. Their findings show that women's attachment to the labor force was stronger in countries with generous state support in the form of home care or residential homes. Total expenditure on LTC did not make a difference, suggesting according to the authors, that the *type of policy* rather than the total effort is a crucial determinant of carers' employment career. Services such as home help and institutional facilities enable carers to be gainfully employed, whereas cash for care schemes encourage carers, particularly those with lower levels of education, to give up work by providing an alternative source of income (Frericks et al., 2014; Leitner, 2003; Saraceno, 2010).

#### **Unfavorable Consequences of Cash for Care Schemes**

Studies on the impact of the type of LTC policy on caregiving in families are starting to emerge, with a specific focus on unfavorable consequences of cash for care schemes. One of them is increased *gender inequality* in intergenerational care (Da Roit, Hoogenboom, & Weicht, 2015; Pavolini & Ranci, 2008). Intended to enable choice in care receipt and to support the activities of informal carers, evidence suggests that cash for care benefits subtly incentivize women to fall back on traditional divisions of roles. Using data from 14 European countries, and confirming earlier findings, Haberkern, Schmid, and Szydlik (2015) show that women were more likely to provide intensive care to aging parents than men are. However, the gender gap in the provision of such care was highest in countries with low provision of professional home care services and high public spending on cash benefits. Additional analyses revealed that professional home care services substituted only for care by daughters, not for care by sons, who showed lower levels of engagement generally. Moreover, cash payments encouraged intergenerational care, but motivated only daughters not sons. Apparently, public services (home help and home nursing) reduced inequality in intergenerational care by reducing the engagement of daughters, whereas cash for care payments increased inequality in intergenerational care by increasing the engagement of daughters. In general, caregiving by sons was hardly influenced by social care policies. Another unfavorable consequence of cash for care schemes, particularly when users can freely spend their benefits, is *unregulated marketization* of care (Lutz & Palenga-Möllenbeck, 2010; Saraceno & Keck, 2010). In Italy, for example, families have increasingly resorted to often undocumented low-paid migrant workers providing around the clock care (Da Roit & Weicht, 2013), a development that is facilitated by a considerable level of undocumented migration and a large underground economy.

## **Conclusion**

The 2017 European Pillar of Social Rights2 lists access to "affordable longterm care services of good quality, in particular home care and communitybased services" as one of its twenty core principles. To what extent do European countries guarantee their aging citizens this right to long-term care? Our overview of developments since the 1990s in LTC systems across Europe revealed "*limited convergence*" (Ranci & Pavolini, 2013, p. 312): while universalistic systems retrenched their provisions, most of the residual care regimes expanded theirs. The exception is Italy, which undertook no major reform in its LTC policies and by default uses cash for care schemes (Costa, 2013; Da Roit & Le Bihan, 2019). Our review has also revealed that, notwithstanding the "limited convergence", several countries in Europe, particularly in Southern and Eastern regions, do not ensure that their aging citizens have access to timely and affordable long-term care of appropriate quality. In

<sup>2</sup>https://ec.europa.eu/commission/priorities/deeper-and-fairer-economic-and-monetary-union/eur opean-pillar-social-rights\_en

these countries, the more affluent can purchase care services at market price, whereas poorer people have few other options than to turn to their families.

Across Europe, the broad changes in long-term care provision have involved shifts toward more home care, more cash for care, less residential care, and greater targeting to those with the most severe needs (Ranci & Pavolini, 2015). In countries with universalistic systems, the driving forces were not only cost containment but also served the purpose of meeting demands for free choice and consumer direction. In countries with residual care regimes, the reforms offered new entitlements but were also aimed at supporting the caring role of families. Throughout Europe, there has been a trend toward *re*-*familialization* of care, that is, shifting responsibility for long-term care from the state to individuals and their families (Ranci & Pavolini, 2013). Both "passive" (i.e. withdrawal by the state) and "active" (i.e. introduction of cash for care benefits) re-familialization have occurred (Leibetseder, Anttonen, Øverbye, Pace, & Vabo, 2017). In addition, there has been a trend toward *marketization* of care, where those in need of long-term care receive publicly funded services from private providers or pay for services out-of-pocket, with some financial compensation through tax rebates (Ranci & Pavolini, 2013). Our review has revealed that re-familialization and marketization bring the risk of a *dualization* of care (Szebehely & Meagher, 2018), where high-resource older adults find the best providers and low-resource older adults are faced with declining public service coverage.

As noted by Colombo and colleagues (2011), cash benefits should not be regarded as the sole policy option to support family carers. There is a trade-off between financial incentives for family caring and possible inappropriate use of cash benefits or the emergence of unregulated gray labor markets. There is also the risk of trapping family carers (predominantly women) into low-paid roles with few incentives for participating in the labor market. Services to support family carers are also needed, such as the provision of information, basic training, work reconciliation measures, and flexible respite options.

The shift toward more home care and less residential care fits efforts to enable older people to "age in place" (Lawton, 1982): to live independently in their own homes for as long as possible. There is a crucial distinction, however, between "ageing in place" and simply "*staying put* " (Boldy, Grenade, Lewin, Karol, & Burton, 2011). Services must be available to enable older people to live in their own "place". Moreover, for those facing poor housing conditions, the home is not an appropriate environment to "age in place". Coping at home for too long can result in great harm, leading to physical and mental exhaustion for both the older people and their carers (Horner & Boldy, 2008). Coordination between multiple care providers is necessary to avoid that older adults living in the community fail to be noticed or assisted.

Research on the ways in which state support frees family members from caring responsibilities or enables them to care for older generations has made great strides. Nevertheless, while covering the literature, we have identified areas requiring further study. The first is how family policies might reduce the gendered division of caregiving. The strong focus on women in research on the intersecting spheres of work and family overlooks the question of what combination of care-friendly policy could, even over a long-term, decrease gender inequality. Research is needed on the processes by which *men* become more involved in caregiving, and how countries can provide incentives. Part of this research requires theoretical specification of the connections between public provisions (or their absence) and expectations, obligations, rights, and vulnerabilities in the gendered family realm. Another part involves more critical empirical assessments of theoretical mechanisms. Yet another part concerns the uncertain relationship between policies for gender equality and policies that support family care (Oliker, 2011). There is a need for careful investigations of politics and policy, interrogating the tensions between gender equality in labor market participation and gender equality in care work.

Second, there is a need for comprehensive cost/benefit analyses that *cut across* policy domains: increased funding in one budget area might have savings elsewhere, or, cutbacks in one budget area might lead to problems elsewhere. One of the research gaps concerns the implications of investments in home care, aimed at enabling "ageing in place", for the health and well-being of older persons and their carers. Another concerns the trade-offs between encouraging people to remain in work longer, continued reliance on families for the provision of long-term care and an expanding role for grandparents in caring for their grandchildren. These cross-cutting research topics do not allow for easy solutions—but with increasing access to national registry data, European countries offer a unique laboratory for a comprehensive assessment of policy impacts. Natural experiments, linking changes in types and levels of public provisions to intergenerational family practices, incomes, and health indicators, should become standard practice.

**Acknowledgements** Financial support for work on this chapter by the first author came from the European Research Council Advanced Investigator Grant (ERC, 324211) "Families in Context".

Djundeva worked on this chapter while she was a postdoctoral researcher in the Department of Public Administration and Sociology at the Erasmus University Rotterdam. Funding for her work came from the Erasmus Initiative "Vital cities and citizens".

## **References**


*2030* (pp. 189–258). Paris, France: OECD Publishing. https://doi.org/10.1787/ 9789264168367-en.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **15**

# **How Well Do European Child-Related Leave Policies Support the Caring Role of Fathers?**

**Alzbeta Bartova and Renske Keizer**

During the last 50 years, fathers have become more and more involved in parenting (Hook, 2006; Maume, 2010; Yeung, Sandberg, Davis-Kean, & Hofferth, 2001). Although scholars have shown that breadwinning remains a strong component of the cultural conceptions of men's parenting (e.g., Christiansen & Palkovitz, 2001; LaRossa, 1997), this role is increasingly being complemented with a diversity of social roles, such as caregiver and nurturing parent. Despite the prevalent "new father" discourse, however, parenting remains gendered. Fathers continue to spend less time taking care of their children than mothers, even when both parents work full time (Dermott & Miller, 2015; Doucet, 2013; Kan, Sullivan, & Gershuny, 2011), and they do less "solo" parenting than mothers (Raley, Bianchi, & Wang, 2012; Yeung et al., 2001).

However, the fact that fathers spent relatively little time with their children does not stem from reluctance or unwillingness (see Grunow & Evertsson, 2016, 2019). Most fathers indicate that they want to be more involved in the lives of their children. Data from the International Men and Gender Equality Survey (IMAGES) show that most fathers report that they would work less if it meant that they could spend more time with their children. However, various factors, such as lack of institutional support for paternal

A. Bartova (B) · R. Keizer

© The Author(s) 2020 R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_15

Erasmus University Rotterdam, Rotterdam, The Netherlands e-mail: alzbeta.bartova@kuleuven.be

R. Keizer e-mail: keizer@essb.eur.nl

childcare, financial constraints, or worries about employer discrimination, often refrain men from actually spending more time with their children (Grunow & Evertsson, 2016, 2019; Wilson & Prior, 2010). Enabling men to become more active in their children's lives facilitates men to be as involved as they want to be, and, even more important, brings benefits to the lives of their children, their partners, and their own lives, and the world around them (Levtov, Van der Gaag, Greene, Kaufman, & Barker, 2015).

In this chapter, we analyze how contemporary family policies in Europe support caregiving role of fathers. We focus our analysis solely on maternity, paternity, parental, and childcare leaves, as these are the policies that have the strongest potential to facilitate fathers' participation in childcare. With an exception of paternity leave, the leave policies were introduced in many countries to primarily protect health of mothers and babies, and to facilitate women's labor market participation (Sigle-Rushton, Goisis, & Keizer, 2012). The caring responsibilities of fathers have not been legally recognized for a long time. Only in the 1970s the first European countries started to recognize fathers' rights to care. Despite that, women remain to be the primary carers in virtually all European countries, and fathers' use of leave entitlements considerably vary across countries. We are interested in the extent to which European countries recognize the caring responsibilities of fathers toward their children and partners, and the value they place on fathers' caring role. To do so, we analyze the designs of individual leave policies and compare them with the available data on their uptake by fathers in 13 European countries that represent different geographical regions to capture the variation in the state approach to caregiving fathers.

# **Why Do We Need Caring Fathers?**

Growing empirical evidence demonstrates that fathers' involvement in childcare has an immense value not only for their children and partners but also for the whole society. Active involvement of fathers in childcare has been found to be important for men's own physical and psychological health, relationship stability, child development, women's labor force participation, and their psychological well-being, but also on the intergenerational transmission of fathers' involvement, which has a potential to strengthen the abovementioned positive impressions in future generations. Below we provide a very short overview. It is important for readers to realize that most of the findings we refer to below are based on research on heterosexual couples. This focus comes with certain limitations. For example, when investigating the impact of father involvement on child outcomes in heterosexual couples, it is very difficult to disentangle whether the effects found are related to the biological sex of the parent or their gender (and gendered division of tasks). As such, more research on same-sex couples, in this example, gay fathers, is needed—see Chapter 16 by Evertsson, Jaspers, and Moberg in this volume for new insights on family formation and leave rights in same-sex parent families.

Studies have stressed the importance of getting men involved in their (unborn) child's life as early as possible. When men are engaged from the start of children's lives, for example, by participating in prenatal care, they are much more likely to remain actively involved in childcare when the child is older (Cabrera, Fagan, & Farrie, 2008). Moreover, men's involvement during pregnancy and in the days and weeks after giving birth benefits women's physical recovery after birth and is shown to contribute to lower rates of post-partum depression (Plantin, Olukoya, & Ny, 2011). Several studies report that being an involved father is good for men themselves as well (Keizer, Dykstra, & Poortman, 2009; Knoester & Eggebeen, 2006). Men who are actively involved in the daily lives of their children live longer, have fewer mental or physical health problems, are less likely to abuse drugs, are more productive at work, and report being happier than fathers who are less involved with their children (Keizer, Dykstra, & Van Lenthe, 2011; Knoester & Eggebeen, 2006). Higher involvement of fathers in childcare activities has also been associated with women's higher satisfaction with the couple relationship (Carlson, Hanson, & Fitzroy, 2016; Risman & Johnson-Sumerford, 1998) and is linked with lower likelihoods of getting divorced (Kotsadam & Finseraas, 2011).

Although there is an ongoing scholarly debate about the roles of fathers and mothers in parenting (Cabrera, Fitzgerald, Bradley, & Roggman, 2014; Fagan, Day, Lamb, & Cabrera, 2014; Grossmann et al., 2002; Lamb, 2010; Paquette, 2004), there is more and more evidence suggesting that fathers' involvement influences children's development in similar ways as the mothers'. There is a general consensus in the literature that fathers' involvement in childcare is positively related to children's behavioral and emotional development (for reviews see Cabrera, Tamis-LeMonda, Bradley, Hofferth, & Lamb, 2000; Lamb & Lewis, 2013; Sarkadi, Kristiansson, Oberklaid, & Bremberg, 2008). Furthermore, fathers' involvement has also been linked to higher cognitive and language development and school achievement of children (Cabrera, Shannon, & Tamis-Lemonda, 2007; Pancsofar & Vernon-Feagans, 2010; Tamis-LeMonda, Baumwell, & Cabrera, 2013). In addition, multiple studies have shown that fathers' interactions with their children are important for the development of empathy and social development in both sons and daughters (Leidy, Schofield, & Parke, 2013). Furthermore, father's involvement is related to daughters' self-esteem (e.g., Keizer, Helmerhorst, & Van Rijn-Van Gelderen, 2019). Finally, studies report that children have better physical and mental health when their fathers are actively involved in their lives (Carlson, 2006).

By being more involved in caregiving work, men also facilitate women's re-entry on the labor market. By increasing their involvement at home, their partners are able to return to the labor market sooner and for longer hours and therefore increase their labor market outcomes (Ory, 2019). The benefits of father's involvement may also reach beyond the individual household. A study finds that if women participated in the labor market at the same rates as men do, this would lead to an average increase in gross domestic product (GDP) of 12% by 2030 across the OECD countries (Thévenon, Ali, Adema, & Salvi del Pero, 2012).

Father's stronger involvement in childcare may also have implications for how the next generation divides childcare responsibilities. According to the social cognitive theory of gender development (Bussey & Bandura, 1999), children look to models in their environment for information about genderappropriate behavior. Parents who divide childcare responsibilities equally have children with more flexible attitudes about gender (Croft, Schmader, Block, & Baron, 2014; Fulcher, Sutfin, & Patterson, 2008), and daughters who are less vulnerable to gendered achievement patterns (Updegraff, Mchale, & Crouter, 1996) and who aspire to less traditional roles (Croft et al., 2014). Men who have seen their own fathers engage in domestic work are themselves more likely to be involved in household work and caregiving as adults (Ory, 2019). In the following section, we will discuss child-related leave policies that set the rules for fathers' taking time off work to provide childcare.

# **Father's Statutory Right to Care**

Across European countries, the right to care for small children is most often guaranteed through Labor and Social Security Codes. These legislations offer fathers a time off work to provide care and financial compensation for their foregone earnings while on leave. Depending on the country, fathers can draw their entitlements from maternity, paternity, and parental leaves. These policies differ primarily in the purpose for which they were introduced and in the intended recipient. The first leave policy—maternity leave, was introduced in Germany in 1883 (Kamerman & Moss, 2009) and by 1919 it was recognized by the International Labour Organisation which formalized the minimum standards for maternity leave in its Maternity Protection Convention (ILO, 1919). Its main purpose then and now is to provide health, job, and income protection to pregnant working women and women recovering from childbirth. Although the maternity leave is primarily designed for mothers, in some countries, it can be transferred to the father in case the mother died or experienced complications during birth. More recently, some countries allowed mothers to transfer part of their maternity leave to fathers regardless of their health conditions (e.g., Czechia and Slovakia).

Parental leave was introduced several decades later. The first parental leave that allowed both parents to provide care for their young child was introduced in Sweden in 1974 (Sundström & Stafford, 1992). In general, the purpose of parental leave is to provide care to a young child and is available to both parents after maternity leave (International Labour Organization, 2014). Paternity leave is the newest addition to the leave policies and was designed specifically for fathers to be taken around the time of birth of their child. The main purpose of paternity leave is to grant fathers leave from work to provide care and support for their partner and newborn child. Apart from the three main types of leave, fathers can be entitled to a leave when their child falls ill, and in a small number of countries they can apply for a socalled childcare or home care leave, which is an alternative for parents who wish to continue providing care for their children after parental leave instead of enrolling them in a formal childcare (Blum, Koslowski, Macht, & Moss, 2018; OECD, 2017).

As Moss, Koslowski, and Duvander (2019) recently pointed out, there is a huge variation in leave designs across countries, which makes a meaningful comparison a challenging task. The great variation in the leave policies is primarily due to the differences in timing and motivations behind the introduction of these policies. It is important to remember that leave policies were created and modified in a response to a particular social, economic, and political needs of each country which inevitably affected their designs in a very specific way. For instance, in Czechoslovakia, the motivation behind the introduction of parental leave policies was to increase fertility which supposed to be achieved through an extended period of leave for mothers (Frejka, 1980). On the other hand, in Sweden, the introduction of parental leave policy was motivated by the achievement of greater gender equality in society (Duvander & Johansson, 2012). This motivation set the country on a path of individual rights to leave for both parents and normalized the caring responsibilities of fathers over time. As a result, maternity, paternity, and parental leave policies can mean different things in different countries. By following the established categorization of leave policies into maternity, paternity, parental and childcare leave, we would substantially constrain the scope for a cross-national analysis of these policies and what they mean for mothers, fathers, and children across countries. Furthermore, the classification is also becoming increasingly inaccurate as more countries adopt leave policies that do not recognize such categories (e.g., Portugal, Iceland, Denmark, Norway, and Sweden).

We propose to abandon the potentially misleading classification of leave policies and to treat them as a blended or composite policy measure that combines leave entitlements for mothers, fathers, and parents. The advantage of this approach is in its ability to reach beyond the country-specific characteristics of leave policies. It allows to decompose the national policies and to create a clearer picture of the leave entitlements for parents across countries. In our chapter, we will refer to the composite leave policies as to child-related leave policies. We will focus on several important dimensions, which are characteristic for all leave designs, to analyze how the 13 European countries in our sample support the caring role of fathers.

## **Dimensions of Child-Related Leave Policies and Their Implications for Fathers' Leave Uptake**

Each child-related leave policy in every country has several dimensions that are central to their designs. Their characteristics vary across countries and have implications for leave uptake and the share of leave between mothers and fathers. Moreover, these characteristics also have a wider implication for maternal employment, the gender wage gap, health outcomes, or child development (Chatterji & Markowitz, 2012; Cukrowska-Torzewska, 2017; Galtry & Callister, 2005). There are four important dimensions of child-related leave policies that have implications for fathers' leave uptake—the type of entitlement, transferability, generosity, and flexibility.

The type of *entitlement* to child-related leave policies is either individual or family-based. In other words, the type of entitlements determines whether the leave is designed specifically for a mother or a father (individual entitlement) or whether it is intended for both parents to share according to their preferences (family entitlement). The type of entitlement applies to both time off work (leave) and to financial benefits for the time spent on leave. While maternity and paternity leave policies are usually accompanied by individual entitlements to financial compensation for foregone earnings, this is not the case for parental and childcare leave policies. In some countries, working parents are entitled to an individual right to parental leave but

the financial benefits are based on family entitlement (e.g., France, Czechia). This means that although both parents could use parental leave at the same time, they would be entitled to only one benefit. Four combinations of entitlements to leave and benefits are possible: (1) an individual entitlement to leave with an individual entitlement to benefits, (2) individual entitlement to leave with a family entitlement to benefits, (3) family entitlement to leave with an individual entitlement to benefits, and (4) family entitlement to leave with a family entitlement to benefits.

Another important dimension is the *transferability* of leave entitlements. Since family entitlements are transferable entitlements by default, this condition only applies to individual entitlements. It determines whether one parent's leave entitlement can be transferred to the other parent (transferable) or whether the amount of leave dedicated to one parent will be lost if not claimed (non-transferable). The transferability applies to both time (leave) and financial benefits.

*Flexibility* is a dimension of child-related leave that refers to the freedom of choice parents have when deciding on when and how to use their leave entitlements. The flexibility is often determined by a time frame within which parents can use their entitlements. For instance, when fathers are entitled to five days of paternity leave which they need to use within four weeks following the birth of their child (The Netherlands). Another form of flexibility, that is characteristic for parental leave policies, is the regulation of the mode in which parents can use their entitlements—whether they must be used consecutively, whether both parents can use their entitlements at the same time, whether parents need to use their whole entitlement in one go or whether they can break it into several time blocks or take the leave on a part-time basis (e.g., one day a week).

The *generosity* of the financial benefits has for long been presented as a crucial for fathers' leave uptake. Although there is no consensus about the threshold below which fathers are not willing to use their leave entitlements, the general rule is the more generous the more attractive the prospect of childcare leave for fathers (Boll, Leppin, & Reich, 2013). The financial benefits for the time spent on leave are either proportional to the earnings or paid as a flat-rate benefit. In the following section, we will discuss the implications these dimensions have for the uptake of the leave policies.

### **Leave Designs and Their Implications for Fathers' Leave Uptake**

The type and transferability of entitlements to child-related leave and the generosity of the financial benefits for fathers are exceptionally important for their leave uptake. Despite the rising numbers of caring fathers and working mothers, men's identity and the expectations society places on them are still largely associated with the labor market (Christiansen & Palkovitz, 2001). In many countries, men and fathers are still often perceived by themselves or by others as providers who are responsible for ensuring financial stability for their family, and as ideal workers who are not constrained by their children's care needs (Lott & Klenner, 2018). By granting fathers an individual nontransferable entitlement to child-related leave, countries recognize fathers as carers and equalize their caring responsibilities with those of mothers. The statutory entitlement to individual non-transferable leave also sends a signal to employers that they need to acknowledge the caring responsibilities of their male employees (Brandth & Kvande, 2019). Moreover, the statutory entitlement to individual non-transferable leave strengthens fathers' negotiating position when requesting leave compared to requesting leave based on family entitlement (Haas & Hwang, 2019; Moran and Koslowski, 2019). The individual non-transferable right to leave also gives fathers a stronger position in the home when negotiating the division of care with the mother (Närvi & Salmi, 2019). At the same time, the importance of men's contribution to family income, which increases in families with lower socio-economic status, cannot be overlooked. There is a strong agreement in the international research that fathers are most likely to use their entitlements to child-related leave if they are based on an individual non-transferable entitlement that is accompanied by generous financial benefits (Boll et al. 2013).

Flexibility in the use of leave is a dimension of child-related leave policies that is rarely discussed but which may potentially have a strong implication for fathers' leave uptake. Flexibility is often determined by the difference between the duration of the leave entitlement and the time frame within which this entitlement has to be used. Flexibility is greater the larger is the time frame. The leave duration for each parent is most commonly defined by a specific time unit (hours, days, weeks, months) but in some countries the leave duration is determined by the age of the child for whom the parents draw their entitlements. This formulation of leave duration carries some characteristics typical of family-based entitlements. When it is paired with an individual non-transferable entitlement to leave, it transforms its meaning. It is because despite having an individual non-transferable entitlement, both

parents are entitled to the same duration of leave within the same time frame. It means that parents need to negotiate who and when will take the leave. This design lacks the clear recognition of fathers' caring roles typical of nontransferable leave entitlements and may weaken fathers' negotiating position with their partners and employers. Therefore, apart from the type of entitlement, its transferability and the generosity of the financial benefits, we also need to consider how is the leave duration determined and what is the time frame within which can parents draw their entitlements. Child-related leave policies that (1) are based on individual non-transferable right to time off work, (2) come with generous financial benefits, and (3) that are defined in time units with extended time frame for uptake represent the leave designs that are likely to meet with the highest use among fathers.

# **Data and Methodology**

To analyze the child-related leave policies, we use two sources of legislative information. One of the data sources is the 14th International Review of Leave Policies and Related Research 2018 (Blum et al., 2018) where we use the information on maternity, paternity, parental, and childcare leave. The second source of data is the Mutual Information System on Social Protection (MISSOC) database from which we use information on maternity and paternity leave (Section IV. Maternity/Paternity leave) and on parental leave and benefits (Section IX. Family Benefits) from the 1 July 2018 (2018-07- 01 update). Our analytical framework is based on the decomposition of maternity, paternity, parental, and childcare leave policies in individual European countries based on the dimensions of child-related leaves discussed in previous sections. The purpose of the policy decomposition process is to gain complete information on the fathers' rights to child-related leave regardless of the specific policy measure they can be drawn from in each country.

To analyze the fathers' child-related leave entitlements, we first pool maternity, paternity, parental, and childcare leaves in each country. We include maternity leave in our analysis because part of the leave may be transferable to the father. However, we recognize the transferable maternity leave as father's entitlement only if the transferability is not conditioned on the death or health conditions of the mother. In the second step, we differentiate between individual and family entitlements to child-related leave. In the third step, we look at the type of entitlement to financial benefits whether they are based on an individual or family entitlement. As we already mentioned in the previous section, the combination of our third and fourth steps yields four categories of fathers' entitlements to child-related leave—(1) an individual non-transferable entitlement to leave for fathers with an individual entitlement to benefits, (2) individual transferable entitlement from mother to father with an individual entitlement to benefits, (3) individual non-transferable entitlement to leave with a family entitlement to benefits, and (4) family entitlement to leave with an individual entitlement to benefits. In the next step, we specify how many weeks of leave corresponds with each of the type of leave entitlement. In case of family entitlements to leave, we assign the total number of weeks of the family entitlement to the father. In the last step, we focus on the financial benefits associated with the leave. We break down the total entitlement to child-related leave for fathers into weeks and the generosity of the financial benefits. We use replacement rate to measure the generosity of child-related leave entitlement for fathers. This indicator shows the proportion of earnings that is replaced by the financial benefits.

During the operationalization process we came across several variations in policy designs that were not readily comparable across countries. First of all, when the duration of leave reserved for one or both parents was determined by the age of the child, we recalculated the total duration of this leave into the number of weeks. We calculated 52 weeks for each year of the child's age and from this number we subtracted the amount of postnatal leave that is reserved for mother (regardless of whether part of the postnatal leave for mother was transferable to the father or not). In case the leave duration was specified in the number of months, we recalculated this value into number of weeks (4.3 weeks in 1 month). Second, in some countries the financial benefits are provided as a flat-rate payment. In this case, we used Eurostat data on men's median monthly earnings in industry, construction, and services<sup>1</sup> from 2014, which is the most recent data currently available, to calculate the replacement rate. In countries where flat-rate benefits are granted based on income level (France), we use the basic benefit level to calculate the replacement rate. We use the most generous benefit in countries where parents can choose from different levels of flat-rate or income-related benefits.

Where available, we contrast the leave designs with fathers' use of leave. We source the data on leave uptake from the Annual Reviews produced by the International Network on Leave Policies and Research.<sup>2</sup> Unfortunately, the data are limited and often collected in different years, which restrict their power for comparative analysis. Moreover, they tend to refer to a specific

<sup>1</sup>The data were collected from the "Structure of earnings survey: monthly earnings" (earn\_ses\_monthly).

<sup>2</sup>https://www.leavenetwork.org/annual-review-reports/country-reports/, accessed 17 December 2019.

policy (e.g., paternity or parental leave) and the data on uptake are often unavailable for each of the leave in every country. Despite that, we decided to report this information alongside our analysis of child-related leave designs to consider possible differential impact on the fathers' leave uptake and to emphasize the need for better-quality data. In these cases, we always mention the type of leave policy the data refer to.

# **Results**

Figure 15.1 plots the total amount of child-related leave entitlements for fathers broken down by their type. Overall, in the majority of the countries we study, fathers are entitled to relatively long leaves that exceed one year in duration. However, the Figure also reveals a considerable variation in the quality of leave entitlements for fathers across countries. Although fathers have access to individual non-transferable leave in almost all 13 countries, the length of the leave varies anywhere from 1 week in Czechia to 156 weeks in Spain and Slovakia. Only Slovak fathers have not gained access to individual non-transferable leave. Nonetheless, this may change soon since the European Union Directive on work-life balance for parents and carers came into effect in 2019. The Directive requires all Member States to introduce policies

**Fig. 15.1** Distribution of leave entitlements to child-related leave for fathers (in weeks), 2018 (*Note* Based on authors' own calculations. *Source* 14th International Review of Leave Policies and Related Research 2018; MISSOC 2018-07-01)

that grant fathers with at least 10 working days of individual non-transferable right to paternity leave.

The second most common type of leave for fathers is based on family entitlement to leave and individual entitlement to financial benefit. We found this type of entitlement in Germany, Poland, Portugal, Iceland, and Norway. The remaining types of child-related leave entitlements are much less common. We found the individual entitlement to leave with family entitlement to benefits only in Czechia, Slovakia, and Norway. Mothers are allowed to transfer some of their entitlements to fathers in Sweden, Czechia, and Slovakia. Finally, Sweden is the only country in our sample that allows fathers to transfer some of their individual leave entitlement to mothers. The empirical evidence shows that fathers are more likely to use their individual non-transferable leave if it is paired with sufficiently high financial compensation (Boll et al., 2013). Figure 15.2 shows the distribution of financial compensation for fathers across the total amount of their leave entitlements in cross-country comparison. The data reveal considerable variation in the value countries place on fathers' time on childcare. Fathers are not guaranteed any statutory financial compensation for a considerable share of their leave entitlements in 5 out of the 13 countries.

We identified two distinct approaches to leave entitlements for fathers— (1) leave with dominant individual non-transferable rights and (2) leave

**Fig. 15.2** Distribution of financial compensation across the total leave entitlements (in weeks), 2018 (*Note* Based on authors' own calculations. *Source* 14th International Review of Leave Policies and Related Research 2018; MISSOC 2018-07-01)

composed of several different types of entitlements. The first type is strong in Western and Southern European countries. The second type, on the other hand, is prevalent in Northern and Central and Eastern European countries. However, we also identified variations in these two main approaches to leave for fathers. Among the countries with composite leave entitlements, Sweden, Iceland, and Norway provide fathers with generously compensated individual non-transferable right to leave. In contrast, among the countries with strong emphasis on individual non-transferable leave entitlements, Germany, Portugal, and Italy offer additional months of leave which vary in the type of entitlement and generosity. We capture this variation in dividing the analyzed countries into four groups: countries with dominant individual non-transferable leave entitlements, countries with dominant individual nontransferable leave entitlements and additional incentives for fathers, countries with composite child-related leave entitlements and strong father's quota, and countries with composite child-related leave entitlements and limited father's quota.

#### **Countries with Dominant Individual Non-Transferable Leave Entitlements**

The countries where fathers are solely entitled to individual non-transferable leave are Spain, The Netherlands, The United Kingdom (UK), and France. In Spain, where we found the longest individual non-transferable leave entitlement, fathers are fully compensated for only 4.4 weeks of leave. The remaining 151.6 weeks are without a statutory guarantee of financial benefits. The situation is similar for fathers in The Netherlands and the United Kingdom. In The Netherlands in 2018, fathers were fully compensated only for 0.4 weeks of their leave entitlements and the remaining 26 weeks were without a statutory guarantee of financial benefits. In contrast, fathers in the UK are not fully compensated for any part of their leave entitlements. They are entitled to financial benefit that corresponds with about 18% of their earnings and that is only for one week of their total entitlements. For the rest of their leave entitlements, they are not guaranteed any financial compensation. In France, where fathers are offered 56.2 weeks of individual non-transferable child-related leave, they are fully compensated for only 2.2 weeks of this entitlement. For the remaining duration of their leave entitlement, fathers are entitled to financial benefits that correspond with about 17% of their earnings.

When we look at fathers' leave uptake, we can identify some patterns in the fathers' preferences for well-compensated leave entitlements over unpaid or poorly compensated leave that have been described in the literature. In 2012, 74% of Spanish fathers took at least some of their 4.4 weeks of fully paid leave (Meil, Lapuerta, & Escobedo, 2018). However, the remaining share of the leave that corresponds with parental leave entitlements has generally very low uptake both among women and men. In 2016, only 9.9% of all parents took some parental leave and out of this 9.9%, about 93% were mothers and about 7% were fathers (ibid.). In France in 2016, 62% of fathers used their entitlement to the fully paid leave (Boyer & Fagnani, 2018). Out of all users of the remaining share of leave, which is also available to mothers, only 4.4% were fathers (ibid.). The situation is similar in The Netherlands where 83% of fathers took the fully paid leave entitlement (0.4 weeks) in 2013 but only 11% of fathers took some of the remaining leave entitlement that is not accompanied with a statutory guarantee of financial benefit (Den Dulk, 2018). However, it is important to point out that some Dutch employers offer financial compensation for this type of leave, which is likely to influence the statistics on leave uptake. This is the case also in the UK, where in 2006 about 30% of fathers took at least some of their leave entitlements without a statutory guarantee to financial compensation. Out of the 30%, about 17% of fathers were fully compensated for their time on leave, 6% were partially compensated and 7% of these fathers took unpaid leave (O'Brien & Koslowski, 2018). About 74% of fathers in the UK took at least some of their one-week entitlement to paid leave (O'Brien & Koslowski, 2018).

Although the uptake of unpaid leave entitlements is low, fathers do not necessarily withdraw from participation in childcare. This is particularly true for the period around the time of birth. The Netherlands and the UK are good examples of fathers adopting alternative strategies to provide support to their families. In these countries, a considerable number of fathers opt for using their annual leave either as the sole source of leave around the time of birth or as an additional leave to their statutory child-related leave entitlements. In the UK in 2009, 18% of fathers who took any form of leave opted for other forms of paid leave than their child-related leave entitlements (e.g., annual leave) and 25% used both their entitlements to child-related leave and some other form of leave entitlements (O'Brien & Koslowski, 2018). In The Netherlands, in 2013, about 60% of fathers who used their entitlements to the fully paid child-related leave also used some of their annual leave around the time of birth (Den Dulk, 2018).

#### **Countries with Composite Child-Related Leave Entitlements and Strong Father's Quota**

The Nordic countries in our sample (Sweden, Norway, and Iceland) are countries that have composite leave entitlements and provide strong incentive for fathers' leave uptake. Sweden is the only country in the whole sample with composite leave entitlements that does not provide family entitlement either to leave or to financial benefits. Swedish fathers are entitled to 20 weeks of individual non-transferable leave, 30 weeks of leave that can be transferred to the mother, and an additional 30 weeks that mother can transfer to the father. In total, Swedish fathers can spend up to 80 weeks with their small child. They are offered financial benefits that compensate for 78% of their earnings for the majority of their leave (62 weeks). For the remaining period, they are entitled to financial benefits that correspond with about 11% of their earnings. In 2004, about 88% of Swedish fathers took at least some child-related leave but according to data from 2008, fathers on average took 106 days of leave, which corresponds with about 21.2 working weeks (Duvander & Haas, 2018). This is only slightly more than the 20 weeks of their individual nontransferable entitlement, which can mean either that fathers take some of the entitlements that can be transferred to their partner or that some fathers work more than 5 days a week.3 Nonetheless, the statistics on fathers' leave uptake suggests that the fathers tend to use only their individual non-transferable entitlement and they transfer the rest to their partner (in 2008, mothers took on average 342 days, or 68.4 weeks) (ibid.).

We can observe similar outcomes in Iceland and Norway. In Iceland, fathers are provided with 30.1 weeks of individual non-transferable right to leave. However, they are relatively generously compensated (85%) only for 12.9 weeks and the remaining 17.2 weeks are unpaid. Fathers are also eligible to 12.9 weeks of family entitlement to leave with an individual entitlement to benefits paid at 85% replacement rate. Data from 2015 show that 81% of fathers used at least some of their child-related leave entitlements and on average, they took 88 days (17.6 weeks). However, 34.3% of fathers took less than the 12.9 weeks of their individual non-transferable leave, and only 13.3% of fathers took some of the family entitlement to leave (Eydal & Gíslason, 2018). In Norway, fathers are entitled to 10 weeks of individual non-transferable leave that is fully compensated and to 2 weeks of individual non-transferable leave that is not accompanied by any financial

<sup>3</sup>We used 5-day working week to recalculate the 106 days into weeks. If we were to calculate 6 days or 7 days working weeks, the transformation into weeks would yield a lower number than 21.2 weeks.

benefits. Apart from that, they are also entitled to 36 weeks of family entitlement to leave with individual entitlement to benefits (100% replacement rate), and to 52 weeks of individual entitlement to leave with a family entitlement to benefits (about 14% replacement rate). The statistics on the use of leave by Norwegian fathers are not regular but the uptake is reported to vary between 70 and 90% (Brandth & Kvande, 2018). The uptake is therefore very similar to those observed in Sweden and Iceland. In 2014, fathers who used some child-related leave took on average 49 days (9.8 weeks), which again corresponds with the individual non-transferable entitlement to leave (ibid.).

### **Countries with Composite Child-Related Leave Entitlements and Limited Father's Quota**

The remaining countries with composite child-related leave entitlements for fathers who offer only limited incentives for fathers' leave uptake are Czechia, Slovakia and Poland. In Czechia, fathers gained individual non-transferable entitlement to leave only in February 2018 (1 week). Apart from that, Czech fathers are entitled to about 29 weeks of individual entitlement to leave with family entitlement to benefits which compensates for up to 100% earnings,4 and to 16 weeks of individual transferable leave from mother (70% replacement rate). In 2018, 43,442 fathers used their entitlements to the one week of individual non-transferable leave (Kocourková, 2018), which corresponds with about 38% of all births in that year. The uptake of the individual entitlement to leave with family entitlement to benefits varies between 1 and 2% (ibid.). In Slovakia, fathers are only entitled to 28 weeks that can be transferred from mother and these compensate for 75% of earnings. In addition, they are also entitled to 128 weeks of individual entitlement to leave and family entitlement to benefits (about 26% replacement rate). According to existing data, in 2017, only about 7% of all recipients of maternity leave were men (Gerbery, 2018). Unfortunately, data on the uptake of the individual leave entitlement with family entitlement to benefits are not available. In Poland, fathers are entitled to two weeks of fully compensated individual non-transferable leave and to 142 weeks of family entitlement to leave with

<sup>4</sup>Parents are entitled to 220,000 CZK over a period of 4 years. Parents can choose freely the duration of leave, which determines the amount of benefits they receive every month. The maximum benefit parents can draw is 32,640 CZK (about e1.263) that can be drawn over a period of 29 weeks. The maximum amount of monthly benefit payment is higher than the median monthly salary in Industry, Construction and Services in 2014 (e851). We, therefore, used 100% replacement rate to mark the generosity of the maximum benefits.

individual entitlement to benefits. Out of the 142 weeks, for 32 weeks are Polish fathers entitled to 80% of their earnings but the remaining 110 weeks are unpaid (small financial benefit is available to families with very low income). In 2017, 174,200 fathers used their individual non-transferable entitlement to leave, which represents about 43% of all births in that year, and used 12.6 days of leave on average (Kurowska, Michon, & Godlewska- ´ Bujok, 2018). The uptake of the remaining leave entitlements is about 1% (ibid.).

#### **Countries with Dominant Individual Non-Transferable Leave Entitlements and Incentives for Fathers**

Three countries with individual non-transferable entitlements (Germany, Portugal, and Italy) provide incentives to increase fathers' uptake of childrelated leave entitlements which are distinct from the father's quota observed in the Nordic countries. These incentives are referred to as *bonus months*. This means that if fathers take a specific amount of child-related leave, the total amount of leave available per child, or per family, is extended by a certain number of months. The bonus months are a particularly interesting characteristic of child-related leave entitlements because they try to mimic the success of father's quotas in Nordic countries through distinct policy design. Although all three countries adopted bonus months in their leave designs, each did it in their own way. We will now look more closely at this feature in the three countries and will discuss how successful they are in encouraging fathers use of child-related leave entitlements.

In Italy, fathers are entitled to 26.6 weeks of individual non-transferable leave but only 0.8 weeks are fully compensated (paternity leave). The remaining 25.8 weeks are compensated at 30% of earnings. Both father and mother are entitled to 25.8 weeks of individual non-transferable leave (parental leave) but they can only use 43 weeks of leave entitlements per child. It means that although the leave entitlements are individual and nontransferable, parents still need to negotiate who and when will take the leave because the total allowed leave duration is shorter than the sum of the individual entitlements of both parents. Therefore, although the leave design is based on an individual non-transferable entitlement, it carries characteristics that are typical for family-based entitlements. To encourage fathers' uptake of child-related leave, Italy introduced a condition that guarantees the family additional 4.3 weeks (i.e., bonus month) if the father takes at least 12.9 weeks of his entitlement. However, the bonus month does not specify who is it intended for, which makes it a family entitlement.

At the beginning of our chapter, we argued that flexibility in leave duration and the length of the time frame within which parents can use their leave entitlements is crucial for fathers' uptake of child-related leave policies. A time frame that is as long as the total leave entitlement of both parents, or is shorter than the total leave entitlement, transforms the individual non-transferable leave into a family entitlement. This is because the limited flexibility in leave use is forcing parents to negotiate their leave. We also expected that the leave policies with such designs will meet with low uptake among fathers. We found some support for this argument, although our analysis cannot directly link the data with the bonus month. In 2016, out of all parental leave recipients in Italy, only 17% were fathers (Addabbo, Cardinali, Giovannini, & Mazzucchelli, 2018).

Germany is somewhat a special case among the countries with dominant individual non-transferable entitlements to leave. It is because German fathers are entitled to an individual non-transferable right to leave and to benefits that are proportional to their income (65%) until their child is two years old. This entitlement is also available to mothers. However, the framing of the leave duration based on the child's age transforms the leave design into a hybrid between individual and family entitlement to leave. This is because each parent does not have a specifically allocated time for leave but instead have to negotiate with the other parent when and how long a leave to take, although taking the leave simultaneously is also a possibility. If both parents use at least two months of leave, the family is provided with two additional months of leave. As in Italy, the law does not specify who are these bonus months intended for, which, strictly speaking, makes it a family entitlement. Even though the bonus months are not specifically designed for fathers, the information on leave uptake suggests that the majority of fathers who took some child-related leave did not take more than two months (78.9% of male recipients in 2016) (Reimer, Erler, & Blum, 2018). However, in 2014 only 35.7% of fathers used their leave entitlements.

Finally, in Portugal, fathers are entitled to 17.9 weeks of individual nontransferable leave but only 5 weeks of this entitlement are fully compensated. The remaining 12.9 weeks compensate for 25% of fathers' earnings. Apart from that, fathers are entitled to 11.4 weeks of family entitlement to leave with individual entitlement to benefits (100% replacement rate). In case, fathers use at least 4 weeks out of this family entitlement, the family is entitled to an additional 4 weeks of leave (i.e., bonus month). In 2017, about 71% of fathers took the 5 weeks of individual non-transferable leave that is fully compensated. However, only about 34% of families received the bonus month and the vast majority of fathers took no more than one month of the family entitlement to child-related leave (Wall & Leitão, 2018). This is somewhat surprising because the financial compensation for fathers' leave is exceptionally generous in cross-national comparison. It suggests that neither generous financial benefits may be successful in raising fathers' use of leave policies if the generous benefits are not associated with an individual nontransferable right to leave. This is also supported by the statistic on uptake of the 12.8 weeks of individual non-transferable entitlement that compensates only 25% of earnings—only 8% of all parents used this entitlement in 2017 (Wall & Leitão, 2018).

## **Conclusion**

In this chapter, we aimed to analyze how contemporary family policies in Europe support caregiving role of fathers. We found support for the notion that generous individual non-transferable leave entitlement is the most stimulating form of leave design for fathers' use of child-related leave. Although all countries offer fathers an individual non-transferable entitlement to child-related leave, the policies in the majority of the countries do not sufficiently challenge the gendered distribution of paid and unpaid work. The individual non-transferable leave entitlements are still very limited in Central and Eastern European countries (Czechia, Slovakia and Poland). However, in countries where fathers are entitled to much longer individual non-transferable child-related leaves, they are fully compensated only for a fraction of this entitlement, which negatively impacts on their leave uptake (e.g., The Netherlands, France, Portugal). In many countries, fathers are still not provided with any statutory guarantee of financial compensation for a substantial share of their individual non-transferable entitlements to child-related leave (The Netherlands, the UK, Spain, Poland). The less compensated child-related leave, the more it produces polarized behaviors among parents, mostly based on social class or education (Korpi, 2000; Saraceno, 2011). Without statutory guarantee of financial compensation, mainly fathers from higher social classes are capable to take short periods of (part-time) leave. In line with this idea, research has shown that throughout developed countries, higher educated fathers are more involved in childcare and are more likely to take up paternity leave (Boll et al., 2013; Geisler & Kreyenfeld, 2018). Differences in father involvement by social class may contribute to widening social-class disparities in children's resources and development. A child-related leave that is accompanied by generous financial benefits has the potential to equalize the leave uptake among fathers with different socio-economic background and consequently lead to fewer social-class disparities in children's resources and development.

We also pointed out the importance of flexibility, a rarely discussed component of child-related leaves, and the way policy designs frame the entitlements to leave duration. We found some evidence that these components may have an important effect on fathers' use of their leave entitlements. Although the generously compensated individual non-transferable leaves are the most attractive for fathers, their value seems to decrease with inaccurately formulated duration of leave and the time frame within which can parents use their leave. When both parents are entitled to an individual non-transferable leave, but the total amount of leave they may use per family is shorter than the sum of their individual entitlements, the individual non-transferable leave starts to resemble the leave that is based on a family entitlement. It is because both of the parents cannot use their full entitlements and have to negotiate when and how they will use the leave. Similarly, when the individual entitlement to leave is determined by child's age, the fathers not only have to negotiate with their partners, but they also lose on their negotiating position with their employer because their own leave entitlement is not clearly specified by law.

In some European countries, policymakers attempted to increase leave uptake among fathers through so-called bonus months. This measure provides a family with extra months of leave if both of the parents take a certain amount of leave. However, this incentive has not managed to achieve the success of father's quotas in Nordic countries, which they tried to mimic. In Northern Europe, the uptake of individual non-transferable entitlements hovers around 80%, while in countries that introduced the bonus month the uptake has not been higher than 35%. This difference may be partially explained by the cultural differences between the countries. While Nordic countries have a strong tradition of promoting gender equality, Germany, Portugal, and Italy are countries with rather a traditional approach to family and gender roles. However, the staggering difference is also likely to be caused by the design of the bonus months that are not built on an individual non-transferable right to child-related leave. For instance, the bonus months in Germany were already introduced in 2007, yet ten years later only 35.7% of fathers used parental leave. In contrast, in Norway the father's quota was introduced in 1993. Before that only about 3% of Norwegian fathers used some entitlement to parental leave. However, only seven years after the introduction the fathers' uptake of parental leave rose to 85% (Brandth & Kvande, 2019). Therefore, individual non-transferable leave entitlements that clearly specify the period of leave are crucial for the fathers' leave uptake. Generous financial compensation for the time spent on childcare is particularly important for leave uptake among fathers with lower socio-economic status. Nonetheless, the use of child-related leave policies by fathers is also likely to be shaped by the gender norms dominant in individual countries irrespective of the quality of their policy entitlements. On the other hand, institutional support for caring role of fathers normalizes fathers' active involvement in childcare, which itself may encourage shifts in common gender norms.

Finally, it is important to point out that the European landscape of fathers' child-related leave entitlements will likely change in the following years in response to the Work-Life Balance Directive.<sup>5</sup> The Directive requires all Member States to provide at least 10 working days of paternity leave and 4 months of individual leave for working parents out of which 2 months must be non-transferable. However, the financial compensation for the time spent on leave is left entirely in hands of the Member States. In consequence, the extent of the potential impact is debatable. Our results revealed that individual non-transferable entitlements to leave are common in all countries in our sample with the exception of the Central and Eastern European countries. Therefore, we expect that the new requirements on parental leave will primarily affect these countries but will trigger no to limited change in the remaining countries in our sample. We can expect more changes in paternity leaves. We expect paternity leave to be introduced in Germany and Slovakia in the coming years, and extensions to existing paternity leave in The Netherlands, Italy, and Czechia. The introduction of the Work-life Balance Directive certainly raises a question of how countries with individual leave entitlements framed by the age of child will deal with the condition of non-transferable entitlement for each parent. This also applies to countries with individual entitlement to leave and a time frame which prevents both parents from using their full individual entitlements. The condition of non-transferability can potentially lead to adjustments in the specification of leave duration in countries like Germany, which could increase fathers' bargaining power and ultimately lead to greater use of leave entitlements among fathers. On the other hand, most of the countries in our sample already have individual entitlements to leave in place without any reference to transferability. From a legal point of view, it can be argued that the conditions of the Directive concerning non-transferability of parental leave have already been fulfilled. In such case, the impact of the Directive on fathers' use of their leave entitlements will be severely limited.

<sup>5</sup>Directive 2019/1158 of the European Parliament and of the Council of 20 June 2019 on work-life balance for parents and carers.

**Acknowledgements** The present study was supported by a grant from the Netherlands Organization for Scientific Research (NWO MaGW VIDI; grant no. 452- 17-005) and by a grant from the European Research Council (ERC StG; grant no. 757210).

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **16**

# **Parentalization of Same-Sex Couples: Family Formation and Leave Rights in Five Northern European Countries**

**Marie Evertsson, Eva Jaspers, and Ylva Moberg**

As discussed extensively in this volume, policies have important implications for how families are formed and how they live their lives. Family leave policies, among other things, enable parents to take job-protected leaves from work to care for a newly born or adopted child and/or to take time off from work when a child is ill. However, factors such as the length of any job-protected leave, the levels of reimbursement, and the number of eligible policy users vary among countries. The degree to which individuals can make use of family leave policies also varies and is partly dependent on the extent to which they fit the legally recognized version of a parent. The family policy system was designed to cater to the heteronormative family, i.e., a (preferably married) mother and father with joint children. As new family constellations emerge and are legally recognized, it is important to study whether and to what extent the policies designed for the heterosexual, nuclear family also encompass other families.

M. Evertsson (B) · Y. Moberg

Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden e-mail: marie.evertsson@sofi.su.se

Y. Moberg e-mail: ylva.moberg@sofi.su.se

E. Jaspers

Department of Sociology/ICS, Utrecht University, Utrecht, The Netherlands e-mail: e.jaspers@uu.nl

In this chapter, we enhance theories of welfare state development by introducing the concept of *parentalization*. Parentalization is defined as the ability to become parents and be recognized as such, both legally and via social policies. To demonstrate an application of this new concept, we examine how states facilitate or hinder parentalization through laws and policies for same-sex parent families. We focus on five countries that were among the first to legalize same-sex unions/marriages and parenthood: the four larger Nordic countries (Denmark, Finland, Norway, and Sweden) and the Netherlands. Whereas the Netherlands is known as a LGBTQI-friendly<sup>1</sup> nation with moderately generous family leave policies, the Nordic countries are considered among the most family-friendly in the world, partly due to their generous family leave policies in terms of length as well as reimbursement levels. By mapping the paths to legally recognized same-sex parenthood and family policy rights, we can identify obstacles and hurdles on the road to full parentalization for couples in these countries.

Parentalization is operationalized as the extent to which individuals can (i) transition to legally recognized parenthood and (ii) make use of family leave policies to care for the child (in our example, parental, maternity, and paternity leaves). Both (i) and (ii) can be accessible in theory but more or less hard to achieve in practice. Parentalization, or the notion of who has the rights and possibilities to become a parent and who is excluded from parenthood and/or the policies that come with it, is closely linked to gender and norms regarding motherhood and fatherhood. We discuss family leave policies from the perspective of same-sex parent families in the countries in focus, demonstrating how the parentalization concept and the resulting analysis facilitate an understanding of how national policies contribute to or create (in)equalities between various groups of parents(-to-be). In conclusion, we briefly discuss the parentalization concept from the perspective of family constellations not analyzed in this chapter and arrive at a research agenda for the future.

# **Parentalization: A First Restricted Analysis**

Variation in the definitions of what constitutes a family, both socially and legally, facilitates for certain groups and hinders others from forming legally recognized unions, having children and being the kind of parent they would like to be, for instance, by making use of family policies. Family law, its

<sup>1</sup>LGBTQI is short for lesbian, gay, bisexual, transgender, queer, and intersex.

rules and regulations, formalizes not only individuals' rights, and obligations towards one another but also the extent to which society and the state have obligations toward the family and its members. The definition of a family that the state applies is based on the legal recognition of a parent–child relationship and determines who has authority over a child, who may act as a child's representative and who is eligible for family benefits. Not being legally recognized as a parent may, among other things, mean not being able to pick up prescribed medicine at the pharmacy if the child is ill, risking losing custody in case of divorce or the other parent's death and not being able to pass on financial assets via inheritance to the child without a written testimony.

Parentalization as a concept can be used to analyze a number of various groups and family constellations. The focus on same-sex couples in this chapter is motivated by the aim to link the analysis to access to family leave policies, which are more readily available and commonly used when children are small. Notably though, same-sex parent families often find it easier to fit within the definition of a family than do other groups such as single gay fathers, transgender individuals, or non-biological parents in multi-parent families2 (e.g., Carroll, 2018; Downing, 2013). When we refer to same-sex couples, we base the definition on the categories of woman/man as defined in population registers i.e., a person's legal gender. Hence, we may include individuals defining themselves as man, woman, both or neither. We do not distinguish identities linked to ethnicity, social class, or other aspects even though they in many ways structure an individual's transition to and experiences of parenthood (e.g., Carroll, 2018; Moore, 2011).

Before we turn to the practical details of how policies work, we briefly discuss the welfare regime and social policy frameworks. The aim is to show how the concept of parentalization expands on these frameworks by recognizing the problems that same-sex couples and other couples may face in realizing their (any) parenthood ambitions. In other words, to be parentalized both (i) through the legal system and (ii) through the policy system.

<sup>2</sup>We define a multi-parent family as a group of individuals linked by parent–child relations in a context where three or more parents share in the active parenthood for the same child/children. None of the countries in focus in this chapter recognize more than two legal parents. Still, it has not been uncommon among same-sex couples to form families that include more than two parents (willing and committed to act as parents for the child), especially in periods before joint adoption and assisted procreation was legally recognized for same-sex couples.

# **Queering the Welfare Regime Perspective: A Call for a New Focus**

In early works, Titmuss (1958) categorized social policies as 'residual' or 'institutional,' where the former term refers to weaker policy systems based on a safety-net in case of market failure, and the latter term refers to more comprehensive institutional systems that are often based on notions of social rights. Building on this categorization, the modern welfare regime literature accumulated, especially during the 1980s and 1990s. A cornerstone in this tradition is Gøsta Esping-Andersen's *Three Worlds of Welfare Capitalism* (1990), in which he outlines three main types of welfare regimes, i.e., the liberal, the social democratic, and the conservative, in which he claims that modern developed capitalist nations cluster.<sup>3</sup> He states that the ways in which nations choose to divide welfare responsibilities between the family, the market and the state are important. Esping-Andersen also discusses the concept of *decommodification*, i.e., social transfers and social insurances that enable workers to survive during periods when they are unable to sell their labor as a commodity in the market (Esping-Andersen, 1990, 1999).

Esping-Andersen's categorization has, among other things, been criticized for neglecting gender and for departing from an implicitly male, social class perspective (e.g., Lewis, 1992; Orloff, 1993). Lister (1994, p. 37) has termed the concept of *defamilialisation* (from here on referred to as defamilialization) to capture … *the degree to which individuals can uphold a socially acceptable standard of living, independently of family relationships, either through paid work or through the social security system*. This concept parallels the concept of decommodification. It highlights how many women would like to be commodified i.e., to have access to paid employment that frees them from their unpaid work and their economic dependency on a spouse, as well as to social security provisions in periods when they are unable to work (see Chapter 6 by Zagel and Lohman in this volume for a more detailed discussion of how this concept has been defined and used). Since then, the welfare and gender regime literature has been updated and modified, resulting in a number of publications on social citizenship, family policy frameworks, and gender regimes. A significant amount of this literature has focused on mothers' ability to combine periods of (infant) care with long-term access to paid work and to avoid falling into poverty if they are in an autonomous household (e.g., Keck & Saraceno, 2013; Korpi, Ferrarini, & Englund, 2013; Leitner, 2010, Misra, Budig, & Moller, 2007). An ever-increasing field of

<sup>3</sup>The categorization was far from encompassing and had a distinct Western European focus.

literature has also explored the extent to which family policies enable men to be caregivers as well as earners (e.g., Eydal et al., 2015; Grunow & Evertsson, 2016; Haas & Rostgaard, 2011; Saraceno & Keck, 2011). Taken together though, the literature largely implies the perspective of the heteronormative family, assuming one male and one female as partners in households that raise children.

The five countries in focus in this chapter differ to some extent in terms of their welfare and gender regimes. The Nordic countries belong to the social democratic welfare regime, with universalistic systems that promote the equality of good standards instead of minimal needs. This implies decommodifying welfare state services that reduce inequalities that are introduced by market-based access to services. The degree of defamilialization is considered high because of long job-protected parental leaves and highly subsidized public childcare. However, as typologies refer to one time point and are ideal types, there are ambiguous cases as well as transmutations (cf. Esping-Andersen, 1999, p. 86). In the Nordic countries, Finland has been discussed as a country that may be closer to the conservative cluster than the others. The Netherlands was originally part of the conservative regime cluster shaped by traditional and gendered family values that encourage family-based assistance dynamics. Over time, the Netherlands has moved toward a more mixed regime, wherein some parts of care work have become defamilialized (especially care for the elderly population) (Reimat, 2019).

The welfare and gender regime perspectives have had important theoretical as well as practical implications for how family policies were designed and implemented. However, today, new issues are on the agenda. Among these are a need to focus on the growing share of families that do not fit the heteronormative family norm. Historically, states have granted rights to men and women based on their statuses in families (Lewis, 1992). As wives and mothers, women have been granted certain rights with respect to maternity and parental leave, widowhood pensions and child benefits (Lewis, 1992). As workers, men have been granted with more or less social security rights, depending on the specific welfare regime of their state (Esping-Andersen, 1990, 1999; Korpi, 2000). Family leaves reserved for fathers have been either short or marginal. The gendered nature of policies linked to unpaid and care work, combined with heteronormative beliefs about proper gender roles, has contributed to same-sex couples facing different constraints than those faced by different-sex couples in regard to parentalization. The next step in the gender regime literature is thus, as we see it, an analysis of the extent to which same-sex couples are able to realize any parenting desires they may have and the extent to which policies support their shared caring if or when they have children.

## **Parentalization: A Gendered Concept**

We use the term *parentalization* as a concept divided into two parts, where the first has to do with individuals' or couples' ability to become parents. Evidence and experience show that the transition to parenthood is not only biologically but also socially facilitated for heterosexual couples. When different-sex couples have been unable to conceive, other avenues have been opened up, for instance, by adoption or fertility treatments. Access to these treatments was later granted to same-sex couples and has been implemented in ways that are not always appropriate for them. As an example, biological mothers-to-be often are given fertility stimulating medication that they may not need in order to become pregnant (Berg Hulthén & Nordqvist, 2017).

Parentalization is a gendered concept, in the sense that women, as a rule, are often deemed better fitted for care work than men. Evidence of a preference for women as parents over men can, for instance, be found in European court rulings linked to family policies, which privilege the mother– child relationship (McGlynn, 2001). Women also seem more motivated and able to negotiate family-friendly work conditions than do men (e.g., Harris & Estevez, 2017; McGlynn, 2001). In couples where there are either two mothers or two fathers, the norms and expectations linked to both heterosexuality and gender/parenthood are challenged. As we will show below, becoming a parent and making use of the policies facilitating the care of a child are currently, and have been, easier for female than for male same-sex couples. With a concept such as parentalization, we can critically evaluate the extent to which not only same-sex couples but also a broader range of couples who are not at the mental forefront of policy makers have the legal as well as the social rights that come with parenthood in various contexts and nations.

# **Parentalization in Practice. Part 1: Who Can Become a Parent and How?**

Same-sex couples' access to medically assisted insemination (MAI) has varied over time and among countries. Same-sex parenthood is legally recognized in the majority of European countries (Waaldijk, 2018). However, even when it is recognized, same-sex couples often have to go through time-consuming and seldom inexpensive processes to become parents. Once the child/children arrive, they are often longed for and the parents are well prepared. Moreover, research indicates that the children brought up in these families fare at least as well and manage school as well as children of different-sex parents (e.g., Aldén, Björklund, & Hammarstedt, 2017; Mazrekaj, De Witte, & Cabus, 2019; Watkins, 2018).

To determine whether and how same-sex couples can (i) become parents and (ii) realize their ideal ways of parenting (which may include sharing parental leave, care for sick children, etc.), we start from an overview of when and by which avenues same-sex parenthood became legally available in the countries in focus.4 It is worth noting that, of course, many same-sex couples became parents long before this, even if it meant not being legally recognized as parents.

#### **Marriage and Parenthood Rights, Joint and Second-Parent Adoption**

The first step toward legally recognized parenthood for same-sex couples has, in many cases, been registered partnerships and/or marriage rights, due to the link to adoption rights. Table 16.1 gives an overview of when the rights and legislation concerning same-sex couples' ability to form families were implemented in the four Nordic countries and the Netherlands. In 1989, Denmark became the first country in the world to legally recognize same-sex relationships and was soon followed by Norway (1993), Sweden (1995), and the Netherlands (1998) (Frantzen, 2011; Waaldijk, 2017).5 Finland legalized registered partnerships in 2002. The Netherlands was the first country to allow same-sex marriage in 2001. Joint within-country adoption has been available to same-sex couples in the Netherlands since 2001, and international adoption has been available since 2009.6 In Denmark, Finland, and Norway, the right of second-parent adoption, i.e., the possibility of adopting

<sup>4</sup>The information we provide herein will in time be outdated, as laws and regulations are continuously revised. Hence, we provide this overview mainly as a description of the different routes to same-sex couples' parentalization in these countries.

<sup>5</sup>Sweden had a law recognizing cohabiting homosexual couples since 1988. This law did not give same-sex couples rights similar to those of married couples, as did the partnership law of 1995, but it gave legal recognition similar to that of cohabiting heterosexual couples. https://www.qx.se/histor ien/110578/sambolagen-for-homos-infors/.

<sup>6</sup>In order to adopt, couples must have lived together for three years prior to the request (this also applies to different-sex couples).



aDalager Kjaer (2017)

bThe Adoption Act was changed in 2009; however, in 2008, courts started granting adoptions to lesbian couples who had undergone inseminations at Danish clinics soon after birth in anticipation of the law change. *Source* 2008/1 LSF 105 "Forslag til Lov om ændring af adoptionsloven og forskellige andre love (Adoption uden samtykke, stedbarnsadoption af registreret partners barn fra fødslen m.v.)" (https://www.retsinformation.dk/eli/ft/200812L00105)

cIn 2006, the Norwegian government changed the interpretation of the guidelines for second-parent adoption (Q-1045) so that if the child was conceived at a clinic using a non-anonymous donor, and the parents were in a stable registered partnership, then the adoptions process could start directly after childbirth. However, between 2002 and 2006 this principle was already being used by some regions and case workers due to unclarity in the original guidelines. *Sources* Barne-, likestillings- og inkluderingsdepartementet (2016) and correspondence with Martin Eckhoff Andresen. The new directives from 2006 can be found here: https://www.aftenposten.no/norge/ i/G3pn9/Brev-om-stebarnsadopsjon

dCouples that are married, cohabiting or in a registered partnership can adopt. Same-sex partners cannot adopt their partner's child if the child was originally adopted from a country that does not permit same-sex adoptions. *Sources* Barne-, ungdoms- og familiedirektoratet (bufdir.no) and correspondence with Torunn Elise Kvisberg


a spouse's child, proceeded the right of joint adoption (see Table 16.1).7 In Sweden, joint and second-parent adoption was made possible for same-sex parents in 2003. The right to joint adoption has been more of a formal than a practical right, as very few same-sex couples have been able to adopt a child. In some countries, second-parent adoption has been an increasingly important avenue to parenthood for male couples as a result of surrogacy arrangements.

### **Medically Assisted Insemination/IVF and Female Couples' Transition to Parenthood**

For female same-sex couples, the right to medically assisted procreation has facilitated the transition to parenthood significantly. In the Netherlands, MAI is and always has been available to all women; there is no legislation limiting the categories regarding who can receive MAI or in vitro fertilization (IVF) (Nikolina, 2017a, 2017b). However, until 1998 (for MAI) and 2001 (for IVF), women in same-sex relationships or those without a partner were dependent on hospitals that were willing to assist them in medical procedures. Since 2002, the wife or registered partner of the birth mother can become a legal parent if the donor is unknown to the parents (i.e., from a sperm bank) but not anonymous (the child can determine who the donor is at age 16), by either legally recognizing the child before birth or via adoption after the birth. Since 2014, the wife/registered partner automatically becomes the child's legal parent in the case of an unknown donor. In practice, this means that the marriage presumption that covers different-sex couples—which by default assumes that the man of a married birth mother is the child's other parent—as of 2014 also includes female same-sex couples. Using a private, known donor, such as a friend or a relative, is more common in the Netherlands than in the Nordic countries; private insemination at home has been widely advocated in the Netherlands for a long time.8,9

<sup>7</sup>It is worth noting that in Finland and the Netherlands, the non-biological parent could also gain parental authority, i.e. represent the child and make decisions on his or her behalf, under certain conditions before formal adoption was legally available, if the legal parent consented (Nikolina, 2017b; Valleala, 2017).

<sup>8</sup>See for example: https://www.freya.nl/brochures/zelfinseminatie/.

<sup>9</sup>In addition, it seems private donors can more easily choose not to legally acknowledge biological fatherhood in the Netherlands. A biological father who is not married to or in a registered partnership with the birth mother needs the birth mother's written permission in order to legally acknowledge the child and cannot do so if the mother's (male or female) legal partner has already acknowledged the child. In Denmark, a known donor can avoid legal fatherhood, provided that the insemination is performed at a fertility clinic (Dalager Kjaer, 2017, p. 8). In Sweden, the same thing is possible if the donation is done through the public health care system in which the donor is first approved. In

As Table 16.1 shows, MAI and IVF have been possible for female samesex couples since 1992 in Denmark, 1997 in Finland,10 2005 in Sweden, and 2009 in Norway. In Denmark, this right was taken away from lesbian couples in 1997, until private clinics found a loophole in the law in 1999.<sup>11</sup> In the Nordic countries, there is no marriage presumption for same-sex couples, and the social mother has to go through a process similar to the one that cohabiting different-sex couples go through in order to legally verify their parenthood (see for Denmark: Dalager Kjaer, 2017; Norway: Eckhoff Andresen & Nix, 2019; Sweden: www.rfsl.se; and Finland: Moderskapslag 20.4.2018/253). In Sweden, until 2019, if the child was conceived abroad,12 the social mother had to adopt the child, and for her to adopt, the couple needed to be married/registered partners. Today, if the donor is non-anonymous and the MAI/IVF is carried out at a certified clinic, then the social mother can confirm her parenthood and adoption is not needed (Zimmerman & Nordqvist, 2018). In Denmark, the social mother initially had to adopt the child to become legally recognized as a parent (Table 16.1). However, since 2013, adoption is not needed when the child is conceived with MAI/IVF. In Norway, the birth mother's partner (via marriage or in a stable cohabiting relationship) can register as the child's legal parent from birth given that the biological mother consents and the child was conceived at a clinic using a non-anonymous donor (Eckhoff Andresen & Nix, 2019; Frantzen, 2011; SOU, 2016:11).13 Being somewhat of a laggard, in April 2019, Finland passed a citizens' initiative that made it possible for both women in same-sex couples to be recognized as mothers from the moment their child is born (seta.fi, Moderskapslag 20.4.2018/253).

It has not been uncommon for same-sex couples to jointly have children with a single mother/father or another couple. However, a child can

Norway and Finland, however, a man who donates to e.g., a friend cannot avoid being recognized as the father. If a known donor is used outside the above-mentioned conditions, he is viewed as the child's father by law, and the social mother has to adopt the child with the donor's consent (while he gives up his legal parenthood).

<sup>10</sup>In Finland, medically assisted reproduction has been legally available to all women since 1997 but has in practice only been available to lesbian couples at (some) private clinics. As of fall 2019, public clinics are also opening up for these couples (Valleala, 2017 and correspondence with Anna Moring, Monimuotoiset perheet (*Finland for all families*) and Juha Jämsä, Sateenkaariperheet (*Rainbow Families Finland* ).

<sup>11</sup>In October 1999, midwife Nina Stork opened a private infertility clinic in Copenhagen. The clinic, *StorkKlinik*, offered donor insemination regardless of "the ethnic background, religion, civil status or sexual orientation" of its clients (Nordisk Ministerråd, Nordisk Råd, & Nordic Committee on Bioethics, 2006, p. 59).

<sup>12</sup>Many go to Denmark to conceive to shorten the wait, to be able to have an anonymous donor, or in order to meet more detailed preferences regarding the donor.

<sup>13</sup>This was possible from 2009 onward (and adopted by a law amendment in June 2008, according to personal communication with Kirsten Sandberg, Institutt for offentlig rett, University of Oslo).

only have two legal parents in the Nordic countries. Starting in December 2019, a child can have more than two (but not more than four) recognized guardians in Finland.14 This policy, among other things, helps individuals who have what is referred to as a special and parental-like relationship with a child to spend time with the child and have the child stay with her/him (*Lag ang. vårdnad om barn och umgängesrätt* 8.2.2019/290). Similarly, in the Netherlands, children might soon be able to have up to four guardians. In the current governmental proposal, non-biological parents are given 'partial custody,' which enables them to take the child to doctor's visits if he/she is sick and to have input on where the child goes to school, among other things (Pieters, www.NLTimes.nl, July 12, 2019).

#### **Male Couples' Transition to Parenthood**

For male same-sex couples, the possibilities by which they can become parents have always been more limited. Although joint adoption is allowed in the countries in focus in this chapter, the waiting lists have been long due to the very low numbers of children put up for within-country adoption and many countries internationally not allowing their children to be adopted by samesex couples. In addition, having a child via surrogacy is controversial and highly debated, although more in some contexts than others. In the Netherlands, commercial surrogacy15 is illegal, yet altruistic surrogacy arrangements have been possible for different-sex couples under strict conditions since 1997 (Vlaardingerbroek, 2003). The surrogate should be someone known to the prospective parents, as they are not allowed to advertise for a surrogate. However, until January 2019, same-sex couples did not have access to highly technological surrogacy, which means that both the egg and sperm were donated and carried by a third person.

In the Nordic countries, surrogacy is either illegal or unregulated. All Nordic countries prohibit fertility treatment if the woman receiving the treatment is planning to give the child up for adoption, although this was not legally regulated in Denmark and Finland until 1997 and 2006, respectively (SOU, 2016:11). Today, commercial surrogacy (i.e., paying more than medical costs to a surrogate mother) is illegal in Denmark and Finland, and

<sup>14</sup>A legal guardian is not the same as a legal parent, even though the roles may overlap.

<sup>15</sup>Altruistic surrogacy occurs when a surrogate carries a child with no additional compensation other than reimbursement for medical costs and other reasonable pregnancy-related expenses. She should also be personally known to the prospective parents. Commercial surrogacy is when a surrogate is compensated for her services beyond such expenses. Commercial surrogacy also includes surrogates offering their services openly or parents openly announcing for the services.

if the court finds that a couple has used a commercial surrogate, it may result in the adoption not being granted to the non-biological parent.<sup>16</sup> Swedish and Norwegian legislation does not prohibit paying for a surrogate but at the same time does not provide any legal framework for this type of arrangement (SOU, 2016:11). Hence, for Swedish and Norwegian couples, paying for surrogacy in another country is not illegal, and this has helped registered partner/married gay couples—not the least if they are fairly well off—to become parents (see Malmquist & Spånberg Ekholm, 2020, for Sweden). If the surrogacy arrangement took place abroad, then the paperwork (including the migration of the child) may leave the child without a legal parent or recognized guardian in the country for a while (SOU, 2016:11). Shared parenthood may be easier to realize, and it is not uncommon for male couples to privately arrange for and have children with another couple or a single woman. Again though, in the countries discussed herein, a child can have only two *legal* parents.

### **Number of Children Less Than One Year of Age in Same-Sex Couples in Five Countries**

In the following section, graphs are presented that show the number of children zero years of age in female and male same-sex couple households in the countries in focus. The graphs are based on population register data from each country.17 They start in the year that the union formation of same-sex couples was legally recognized or when cohabiting same-sex couples with children can be identified in the data.18 Cohabiting couples have been included since 1990 in the Netherlands and since 2009 in Norway, and female cohabiting couples have been included since 2005 in Sweden. It is worth noting that the size of the populations in these countries differs considerably. The Netherlands has the largest population (17 million in 2017) and here we also see the largest number of female couples transitioning to parenthood (Fig. 16.1). The trend starts before registered partnership was legally recognized, and it does not

<sup>16</sup>In all four Nordic countries, a woman who gives birth is considered to be the child's legal parent and, if she is married to a man, her husband is considered to be the legal father. Hence, achieving joint parenthood through adoption by a social parent requires the surrogate mother's consent. If she is in a heterosexual marriage, her husband also needs to consent. If none of the parents-to-be are the biological parents, then the couple needs to jointly adopt the child.

<sup>17</sup>Data was delivered from Statistics Denmark, Statistics Finland (*Statistikcentralen*), Statistics Norway (SSB) and Statistics Sweden (SCB). In addition, data from Statistics Netherlands was used.

<sup>18</sup>As legal changes that recognize same-sex couples and parents make them easier to identify in the data in the later years, we most likely underestimate the number of children more so in the earlier than in the later years.

**Fig. 16.1** Number of infant children (zero years old) in households with a female same-sex couple (*Note* Due to data limitations, only couples who are in registered partnerships or are married can be identified in Denmark and Finland. In Sweden [from 2005], Norway [from 2009] and the Netherlands [from 1990] cohabiting couples can be identified and are thus included in the figure)

seem to be affected by legal changes. In 2018, 700 children less than one year of age were living with two mothers. In Denmark, which was the first country to legalize registered partnerships in 1989, a slow but steady increase in the number of (new) children in lesbian couples can be noted, mainly since 1995. Second-parent adoption became possible starting in 1999 and from 2007 onward, when MAI/IVF was (again) possible in public clinics, we see a more rapid increase in the number of children. For Sweden, there has been a clear increase in the couples transitioning to parenthood since 2002; this is particularly true from 2005 onward, when MAI/IVF was made available to lesbian couples. Sweden has a larger population than that of Denmark, Finland, and Norway (10 million compared to approximately 5.5 million in 2016), and this may explain the steeper increase in the number of children less than one year of age in Sweden. Norway legalized registered partnerships early (1993) but did not allow MAI/IVF for lesbian couples until 2009, after which we see the trend for Norway catch up with the numbers for Denmark. In Finland, the increase in the number of children less than one year of age starts later (registered partnerships were legalized in 2002), and it is similar to the trends in the early periods for both Denmark and Sweden. It is worth

**Fig. 16.2** Sweden: number of infant children (zero years old) in households with a male same-sex couple who are married or in a registered partnership

noting that in Finland, MAI/IVF was not available to lesbian couples in all clinics during the period covered. This may explain the more modest increase in the number of couples becoming parents. Due to the inequalities this has caused and based on threats of financial sanctions for discrimination, public health care and fertility centers are increasingly opening up to lesbian couples and single women (as of fall 2019).19

In Figs. 16.2, 16.3, and 16.4, the number of children less than one year of age in male same-sex couple households in Sweden, Denmark and Norway are presented. Not surprisingly, the number of male couples transitioning to parenthood is much smaller than the number of female couples transitioning to parenthood in these countries (compare the *y*-axes in Figs. 16.1, 16.2, 16.3, and 16.4).20 Sweden and Norway have notably larger numbers of male couples with children than that of Denmark, while in Finland, we identify too few male couples to be able to present a graph.21 Denmark and Finland are also the countries with the most restrictive legislation on surrogacy, which is a common way for male same-sex couples to have children. For the Netherlands (Fig. 16.5), we again find a number of gay couples living with a child less than

<sup>19</sup>Personal communication with Anna Moring, Monimuotoiset perheet, Finland.

<sup>20</sup>Due to data restrictions, we cannot report on years when fewer than three couples make up the bar charts. Consequently, we use bars that are accumulated for some years.

<sup>21</sup>Very few male same-sex couples where both fathers were legally recognized as parents can be found in the data for Finland. The low number has been confirmed by Sateenkaariperheet (*Rainbow Families)* Finland. Some additional couples in which one of the partners became a biological father but where the child cannot be identified as living in the household can be observed but are not included in the graph.

**Fig. 16.3** Denmark: number of infant children (zero years old) in households with a male same-sex couple who are married or in a registered partnership

**Fig. 16.4** Norway: number of infant children (zero years old) in households with a male same-sex couple who are in a registered partnership, married or cohabiting (from 2009)

one year of age from the start of the observation period in 1990 onward. Nevertheless, the number of Dutch male couples living with a child below one year of age has been smaller than that in Sweden in more recent years, most notably since 2010 (cf. Fig. 16.2).22 The difference may be explained by more male couples having children and sharing parenting with single women

<sup>22</sup>Given that very few couples have been able to jointly adopt a child, we do not think that we lose a lot of adoptive couples due to the age restriction applied in the graphs.

**Fig. 16.5** The Netherlands: number of infant children (zero years old) in households with a male same-sex couple who are cohabiting, married or in a registered partnership

or lesbian couples in the Netherlands. In the figures discussed here, as well as those seen in Fig. 16.1, we cannot identify children who partly live in the household if they are registered at another address.

# **Parentalization in Practice. Part 2: Sharing the Care of the Child**

## **Family Leave Rights**

In this section, we discuss same-sex couples' family leave rights and the extent to which they have differed during periods—if at all—from the rights of different-sex couples. This exploration enables an understanding of the second step of the parentalization process, namely, the ability of both parents (or more than two parents) to make use of family policies and share any jobprotected and potentially subsidized leave to care for a child. As a basis for the comparison, we provide information on parental leave policies in separate fact boxes. This information often stems from the yearly reports from the *International Network on Leave Policies and Research* (see: www.leavenetwork.org). We use the terminology applied in the reports (based on the concepts used in laws and policies). For instance, this means that paternity leave would be the leave assigned to the social mother in a female same-sex couple. In addition to the publicly funded family leave policies in the various countries, many parents are covered by collective employer agreements that give them a higher

wage replacement than that which they receive by public social insurance. This means that the income obtained during leave for many is higher than the earlier earnings levels presented in the fact boxes.

Although we generally think of the legal recognition of same-sex parenthood as the first step towards family formation and the practical prerequisites of sharing the care of the child as the second step, this has not been the sequence of events in Sweden. Ever since the right to registered partnership was introduced in 1995, biological parents have been able to share parental leave with a registered partner/married spouse. In other words, if, for instance, a lesbian couple privately arranges for and becomes pregnant in partnership with a gay couple (i.e., there is one biological mother and one biological father who want to be active parents jointly with two social parents), then the biological parents can sign over parental leave days to their partners such that, in total, four parents can share leave for the child if both biological parents are in registered partnerships/married (but not to each other). Parental leave is linked to the child in Sweden and in our example, the total leave time that the four parents can share will be the same as that for a two-parent family.

In Norway, the ability to share parental leave was not linked to registered partnership, marriage or adoption but was granted to same-sex couples in separate legislation in 2007 (Table 16.1). In Denmark, legal parenthood and access to family policies did not coincide. When MAI/IVF was made available to lesbian couples in all clinics in 2007 (it had been available in private clinics since 1999), it was framed as a health law rather than a family law (Jeppesen de Boer & Kronborg, 2012). As such, it was more influenced by health equality principles than by family law and parental rights. This meant that same-sex and adoptive parents could not share parental leave until the Maternity Act was revised in 2009 (Tølbøll, 2014). In Finland, starting in 2007, the registered partner of the birth mother had the right to parental leave if the child was born after the registered partnership started and if the social mother was living with the child and the child's legal parent (Valleala, 2017). In 2017, the law was changed so that a woman who gives birth can share her parental leave with a partner regardless of their gender and civil status. However, biological fathers do not have the same option.

#### **Family Leave in Sweden**

Same-sex couples who are married or in registered partnership have been able to share family leave in Sweden since registered partnership was introduced in 1995. When same-sex parenthood was legally recognized in 2003, the number of female couples transitioning to parenthood started growing. At this time, parents had the right to 480 days of parental leave with a child, which could be shared between the parents, with 60 days being reserved for each parent.<sup>23</sup> Since 2016, 90 days have been reserved for each parent. For 390 of the 480 days, the reimbursement is approximately 80% of the previous earnings, up to a ceiling. In addition, the remaining 90 out of the 480 days are reimbursed at a lower flat rate. Parents also have the right to temporary parental leave at the time of birth, which consists of ten days for the parent not giving birth, to be used within 60 days of delivery. Those who have worked less than 240 days before the expected delivery or who have had a low daily income during this period receive the low flat rate during the entire leave period. When parents adopt a child, they get the same parental leave insurance (in terms of days and benefits) as parents who have a biological child.

*Sources* Chronholm (2005), Duvander and Haas (2018), Korsell and Duvander (2019).

One group that may have difficulties using parental leave are male couples who have become parents with the help of a surrogate abroad. The migration and adoption process may take time and consequently, the child may be without a legal parent in the country for months. For example, in Sweden, when the surrogacy mother is not a citizen, then she cannot make use of the parental leave system and neither can she sign over leave to the fathers. Instead, the fathers may need to use vacation time or leaves of absence to stay home with the baby (Malmquist & Spånberg Ekholm, 2020). This differs from Dutch parental leave regulations that enable a person to take leave for a prospective adoptive child (as well as for a foster child and a step-child), as long as the child is registered at the same address as the person taking the leave (Rijksoverheid, the Netherlands, downloaded on 14.2 2020). In Denmark, only those who are legally recognized as the parents of the child can make use of family leave. Second-parent adoption rules complicate things for some couples due to the 2.5 years that the parent must live with the child before s/he can adopt (see Table 16.1). This policy mainly affects lesbian couples who have not conceived at a clinic and male couples in which one of the partners is the biological father. The rules are similar in Norway, with a few exceptions. First, second-parent adoption can take place after the parent has lived with the child for five years (unless the child was conceived through MAI/IVF at a clinic). Second, if only one legal parent has parental responsibilities or if there are two legal parents but one parent is not using leave, then the spouse (if any) of the legal parent with parental responsibilities may use unpaid family leave from work (but not parental leave benefits) (Eeg, 2017).

<sup>23</sup>This assumes two-parent families and that the reserved days are for the two legally recognized parents. If there are additional social parents, no days are reserved for them (even though they may take leave).

In Finland, only a few gay couples have attained legal parenthood status for both partners after second-parent adoption was made possible in 2009 and have thus been able to both take parental leave.

#### **Family Leave in Denmark**

MAI/IVF has been available to female same-sex couples since 1999 in private clinics and since 2007 in all clinics (apart from the period of 1992–1997 when it was also available). However, it was not until 2009, with the revision of the Maternity Act (Barselsloven), that same-sex parents and adoptive parents were able to share parental leave (Tølbøll, 2014). The maternity leave insurance covers 18 weeks, specifically, four weeks before and 14 weeks after birth. Paternity leave is two weeks. Parental leave is 32 weeks and can be used until the child is 48 weeks of age. Parental leave is individual; each parent can claim at most 32 weeks of leave, but the total cash benefit of parental leave per family cannot exceed 32 weeks (excluding maternity and paternity leaves). The benefit levels of maternity, paternity, and parental leaves were 90% of the previous earnings up to a ceiling in the beginning of the period, and a cash benefit of at most approximately 580 EUR per week before taxes at the end of the period (the benefit cannot exceed the previous earnings). Adoptive parents have the same leave rights as other parents, with the exception that two of the total 48 weeks need to be taken by both parents jointly. Eligibility for leave is conditional on the parent working at least 120 hours in the 13 weeks preceding the leave.

*Sources* Rostgaard (2009), Bloksgaard and Rostgaard (2019).

#### **Family Leave in Norway**

Starting in 2007, a social mother in a female same-sex couple could share the parental leave with the birth mother given that she was recognized as the parent of the child at the start of the leave.<sup>24</sup> At that time, parents could share either 29 weeks (49 weeks as of 2018) of leave at 100% of the earlier income or 39 weeks (59 weeks as of 2018) at 80% of the previous income up to a ceiling, of which six weeks (ten weeks as of 2018) were reserved for each parent and three weeks (*svangerskapspermisjon*) needed to be taken before the birth. To be eligible for the earnings-related benefits, the parent needs to have been employed (or self-employed) for six of the last ten months prior to the birth and to have earned at least a given minimum amount of income during the previous year. Those not eligible receive a flat rate payment. A paternity leave of two weeks after the birth is unpaid by the government. Adoptive parents get the same amount of leave as couples in which one has given birth to the child, with the exception of the *svangerskapspermisjon* (the three weeks reserved for before a birth).

*Sources* Brandth and Kvande (2007, 2018), Eckhoff Andresen and Nix (2019), Eeg (2017).

#### **Family Leave in Finland**

Since 1997, lesbian couples have had access to assisted reproduction techniques in private clinics in Finland. Since 2007, due to a change in the Health Insurance Act, the birth mother's spouse or partner has had access to parental leave entitlements (Valleala, 2017). In 2002, when registered partnership was introduced, maternity leave consisted of 105 working days (a work week was defined as six days), of which 30–50 days must be taken before the birth. Paternity leave consisted of 18 working days, and parental leave consisted of 158 working days per family. Leave benefits were earnings related, averaging 66% of the previous earnings.

As of 2017, a woman who gives birth can share her parental leave with a partner regardless of gender (biological fathers do not enjoy this option). During the first 56 days of the 105 days of maternity leave, the payment is 90% of the annual earnings up to a ceiling; after this point, the payment is 70% of the earnings (as of 2018). Entitlement is linked to residence and is paid to all who have lived in Finland (or have been insured in another EU member state) for at least 180 days before the birth. Mothers who are not employed or who have low annual earnings receive a minimum flat-rate benefit. As of 2018, paternity leave consists of 54 working days, and both paternity and parental leave (the latter consists of 158 working days, as of 2018) are reimbursed at 70% of the previous earnings. The adoptive parents of a child younger than seven years of age but older than 2 months of age are eligible to 200 working days of leave (234 if the child is younger than 2 months old).

*Sources* Salmi and Lammi-Taskula (2005), Salmi, Närvi, and Lammi-Taskula (2018).

In the Netherlands, maternity, paternity and parental leaves are for employees only. Birth mothers and their partner or the person acknowledging the child are eligible (Rijksoverheid, the Netherlands, downloaded on 14.2 2020). Unlike the other countries in this chapter, parental leave is strictly personal and cannot be signed over. This means that single parents have only 26 weeks of leave in total, whereas couples have 52 weeks. Leave is linked to legal parenthood (or acknowledgment of the child, see the fact box). When a child has more than two adults who share parenting roles, then the persons who are not legally recognized as parents cannot take parental leave.

<sup>24</sup>In 2007, the law was changed so that a person who adopts the child of their same-sex partner can use parental leave. Leave could also be taken by a married spouse of the birth mother if there was no other legal parent or if this parent was not using leave. In 2006, the Norwegian government changed the interpretation of the guidelines for second-parent adoption (Q-1045) so that if the child was conceived at a clinic using a non-anonymous donor, then the adoption process could start directly after childbirth. However, between 2002 and 2006, this principle was already being used by some regions and case workers due to unclarity in the original guidelines. *Sources* Barne-, likestillingsog inkluderingsdepartementet (2006, 2016) and correspondence with Martin Eckhoff Andresen and Kirsten Sandberg. The new directives from 2006 can be found here: https://www.aftenposten.no/ norge/i/G3pn9/Brev-om-stebarnsadopsjon.

#### **Family Leave in the Netherlands**

Parental leave was introduced in 2001 in the Netherlands, which was the same year that same-sex marriage and within-country adoption were legalized, and it immediately became available to different-sex as well as same-sex couples. Birth mothers have 16 weeks of fully paid maternity leave, which is regulated by the national government. All female employees have the right to maternity leave, and those who are self-employed are compensated up to the minimum wage. Paternity leave is fully paid for five days, and as of July 1, 2020, an additional 5 weeks at 70% of earlier earnings can be added. This leave is granted to the employed partner of the birth mother or a person acknowledging the child. The birth mother and her partner can each take unpaid parental leave up to 26 times their weekly work hours per child. In 2018, adoptive parents had four, and since 2019, they have six weeks of leave with payment equivalent to maternity leave. Adoptive parents have the same parental leave rights as other parents.

*Sources* Den Dulk (2018), Nikolina (2017b).

### **Parentalization: Summary and Conclusion**

In this chapter, we present the concept of parentalization. Parentalization is defined as the ability to become a parent and be recognized as such, both legally and via social policies. The concept aims to address a gap in the welfare, social policy, and gender regime literature. A key concept in this literature is defamilialization and the extent to which women can be freed from economic dependency on a spouse and covered by social policies when performing care work in the household. The contributions made by researchers engaging with this concept have been imperative in regard to mapping gender inequalities in welfare states. The concept problematizes dependencies and imbalances in heterosexual couples and families but is not applicable to the same extent to other family constellations. One group that faces varying degrees of difficulties regarding (i) being legally recognized as parents and (ii) making use of family policies in various contexts is same-sex couples. In this chapter, we focused on five countries that were among the first to legally recognize same-sex parenthood: the Netherlands, Denmark, Finland, Norway, and Sweden. By mapping the road to legal parenthood and family policy rights for same-sex couples in these countries, we identified obstacles and impediments on the road to parentalization. We also provided evidence of the importance of legal, parental, and policy rights for the transition to parenthood among female and male same-sex couples by graphing the increase in the number of registered partnered/married couples with a child less than one year of age in the study nations. Even if many such couples transitioned to parenthood before same-sex partner and marriage rights were in place (though we are unable to fully map this with the data at hand), the legal recognition of parenthood came with increased certainty and security in the couples' everyday lives, as well as in case of parental separation or divorce. In the following, we sum up and discuss some of the hurdles to full parentalization that same-sex couples have had to overcome and/or to some extent are still facing in the nations focused on in this study.

Starting with the ability to transition to parenthood and be legally recognized as parents, same-sex couples still face legal obstacles to parenthood in many countries. In the countries discussed in this chapter, many obstacles have been removed, but it remains difficult for some to attain parenthood. This leads us to the conclusion that the first prerequisite (i) is to some extent fulfilled, at least for female couples. In practice, the number of possible routes to parenthood differs for same-sex couples compared to different-sex couples. Very few same-sex couples have been able to jointly adopt a child, as adoption rates have decreased (Mignot, 2019) and many countries and adoption agencies are reluctant to put children up for adoption by same-sex parents. These issues have made it difficult for male couples to realize their (any) parenting desires. Another example is the marriage presumption that applies to different-sex but not same-sex parents in many countries. Sweden is one example, where a married social mother is not recognized as the child's parent from birth but must go through a bureaucratic process (the same as that for cohabiting different-sex fathers), in order to be recognized as a legal parent of her child. In Finland, MAI/IVF was not available to lesbian couples in all clinics until recently. Private clinics are often more expensive than public clinics (as costs are more often subsidized in the latter). After threats of financial sanctions based on the discrimination, starting in the fall 2019, public clinics in Finland have been increasingly opening up to lesbian couples and single women.

Turning to the second criteria (ii), some family policies may enable samesex parents to share in the care of a child in the same way as that of different-sex couples, while some may not. In Denmark, there was a delay between providing lesbian couples with MAI/IVF in public clinics and granting both parents with policy rights to take parental leave with the child. In Sweden, same-sex parents who were registered as partners could share parental leave even before they were legally recognized as parents, as parental leave is linked to marriage and cohabitation. Today, legally recognized samesex parents have the same family leave rights as different-sex couples have in all the countries in focus. However, whereas a woman who gives birth can share the leave with a partner independent of gender and civil status in Finland, biological fathers does not have the same option. In addition, second-parent adoption rules may delay parental leave use or render it impossible to use in time in Denmark and Norway, where a non-biological parent needs to wait 2.5 and 5 years, respectively, before they can adopt the child, if the child is not conceived through MAI/IVF at a clinic.

As indicated above, parentalization is more readily available to some samesex couples than others. We argue that this partly reflects the dominant norms on gender and motherhood. Married female couples face fewer barriers to parentalization than other couples do. This is not only linked to the partners' ability to become pregnant and carry a child. We argue that part of the reason is also that female couples more closely resemble the heteronormative family. Ideals about a child's need of a mother are strong, and women are expected to want to have children and to be more child-oriented and better caretakers than men (e.g., Grunow & Evertsson, 2016, 2019). This expectation is partly mirrored in family leave policies where paternity or co-parent leaves are often short and seldom paid (the Netherlands being the obvious example in our case). In addition, the topic of surrogacy is highly debated, with many advocating the rights of women not to be exploited by any form of surrogacy, whereas others are less critical to altruistic surrogacy compared to commercial surrogacy (cf. Baker, 1996; Roach Anleu, 1990). Commercial surrogacy is not legal in any of the countries in focus in this chapter. However, the degree to which it is legally punished varies and partly seems to influence gay couples' abilities to become parents. Examples include Sweden and Norway, where surrogacy is not legally regulated for those who go abroad, which seems to contribute to a larger increase in gay parenthood in these countries (cf. Malmquist & Spånberg Ekholm, 2020; see also Figs. 16.2 and 16.4). Currently, altruistic surrogacy is allowed in the Netherlands, and it seems more accepted (although it is still not legally regulated) in Denmark than in the other Nordic countries, as indicated by media coverage.25 Studies analyzing the practices that surround the surrogacy process in countries that ban commercial surrogacy, in countries that allow altruistic surrogacy, and in countries without surrogacy regulations would enable a more informed discussion about the pros and cons—and intended and unintended consequences—of various laws and regulations for parents(-to-be) and surrogates, as well as children.

<sup>25</sup>Examples are CPH Post Online: http://cphpost.dk/?p=86936, DR.dk: https://www.dr.dk/nyheder/ indland/overblik-maa-man-bruge-rugemor-i-danmark-her-er-reglerne, both downloaded on September 13, 2019.

Even if male same-sex couples have greater difficulties in becoming parentalized than do female same-sex couples, the former still find it easier to fit the definition of a family than do other groups such as single gay fathers or multi-parent families (e.g., Carroll, 2018; Downing, 2013). Hence, the analyses and discussions in this chapter only scratch the surface in regard to parentalization in a broader range of couples and families. In the countries in focus in this chapter, a child can only have two legal parents; thus, any social parents in a multi-parent family are not recognized by law. However, even in this context, we see signs of change in the Netherlands, where children soon should be able to have up to four guardians, and in Finland, where a child can have three or possibly four guardians, as decided by the court. These are important advances that will influence not only so-called pink or rainbow families but also heterosexual, step- and bonus-parent families. Facilitating for the latter families was also one of the main motives spurring these changes.

Another group of families that can be analyzed in terms of parentalization is that of single-parent families. As indicated above, in the Netherlands, parental leave rights are individual. For single parents, this may mean that they may be forced to send the child to day care earlier than hoped for, or it may force them to drop out of paid work altogether. In this case, the second part of parentalization, i.e., the access to parental leave rights, is in place, but in practice, such access produces inequalities for children based on family type and it complicates for some parents to realize their parenting and care ideals. Another group of parents that can be analyzed within the parentalization concept are couples in which one or both partners have migrated. The extent to which both partners can share legal parenthood and care via social policies, as well as any waiting periods prior to the time when such sharing is realized in various contexts, could efficiently be mapped using the parentalization concept. Here, we can think of different potential problems linked to whether the couple is married or not, whether the birth mother (if any) or the other parent is a citizen or not, or whether they have both migrated. Lesbian couples who migrate to Sweden risk losing their joint legal parenthood due to a Swedish law stipulating that the social mother can only be the parent from birth if the child was inseminated either at a Swedish clinic or more recently—at a foreign, certified clinic using a non-anonymous donor. Focusing on couples that migrate no doubt adds layers of complexity in terms of parentalization that we have not dug into at all in this chapter. Hopefully though, our first attempt to establish and apply the concept will spur such analyses.

Parentalization, which is the ability to (i) transition to parenthood and (ii) make use of family policies and share the care of the child, varies considerably between countries and groups of parents, even when we focus on five countries that are at the forefront in regard to same-sex couples' legal rights to marry and have children. Lurking in the background are norms linked to the heterosexual family ideal, to motherhood and fatherhood, making parentalization more or less challenging for the groups of parents who do not fit within the regular mold. Future empirical research of same-sex couples could study patterns of fertility and parity and possible inequalities within this group, for instance, by family type or by family income. By mapping out and understanding how the transition to parenthood and the process of parentalization work for same-sex couples, we can evaluate how countries either reproduce normative systems or become more inclusive by going beyond the heteronormative family.

**Grant Acknowledgements** This research has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program awarded to Marie Evertsson (grant agreement No. 771770). **Other Acknowledgments** We would like to thank a number of researchers and country experts for responding to questions on short notice, as well as for providing us with comments and suggestions on the text. Among them are Martin Eckhoff Andresen, Peter Fallesen, Jennifer Hook, Juha Jämsä, Anna Malmquist, Anna Moring, Michael Nebeling Petersen, Nicole Ovesen, Tine Rostgaard, Kirsten Sandberg, Tone Sverdrup, Aslak Syse and Hannah Zagel. Any remaining errors or mistakes are on us. We have also benefitted from comments and suggestions from participants at the conference *Towards Resilient Nordic Welfare States—Challenges, Responses and Consequences* in Helsinki in March 2019 and the *Community, Work and Family in Diverse Contexts and Changing Times* conference in Malta in May 2019. Finally, we are also grateful to the research group at the social policy unit at SOFI, Stockholm University for providing us with comments and suggestions, especially Kenneth Nelson and Sebastian Sirén.

# **References**


*stebarnsadopsjon)*. Retrieved from https://www.regjeringen.no/no/dokumenter/ otprp-nr-5-2006-2007-/id430437/?ch=1.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Part IV**

**Sub-national**

# **17**

# **Breaking the Liberal-Market Mold? Family Policy Variation Across U.S. States and Why It Matters**

**Cassandra Engeman**

In cross-national comparative studies, the United States is understood as the ideal-typical liberal regime that leaves work–life reconciliation for individuals and markets to address (Gornick & Meyers, 2008; Korpi, Ferrarini, & Englund, 2013). Fitting the liberal-market characterization, the United States lags behind most other countries in the world with respect to maternity and parental leave provisions (Heymann, Earle, & Hayes, 2007). The Family and Medical Leave Act (FMLA), the only federal leave law in the United States, provides up to 12 weeks of job-protected leave for self-care or care of a parent, spouse, or child to address a serious illness or health condition. Leave provisions are unpaid, meaning that leave-eligible employees are guaranteed a return to the same position or a similar position with the same employer but do not receive wage-replacement benefits while on leave. Additionally, the FMLA's eligibility requirements effectively exclude workers employed part-time, in new jobs, or in small firms.1 According to a 2012

C. Engeman (B)

Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden

e-mail: cassandra.engeman@sofi.su.se

<sup>1</sup>Eligibility for leave under the FMLA requires employees to work at least 1250 hours in the year prior to taking leave from an (or each) employer. The employer must have 50 or more employees within a 75-mile radius.

study commissioned by the U.S. Department of Labor, the FMLA's eligibility requirements restrict over 40% of the workforce from the law's coverage (Klerman, Daley, & Pozniak, 2014). Additionally, a majority of leave-eligible workers who report not taking leave when needed cite an inability to afford unpaid leave from work (Klerman et al., 2014; Waldfogel, 2001). The lack of paid, job-protected leave is most prevalent among women, low-wage workers, and workers of color (Milkman & Appelbaum, 2013).

However, there is considerable subnational variation in U.S. family policy that cross-national comparative research often overlooks. The FMLA applies across states and provides base-level rights upon which state and local (municipal/county) laws can expand. Given the FMLA's limited coverage and benefits, subnational governments can expand leave rights and provisions in several ways. For example, provisions under state law can include wagereplacement benefits or lower or no eligibility requirements. As of 2017, twenty-five states offered more accessible leave provisions for private and public sector employees than provided under the FMLA (Engeman, 2018). While regularly issued policy reports describe the current status of family policy across U.S. states (see Engeman, Petts, Gatenio Gabel, & Kaufman, 2019; National Partnership for Women and Families, 2016a, 2017a, 2018), less is known about the historical development of state leave laws.

This chapter examines state-level policy variation in the United States and addresses two main sets of questions concerning, first, policy variation across states and then, changes over time. First, what are the different types of statelevel leave provisions and how active have states been on leave policy? Second, how have state-level leave provisions developed over time? What are the main landmarks and trends?<sup>2</sup> My policy research draws from multiple primary sources, including state legislative documents, and secondary sources, such as academic publications on policy developments prior to the FMLA. Using these sources, I identify 72 leave laws adopted by U.S. states between 1942 and 2017. To discuss my state-level findings, I compare state leave legislation with leave policy development across affluent democratic countries.<sup>3</sup> With this comparison, I highlight some often-overlooked, innovative aspects of U.S. family policy and argue that state-level leave policies developed along a distinct trajectory.

<sup>2</sup>Anne Gauthier's (2004 [1996]) seminal work explores the same question for family policy development across affluent countries.

<sup>3</sup>By affluent democratic countries, I am referring to a common set of countries in cross-national comparative policy analysis, namely: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom, and the United States.

Social policies providing time-off from work to address medical needs and to care for new children or other family members have wide-ranging benefits—from improved public health (see Rossin, 2011; Ruhm, 2000; Widera, Chang, & Chen, 2010) to workforce stability (see Pavalko & Henderson, 2006; Rosenfeld, 2007) and family economic security (see Gould, Filion, & Green, 2011; Rossin-Slater, Ruhm, & Waldfogel, 2013). Studies in other country contexts show that leave policies, together with public childcare, help reconcile work and family obligations (Cukrowska-Torzewska, 2017; Ruppanner, 2013), reduce poverty (Maldonado & Nieuwenhuis, 2015), and influence women's labor force attachment (Korpi et al., 2013; Rønsen & Sundström, 2002) and lifetime earnings (Cukrowska-Torzewska, 2017). Yet, establishing parental and family leave provisions in the United States has been a decades-long project that continues to challenge leave-policy advocates. While Democrats have introduced paid family leave legislation in Congress four times since 2013,4 these proposals have yet to move from the early stages of the policy-making process.5 Given the lack of political will at the federal level and the societal importance of medical and caregiving leave, this chapter emphasizes the importance of understanding state-level policy developments and promising reforms.

The chapter proceeds as follows. The following section describes the development of the FMLA and highlights the compromises made in the policy process that left room for subsequent state-level expansions. After a brief description of methods, I identify different types of leave policies and show how a select few states are more active than others. The more active states have more generous provisions, suggesting efficacy for incremental approaches to policy change in some states. Then, I describe state policy development over time and show how leave provisions targeting female employees was abandoned early in favor of gender-neutral individual entitlements. I argue that state-level leave provisions can be distinguished by whether they address gender-neutral health or family caregiving needs. The emphasis on gender-neutral individual entitlements in U.S. leave policy, I argue, creates an opportunity for leave advocates to layer wage-replacement benefits on top of job-protected leave in a step toward gender-egalitarian family policy models found in other countries. I conclude by arguing that subnational variation

<sup>4</sup>Legislation was introduced as the Family and Medical Insurance Leave Act in the 113th and 114th Congress and the FAMILY Act in the 115th and 116th Congress.

<sup>5</sup>Specifically, the bills have been stuck in committee. Any one lawmaker can introduce legislation, making bill introduction a relatively easy achievement for policy advocates. Once a bill is introduced, however, it is referred to one or more committees, and committee chairs, who are appointed by the majority party, decide what bills receive hearings and are brought for a vote in committee. Committees must approve bills before they are sent back to the house (or legislative body) for a vote.

in social policy is a rich area for future research and cross-national comparative analysis of family policy outcomes should consider subnational variation when including federal systems.

# **The FMLA: Setting an Incremental Agenda for Leave Rights in the States**

Although provisions under the FMLA are limited compared to leave provisions in other countries, the FMLA was contentious legislation. A version was introduced in five congressional sessions starting in 1985 and vetoed twice before becoming law in 1993 (Elving, 1995). In the course of negotiating the FMLA, several compromises set the stage for state policy development. First, providing wage-replacement benefits was rejected as a policy idea before legislation was introduced (ibid.). Leave advocates knew that without wage-replacement benefits, access to leave rights would be limited to higherearning professionals who likely already had employer-provided paid leave, and unpaid leave provisions would do little to improve the lives of lowincome workers who needed leave rights most (Bernstein, 2001, pp. 44–45, 96). Yet, while drafting the original bill, lawmakers and a small group of feminist attorneys, union representatives, and other leave advocates agreed to pursue unpaid rather than paid leave (Elving, 1995). The decision to pursue unpaid benefits was based on a consideration of the bill's potential financial impact should benefits be provided (Sholar, 2016) and the partisanship of the executive and legislative branches (Elving, 1995). With a Republican president and a Republican majority in the Senate, proponents thought a paid leave bill would lack credibility (ibid.). Though an increasing number of states are adopting paid leave programs, establishing wage-replacement benefits remains a major hurdle for expanding leave rights.

Second, leave advocates rejected maternity leave early in the policy process in favor of gender-neutral family caregiving leave (Elving, 1995, p. 39). The question of whether to pursue maternity leave or gender-neutral (parental or family) leave conjured debates about whether women needed different or equal treatment and whether equal treatment—specifically, an equal lack of leave rights—resulted in equality (see Shiu & Wildman, 2009). Maternity leave would be easier to pass compared to gender-neutral parental or family leave, because it would be narrower in coverage and have broader appeal as a protective measure for vulnerable mothers (Bernstein, 2001). However, some leave advocates worried that leave rights tied to pregnancy and childbirth would further incentivize discrimination against women (Elving, 1995). Ultimately, advocates decided to emphasize equality over protection and opposed proposals to limit leave rights to women (Bernstein, 2001, pp. 44–45; Elving, 1995, p. 39). Instead, leave proponents broadened the original bill, which proposed parental leave, to include leave to care for spouses, parents, and older children (Elving, 1995, p. 66).6 As I will show in the following sections, inclusive family leave and gender-neutral individual entitlements became the normative form of state-level leave provisions after the FMLA.

Third, proponents agreed to limit the possible uses of leave and add eligibility requirements in an effort to gain Republican support (Elving, 1995). Early in the policy process, proponents agreed to interpret the bill's leave provisions to exclude leave for "non-recurring conditions for which treatment and recovery last no more than a few days" (ibid., p. 57). By excluding shortterm health needs, proponents sidestepped the contentious issue of granting leave for abortions, which would have drawn active opposition, potentially ruining the bill's chances of passing (ibid.). As a result, however, under federal law, American workers lack rights to paid sick leave to recover from minor illnesses. Paid sick leave, therefore, remains an issue that can be addressed under state policies.

Finally, eligibility requirements were introduced into the bill by a supportive Republican member of the House of Representatives. In an effort to gain additional support from fellow Republicans, Representative Margaret Roukema proposed shortening leave duration, adding "minimum service requirements" for leave eligibility, and limiting the law's coverage to establishments with 50 or more employees (Elving, 1995). Congress passed the bill with Roukema's amendments, but President George H.W. Bush vetoed it. Introducing the bill again, proponents agreed to increase the number of hours a worker would need to work before qualifying for leave—from 1000 hours to 1250 hours in the year prior to taking leave. President Bush vetoed the bill a second time, and proponents lacked the votes in the House to override the veto (ibid.). The 1992 elections resulted in Democrats winning the presidency and a majority of seats in both houses of Congress. Despite the new favorable political context, proponents decided to introduce a bill with the previously amended language to assure a swift victory (Bernstein, 2001; Elving, 1995). After sixteen days in office, President Bill Clinton signed the Family and Medical Leave Act (FMLA). Thus, the restrictive eligibility requirements remain, presenting a fourth way for states to expand leave rights.

<sup>6</sup>Including care for relatives broadened the bill's appeal and won active support from organized labor and the American Association of Retired People (AARP). Together, these organizations provided most of the funding for the coalition advocating leave legislation (Bernstein, 2001, p. 99).

Leave advocates planned to return to family policy later to reform provisions incrementally (Elving, 1995). Despite such aspirations, family policy has been a project for the states. In the context of congressional inaction, both before and after the FMLA, states adopted their own leave laws. When a version of the FMLA was first introduced in 1985, it inspired state legislative action on leave issues (Bernstein, 2001, p. 122). In 1987, 28 states introduced family and medical leave bills (Wisensale & Allison, 1989), and preceding the FMLA, eleven states adopted parental and family leave laws (see Engeman, 2018). All of these laws are still in effect, because they are more generous than the FMLA in at least one respect. While the FMLA applies across states, its limitations create opportunities for states to expand leave rights in several ways. Some states provide paid family leave (e.g., California, New Jersey, New York, Rhode Island, and Washington). Other states adopted unpaid, jobprotected leave laws that, compared to federal law, grant longer leave, cover smaller establishments, or include broader definitions of family for caregiving leave purposes. Given the institutional and political challenges of expanding social policy in the United States, the establishment of leave rights marks a notable accomplishment.

Few researchers have examined the historical development of state leave laws, particularly after the passage of the FMLA (for exceptions, see Engeman, 2016, 2018; Sholar, 2016; Williamson & Carnes, 2013). In this chapter, I contribute to previous research by tracing policy developments across states and over time. I show how the FMLA and subsequent state policies established gender-neutral individual entitlements and inclusive family (as opposed to parental) leave as normative elements of leave policy.

# **Methods**

For this chapter, I compiled and coded state leave laws covering both private and public sectors. To identify state leave laws and their enactment dates, I drew from multiple sources. Specifically, I used reports by the National Partnership for Women and Families (2015a, 2016a, 2017a, 2018), the National Conference of State Legislatures (2008), the Women's Bureau of the U.S. Department of Labor (1993, 2017), and a cross-institutional report by Workplace Flexibility 2010 at the Georgetown University Law Center and the Berkeley Center on Health, Economic & Family Security at the UC Berkeley School of Law (2010). I also consulted legal and academic publications to track the development of state laws prior to the FMLA. Using the legal references cited in secondary reports and publications, I searched LexisNexis Congressional, an archive of state laws and statutes, to find original language and enactment dates. When legal documents were unclear, I searched local media reports and legislative archives on state government websites.

States also adopted other types of family policies that were not included in my analysis. For example, some states have adopted leave policies covering state employees only. I did not include these laws in my analysis, because they better reflect labor relations between a state and its workforce than social policy, which is my main interest. I also exclude small necessities leave laws. Small necessities leave is used to address occasional, short-term needs, such as attending a child's school activity, grieving a family member killed in active military duty, or addressing issues related to domestic violence, sexual assault, and/or stalking. The diversity of small necessities leave laws makes them difficult to track comprehensively. I therefore leave this task for future research.

Family policies that are not leave policies are excluded from my analysis. Previous policy reports describe employment protections for pregnant employees (U.S. Department of Labor, 2015) or assess state-level "family friendliness" that consider, for example, space at work to express milk (NPWF, 2016a, 2017b). My analysis also distinguishes between leave provisions and anti-discrimination laws protecting pregnant employees. Antidiscrimination laws do not grant leave. Rather, they require employers treat pregnant employees the same as they would treat other temporarily disabled employees.7 I also distinguish parental and family leave laws from flexible use laws that allow flexible use of employer-provided paid sick leave. For example, California passed a "Kin Care" law in 1999 requiring employers who provide sick leave to allow use of sick leave to care for a sick parent, spouse, or child. Like anti-discrimination laws, flexible use laws do not grant leave rights. Rather, they govern how employees can use employer-based provisions.

After compiling a comprehensive list of state leave laws meeting my aforementioned specifications, I coded the laws by enactment year, state, and type of leave policy and included a detailed description of provisions. Using this coding method, I determined how many laws each state adopted and examined policy development over time (for a timeline of state leave legislation and comprehensive documentation of sources, see Engeman, 2018).

<sup>7</sup>A temporary disability is a non-work-related injury or illness that prevents work for a short period of time while the employee recovers.


**Table 17.1** Types of state leave laws in 2017

# **Family Policy Reforms Across States: Policy Types and Legislative Activity**

Between 1942 and 2017, U.S. states adopted 72 leave laws.<sup>8</sup> Most laws established or expanded unpaid, job-protected leave. I categorize provisions into seven types based on the purpose of the leave and whether wage-replacement benefits are provided (see Table 17.1). Of the unpaid provisions, types of leave include pregnancy and childbirth, parental, family, and family and medical leave.9 Short-term paid sick leave is provided by employers at 100% of earnings. To provide wage-replacement benefits during longer durations of leave,

<sup>8</sup>The leave laws counted here meet the criteria described under the methods section.

<sup>9</sup>Small necessities leave is an additional type of leave provision that is common across states. However, these laws are excluded from the table, because they are excluded from my analysis (see justification under methods section).

states have adopted temporary disability insurance and paid family leave insurance programs.

As of 2017, six states—California, Hawaii, New Jersey, New York, Rhode Island, and Washington—have temporary disability insurance or paid family leave insurance or both.10 As insurance programs, some do not provide job protection.<sup>11</sup> Rather, when an employee needs to take time-off from work to care for family or address a non-work-related injury, illness, or disability, including pregnancy- and childbirth-related disabilities or health issues, the employee can file a claim for benefits under the state insurance program. Rights to return to the same or similar position with the employer, when not provided under the insurance program, are covered under state and federal laws, though eligibility requirements may limit access. Employee payroll contributions fund most insurance programs, though employers and employees jointly fund the temporary disability programs in Hawaii, New Jersey, and New York and paid family leave insurance in Washington. Wage-replacement benefits across all programs are granted as individual entitlements. All paid family leave insurance programs pay benefits during leave to care for a parent, spouse, domestic partner, or child, but many laws cover care for additional family members. Benefits are paid between six and twelve weeks at between 50 and 90% of previous earnings to a cap, and New York's and Washington's benefit levels are calculated based on the state's average weekly wage.

Few states have approached family policy as an incremental project. Six states—California, Hawaii, Massachusetts, New Jersey, Rhode Island, and Washington—are responsible for adopting over half of the 72 leave laws enacted between 1942 and 2017. Figure 17.1 illustrates the unevenness in legislative activity across states. As shown in the figure, private sector employees in half of the states are covered only under the FMLA.<sup>12</sup>

Eleven states have passed only one law. Nine of these eleven states passed their one law within the five years that Congress debated the FMLA. These nine states—Illinois, Iowa, Kansas, Kentucky, Louisiana, Montana, New Hampshire, Ohio, and Wisconsin—passed laws prior to 1993 and have been inactive since. Arizona and Maryland each passed one law after the FMLA.

<sup>10</sup>Massachusetts adopted a paid family leave insurance program in 2018 that will begin paying benefits in 2021.

<sup>11</sup>Paid leave insurance programs in New York, Rhode Island, and Washington also provide job protection.

<sup>12</sup>Some state employees may be covered under other state leave laws. Of course, private sector employees may also have employer-provided leave rights, either voluntarily from employers or under collective bargaining agreements.

**Fig. 17.1** Level of legislative activity in the states: number of leave laws enacted, 1942–2017

Arizona adopted paid sick leave law with a ballot measure in 2016, and Maryland passed parental leave legislation in 2014 to cover smaller establishments than covered under the FMLA. Eight of the nine pre-FMLA laws in one-law states grant leave only to female employees.<sup>13</sup> In these states, family policy is thus stuck in a female-targeted, pre-FMLA policy framework.

The more active states generally have more generous provisions (see Fig. 17.2). Fourteen states have passed more than one law, and each of these states passed laws both before and after the FMLA. These states include: California, Colorado, Connecticut, Hawaii, Maine, Massachusetts, Minnesota, New Jersey, New York, Oregon, Rhode Island, Tennessee, Vermont, and Washington. After the FMLA, these incremental states passed additional laws, expanding FMLA provisions or coverage in some way. Five states—California, New Jersey, New York, Rhode Island, and Washington—adopted paid family leave insurance by 2017. Other laws expanded leave rights in smaller ways. For example, five states—California, Hawaii, Maine, Oregon, and Rhode Island—expanded job-protected family leave to cover care for additional family members and "reciprocal beneficiaries." In contrast, Tennessee and Colorado passed two laws each and expanded FMLA provisions in only small ways. Tennessee provides an extra month of leave with respect to the FMLA but covers establishments of 100 or more employees. Colorado's

<sup>13</sup>The exception is Wisconsin, which adopted family and medical leave in 1987.

**Fig. 17.2** The number of laws passed by the number of leave needs covered (*Note* Combining temporary disability and family leave insurance, this figure includes six categories of leave needs: pregnancy and childbirth, parental, family, [other] medical, paid sick leave, and longer-term paid leave insurance)

law broadens "family" for the purpose of family caregiving leave to include domestic or civil union partners. Although a select few states have achieved more generous provisions with fewer laws,<sup>14</sup> most incremental states improve leave access and affordability by adopting multiple, minor reforms.

# **Family Policy Reforms Over Time: Historical Landmarks and Trends**

In this section, I identify five cross-state trends in leave policy development and discuss them in the context of important landmarks. I argue that in the process of state-level policy development, leave advocates and lawmakers abandoned policies that restrict bonding/caregiving leave to mothers in favor of gender-neutral, individual entitlements to address either medical or family caregiving needs.

<sup>14</sup>For example, New York adopted three leave laws compared to California's eleven, but New York's paid family leave insurance is more generous than California's, offering job protection and longer leave duration (although fourteen years after California).

## **Temporary Disability Insurance, 1942–1969**

From 1942 to 1969, five states—California, Hawaii, New Jersey, New York, and Rhode Island—adopted temporary disability insurance (TDI) programs. TDI programs15 offer wage-replacement benefits during leave from work to address illness or non-work-related injury.<sup>16</sup> TDI programs originally excluded pregnancy and childbirth as qualifying disabilities for benefit claims. These programs are still in effect but now include pregnancy and childbirth as qualifying disabilities, as will be discussed below. As insurance programs, employees can apply for benefits when they miss work, but job protection or the guaranteed return to the same or similar position with the same employer were not included. Instead, job protection was provided by employers on a voluntary basis or under collective bargaining agreements or later, under state and federal laws. Employee payroll contributions fund TDI programs in California and Rhode Island, and employers and employees jointly fund TDI programs in Hawaii, New Jersey, and New York (for a comprehensive summary of state TDI programs, see Workplace Flexibility, 2010, pp. 109– 111). Because the early TDI programs did not cover leave for pregnancy or childbirth, they were not technically family policy. However, these programs laid the foundation for paid disability leave for pregnancy/childbirth and paid family leave insurance programs.

## **Pregnancy Disability Leave, 1961–1985**

Starting in the 1960s, state lawmakers began to consider the rights of pregnant employees. The first leave-related family policies were established by reforms to TDI programs. In 1961, New Jersey added pregnancy and childbirth as qualifying disabilities for receiving TDI benefits. Hawaii (in 1973), California (in 1976), and New York (in 1977) followed with the same reform. Finally, as a result of the Pregnancy Discrimination Act in 1978, Rhode Island's TDI program covered leave for pregnancy- and childbirth-related health issues. The expansion of TDI programs did not create new rights to job-protected leave but allowed female employees who had contributed to state insurance programs to draw benefits for pregnancy- and childbirthrelated disabilities. Thus, the inclusion of pregnancy- and childbirth-related disabilities under TDI programs was consistent with anti-discrimination

<sup>15</sup>TDI is referred to as state disability insurance (SDI) in California.

<sup>16</sup>Workers' compensation laws cover work-related injuries.

measures, because it resulted in equal treatment of temporarily disabled employees regardless of sex.

Lawmakers, feminists, and legal professionals debated whether equal provisions under the law resulted in gender-equal outcomes. At issue was whether job-protected leave for pregnant employees would further incentivize employer discrimination against women and whether the absence of job-protected leave for pregnant employees was discriminatory, resulting in unequal employment outcomes for men and women (Elving, 1995; Shiu & Wildman, 2009). The Civil Rights Act of 1964 made discrimination on the basis of race, color, religion, sex, or national origin unlawful. The inclusion of "sex" as a protected category left ambiguous whether disciplinary actions against pregnant employees were unlawful (Elving, 1995). In 1976, the Supreme Court ruled that discipline or dismissal of a pregnant employee did not violate the Civil Rights Act (Shiu & Wildman, 2009). In response, Congress passed the Pregnancy Discrimination Act of 1978, prohibiting discrimination based on pregnancy. Under this law, employers were required to treat pregnant employees the same as other temporarily disabled employees. Therefore, any employer-provided paid or unpaid, jobprotected leave for disabled employees must also be extended to pregnant employees. Alternatively, an employer could discipline or dismiss a pregnant employee for needing accommodation or time-off if the employer would take the same action against an employee with another type of temporary disability (ibid.).

In the midst of debates over "equal treatment" (e.g., anti-discrimination) versus "special treatment" (e.g., maternity leave rights), states began adopting unpaid, job-protected leave to address medical issues related to pregnancy and childbirth. Between 1972 and 1985, ten states adopted such laws.<sup>17</sup> Many of these early laws were unclear about duration and had low or no eligibility requirements. California's pregnancy disability leave policy provided four months of job-protected leave, and Massachusetts' maternity leave policy provided eight weeks. However, in the remaining eight states, leave duration was unspecified or stipulated as a "reasonable" period of time. Leave provisions in four states had no eligibility requirements, and the other state laws covered comparatively small establishments, starting at employers with 3–8

<sup>17</sup>California's and Montana's pregnancy disability leave laws, enacted in 1978 and 1975, respectively, were later challenged in court for potential violation of the Pregnancy Discrimination Act (Shiu & Wildman, 2009). Both state laws were upheld in court (ibid.), leaving the possibility to adopt job-protected leave rights for pregnancy and childbirth open to other states.

or more employees. Only Massachusetts' provision had a tenure-based eligibility requirement; to be eligible for job-protected maternity leave, employees needed to work at least three months for the employer.

Until 1985, state policies established job-protected leave only for pregnancy and/or childbirth, and these rights were restricted to addressing medical issues. Only two states offered exceptions during this period by providing leave to bond with new children. The Massachusetts Maternity Leave Act of 1972 provided eight weeks of unpaid maternity leave for nursing and to address health issues related to pregnancy and childbirth. Massachusetts then extended the law in 1984 to cover time-off to bond with newly adopted children but restricted this provision to women. Kentucky, in 1982, became the first state to adopt a gender-neutral leave policy, but rights were restricted to the purpose of bonding with a newly adopted child.

### **Gender-Neutral, Job-Protected Leave, 1985–1993**

In 1985, the first bill to propose job-protected caregiving leave was introduced in Congress. The Parental and Disability Leave Act, which eventually became the FMLA, proposed 18 weeks of parental leave and 26 weeks of disability leave and leave to care for sick children (Elving, 1995). Provisions were gender-neutral, unpaid, job-protected, and granted as individual entitlements. With the introduction of this bill, proponents found a way to bypass the equal treatment-versus-special treatment debate by creating leave rights without reference to sex or gender (ibid.).

States responded to policy developments in Congress with their own versions of gender-neutral leave. By 1993, two states provided job-protected parental leave to care for a newborn, newly adopted, or newly placed foster child. Two states provided family leave to care for a new child and a parent, spouse, or child with a serious illness or medical condition. Seven states provided family and medical leave to care for a family member or a new child or for employees to address their own serious illness or medical condition, including pregnancy- and childbirth-related health issues. Three of the four states that adopted parental or family leave policy but not medical leave had previously passed pregnancy disability leave laws; therefore, leave for pregnancy- and childbirth-related health issues were already covered. For the gender-neutral leave policies adopted in this period, provisions ranged in duration from six to twelve weeks.

As congressional lawmakers finalized the FMLA, state policies converged around leave provisions for an inclusive range of care needs. Policies adopted closer to 1993 were more likely to provide leave for temporary disabilities, infant/child care, and family care (i.e., family and medical leave). For example, Oregon and Rhode Island—early adopters of gender-neutral leave provisions—adopted parental leave first (both in 1987) and then expanded provisions (in 1989 and 1990, respectively) to include leave to care for non-child relatives—thus, establishing family leave.

Yet, some states in this period continued on the previous path of policy development by adopting leave provisions for pregnancy and childbirth only. Seven states—Illinois, Iowa, Louisiana, Tennessee, Ohio, Oregon, and Vermont—adopted such policies. Of these states, Oregon and Vermont also adopted gender-neutral leave provisions by 1993. Lawmakers in Tennessee, however, were so adamant about restricting leave rights to female employees they reformed the state's maternity leave law one year later to clarify that the leave was not for gender-neutral "bonding" but for pregnancy, childbirth, and nursing—activities tied to female bodies (Bernstein, 2001, p. 73). Between 1985 and 1993, leave rights in Massachusetts also remained restricted to female employees. Massachusetts expanded its maternity leave policy to include care for disabled children but did not extend leave rights to fathers, even as other states passed gender-neutral laws. In her case study of Massachusetts, Anya Bernstein (2001) shows how leave advocates were determined to establish wage-replacement benefits for family leave but failed. Unwilling to compromise on wage-replacement benefits, leave advocates also failed to adopt any unpaid, gender-neutral leave policy. Bernstein (2001) argues that in other states, like Iowa, Louisiana, and Tennessee, leave advocates adapted their policy goals to their respective state political contexts and won job-protected leave for mothers.

### **Implementation and Assessment, 1993–1998**

After the FMLA passed, the adoption of new state leave legislation slowed. During this time, states adopted only three leave policies. Oregon consolidated provisions from its separate pregnancy disability, parental, and family and medical leave policies. Hawaii broadened its definition of "family" for family caregiving purposes to include reciprocal beneficiaries, and Maine began covering smaller employers (i.e., with fifteen or more employees) under its family and medical leave policy.

## **Paid Leave and Other Expansions, 1999–2017**

In 2002, California became the first state to adopt paid family leave insurance. The state's paid family leave program was created by expanding its disability insurance program to cover leave to care for family members, including bonding with new children. Preceding the adoption of paid family leave insurance, California lawmakers in 1999 increased the level of disability insurance benefits. The state legislature also commissioned a fiscal impact study on the estimated costs of using state disability insurance funds to pay benefits during family caregiving leave. The study, released in 2000, showed that an extension of the state disability insurance program was feasible (Firestein, O'Leary, & Savitsky, 2011).

Following California's example, all of the states with TDI programs except Hawaii—extended their programs to cover family caregiving leave, including care for a newborn, newly adopted, or newly placed foster child. Specifically, these states included: New Jersey (in 2008), Rhode Island (in 2013), and New York (in 2016). Washington adopted a paid parental leave insurance program in 2007, just before the recession. Lacking a preexisting funding mechanism in the form of a TDI program, Washington's original program was never implemented. Ten years later and with new legislation, Washington became the first state without a TDI program to establish a paid family leave insurance program, with benefit payments starting in 2020.

States also adopted paid sick leave legislation. The legislation allows employees to accrue leave—paid by employers at 100% of the employee's previous earnings—to recover from short-term illnesses, like the flu. The first government to adopt earned sick leave was the City of San Francisco in 2006 (NPWF, 2015b), and in 2011, Connecticut became the first state to adopt the policy. Six states followed: California (in 2014), Massachusetts (in 2014), Oregon (in 2015), Arizona (2016), Washington (in 2016), and Vermont (in 2016). The laws—three of which passed by ballot measure—allow employees to accrue hours of leave based on the numbers of hours worked, up to a cap of either 24 or 40 hours and, except for Connecticut's law,18 cover small establishments.

<sup>18</sup>As the first state to adopt paid sick leave legislation, the coverage of Connecticut's paid sick leave law is comparatively limited. The law only covers employees in establishments of 50 or more employees working in "public contact occupations," for example, food service or public health. Proponents of earned sick leave advocated for policy change in the midst of the H1N1 outbreak (Greenhouse, 2009) and in motivating the need for paid sick leave, emphasized the public health risks of lacking access to paid sick leave, particularly in the service sector. Connecticut is the only state to restrict access to paid sick leave in these ways.

During this period, states also expanded unpaid, job-protected leave rights. Eleven laws amended leave rights by expanding leave duration, covering smaller establishments, allowing care for additional family members, or extending leave provisions in some other way. Most notably, Tennessee (in 2005) and Massachusetts (in 2015) extended leave rights to men under their existing state policies. The most frequent type of expansion was broadening the definition of family for the purpose of family caregiving leave. California in 2013 also broadened the definition of family for receiving benefits from the paid family leave insurance program. Under the program, eligible employees can now receive benefits while on leave to care for grandparents, grandchildren, siblings, and parents-in-law. Rhode Island, New York, and Washington—adopting paid family leave insurance programs after California's 2013 expansion—included a similar set of additional family members in their original laws.

## **State-Level Leave Policy Development in Cross-National Comparative Perspective**

U.S. state-level leave provisions differ from provisions in other countries in several ways. First, one of the most apparent differences is the prevalence of unpaid leave. While the United States is decades behind most other countries in the world with regard to paid leave, particularly for mothers (Heymann, Earle, & Hayes, 2007), the availability and need for job protection to address caregiving needs should not be minimized. Prior to the adoption of jobprotected leave laws in the 1980s, employers could terminate employees for missing work to give birth or missing work or needing accommodations for pregnancy-related health issues. With equal treatment under the law, employers could discipline or dismiss pregnant employees if they would take similar actions against other employees for work interruptions due to other temporary illnesses or injuries (Shiu & Wildman, 2009). Thus, researchers have found that even unpaid family leave improves women's labor force attachment (Pavalko & Henderson, 2006).

Second, unlike family policies in other countries, all leave rights and provisions in the United States are individual entitlements. In other countries, separate policies provide paid leave for mothers during pregnancy, for childbirth, and to care for a newborn (maternity leave), paid leave for fathers around the birth of the child (paternity leave), and paid leave for parents to care for a newborn (parental leave) (Koslowski, Blum, Dobroti´c, Macht, & Moss, 2019). To strengthen women's employment and encourage fathers' involvement in caring for new children, governments began—in the 1990s to create individual parental leave entitlements or convert portions of shared leave provisions into individual entitlements for fathers/partners (Brandth & Kvande, 2018; Daly, 2010; Daly & Ferragina, 2018; Karu & Tremblay, 2017). Whereas gender-neutral, individual entitlements are the norm for U.S. state-level leave laws, they are debated and under threat in other countries. Because individual parental leave entitlements incentivize fathers' leave-taking (Brandth & Kvande, 2018; Duvander & Johansson, 2012; Escot, Fernádez-Cornejo, & Poza, 2014; Haas & Rostgaard, 2011, also see Chapter 15 by Bartova & Keizer in this volume) or, in some cases, reduce shared parental leave duration, debates about individual entitlements center on whether governments should play a role in how people divide labor within their families (Borchorst, 2006; Chronholm, 2009; Ellingsæter, 2012; Eydal et al., 2015). While individual entitlements are not debated in the United States, unpaid leave does little to promote gender egalitarianism as men are unlikely to take unpaid leave or leave paid at low levels (see Kaufman, 2018; Moss, 2008). However, should the United States eventually adopt wage-replacement benefits—as have some states—the prevailing norm of gender-neutral, individual entitlements may help to align U.S. federal family policy more closely with gender-egalitarian family policy models.

Third, leave provisions in the United States are distinguished by medical and caregiving needs, and many leave laws provide job-protected time-off to address either or both needs. In contrast, provisions in other countries can be distinguished by who provides care (i.e., mother, father/partner, relative) and who receives care (i.e., self, new child, older child, or relative) (see Blum, Koslowski, Macht, & Moss, 2018; Daly & Ferragina, 2018). For example, in other countries, maternity leave provides paid time-off for female employees during the late stages of pregnancy, for childbirth, and to care for/bond with a newborn, thus combining medical and caregiving needs and emphasizing ties between female bodies and caregiving (see Blum et al., 2018). In the United States, paternity leave, or time-off for fathers surrounding the birth of a child, does not exist as a distinct provision. Instead, fathers, when eligible, have rights under gender-neutral parental or—more commonly—family leave laws, and the duration of leave for fathers (and other non-birth parents) is the same as for (birth-)mothers.19 These laws, therefore, frame leave to care

<sup>19</sup>All current family leave laws cover care for adopted children and children of same-sex couples with legal guardianship. Coverage of same-sex spouses became universal across states when the U.S. Supreme Court overturned state-level same-sex marriage bans in 2015, though some states recognized same-sex marriage or extended leave rights to same-sex partners before the ruling. Some state laws additionally cover children of employees standing "in loco parentis" (i.e., individuals caring for a

for and bond with a new child as meeting non-gendered needs.<sup>20</sup> By distinguishing between medical and caregiving needs and defining pregnancy and childbirth as temporary disabilities among many types of disabilities, leave law in the United States emphasizes the need for leave as something any worker may experience.

Finally, U.S. leave laws often merge provisions to care for new children (parental leave) with provisions to care for older children or other family members (family or carers leave). Given that some states now provide wagereplacement benefits during family leave, some American workers have access to family/carers leave provisions that surpass other countries in terms of duration and benefits. For example, New York's paid family leave insurance program, which will be fully implemented in 2021, is among the most generous state insurance programs. The program covers up to 12 weeks of family caregiving leave paid at 67% of employees' earnings and capped at 67% of the statewide average weekly wage (see Engeman et al., 2019). Countries with annual provisions of paid family (or carers) leave offer two weeks of benefit payments at most (see OECD, 2014).21 Even unpaid family leave under the FMLA and state laws exceed carers' leave provisions in many other affluent countries. For example, carers' leave in Canada, Greece, and Portugal is unpaid and shorter in duration compared to the FMLA (see OECD, 2014).

# **Has the Historical Moment for Paid Leave Passed?**

Countries and states that provide wage-replacement benefits during parental leave often started with a focus on mothers and an acknowledged role for the state in administering paid benefits (see Gauthier, 2004 [1996]). By the time Congress introduced the first version of the FMLA, paid maternity leave had been available in other countries for decades (ibid.). Some of these countries, starting in the 1990s, extended individual entitlements to paid leave

child regardless of legal guardianship), for example, grandparents with primary care responsibilities for grandchildren.

<sup>20</sup>Out of 72 enacted leave laws, only two—in Massachusetts and Vermont—restricted bonding leave to mothers. However, both states later reformed their laws to extend leave rights to fathers.

<sup>21</sup>Some countries offer more generous benefits than state-level programs in the United States Italy offers a two-year lifetime provision of family caregiving leave paid at 100% of previous earnings, and other countries offer "per episode" provisions of well-compensated leave (i.e., 60% or more of previous earnings), ranging from two days to one year (see OECD, 2014). Funding of eldercare services also varies across countries, which may inform the need for private, family care (Brandt, Haberkern, & Szydlik, 2009) and therefore, support for family caregiving through paid leave entitlements or cash-for-care schemes (Bouget, Saraceno, & Spasova, 2017).

to fathers (see OECD, 2014)—around the time the United States adopted unpaid family and medical leave. Additionally, most states that now offer paid family leave insurance likewise started with a focus on mothers, providing paid leave benefits for pregnancy and childbirth through temporary disability insurance programs established prior to 1970.

By the time family policy entered the U.S. national legislative agenda in the 1980s, a focus on mothers was no longer acceptable (see Elving, 1995; Shiu & Wildman, 2009). By acting late on family policy, American lawmakers and leave proponents were restricted to favoring unpaid (gender-neutral) individual entitlements over paid (gendered) maternity leave, possibly limiting opportunities to adopt paid leave in subsequent decades. While other social policies and workplace protections in the United States have started with an appeal to women's special status as child-bearers and family caregivers (Skocpol, 1992), contemporary leave proponents are now in a challenging position to win new rights for workers without starting with a narrow—and historically compelling—appeal to protect mothers.

Yet, recent changes at the state level offer hope for paid parental and family leave expansions. In 2017, Washington established a paid family leave insurance program, making it the first state to do so without a preexisting funding mechanism (e.g., a state disability insurance program). Prior to Washington State, Washington, DC also passed a paid family leave law. Massachusetts, in 2018, adopted 12 weeks of paid parental and family leave and 20 weeks of paid medical leave covering pregnancy and childbirth. Employees can combine the leave for a total of 36 weeks. Lastly, in 2019, Oregon became the most recent state to adopt paid family leave insurance. Its program includes a 100% wage-replacement rate for its lowest earners (i.e., people earning 65% or less of the statewide average weekly wage) (NPWF, 2019).

How were these subnational governments able to do something others have not? Previous research suggests a combination of social movement pressure and allies in government motivate family policy reforms in U.S. states. Qualitative studies suggest that women's organizations (Bernstein, 2001; Sholar, 2016) and organized labor play leading roles in facilitating leave policy adoption (Milkman & Appelbaum, 2013), and quantitative analysis shows that states with stronger trade union institutions are faster to adopt gender-neutral leave rights (Engeman, 2020). Democratic lawmakers and majorities in legislative offices are also important (Engeman, 2020; Milkman & Appelbaum, 2013; Sholar, 2016). Drawing from the research presented in this chapter, Democratic majorities in the state legislature matter for the expansion of leave rights and provisions. Of the 26 state leave laws that passed since the FMLA, 22 (85%) passed when Democrats held a majority of seats in both state houses, and Democratic governors signed 23 (88%) of them.

Leave advocates can also expand leave rights and provisions by circumventing state legislatures. Three states created paid sick leave rights by using ballot initiatives that propose the policy directly to voters. Using ballot measures to establish state-level leave rights is a relatively recent tactic, with Massachusetts in 2014 as the first state to successfully adopt paid sick leave in this way. Counties and municipalities can also establish leave rights. As of 2016, twenty-nine municipalities had paid sick leave ordinances that cover workers within city limits (NPWF, 2016b). In another recent development, San Francisco and Washington, DC require employers to supplement benefits paid by the state/district family leave insurance programs (see Kaufman, 2019).

Ballot initiatives can also motivate reforms via traditional legislative channels. Massachusetts and Washington established paid family leave insurance programs following efforts to pass ballot initiatives. In Massachusetts, the ballot initiative would have provided up to 16 weeks of family leave paid through an employer-funded, state-administered program (Salsberg, 2017). In Washington, voters passed a ballot initiative backed by organized labor that increased the state minimum wage and required employers provide up to 40 hours of paid sick leave per year (Jenkins, 2017). In both states, some business leaders and other usual opponents to leave legislation supported the state bills, because they thought a ballot initiative would pass and they would have more control over the content of legislation using the state legislative process (see Jenkins, 2017; Salsberg, 2018). With ballot initiatives, voters decide whether to pass legislation as written by proponents. With the state legislative process, however, opponents can propose amendments, for example, introducing restrictive eligibility requirements, shortening duration, or lowering benefit levels. Thus, introducing or threatening to introduce ballot initiatives may reduce opposition to reforms and offer another tactic for facilitating state action.

Establishing paid leave provisions or other improvements to leave access is unlikely at the federal level in the near future. While most Americans want access to paid leave, they agree less about how this should happen (Horowitz, Parker, Graf, & Livingston, 2017). In order to establish paid leave provisions at the federal level, research on state policy development suggests a need for Democrats to control the legislative branch of government (Engeman, 2020). Even with this control, however, Congress would need to prioritize paid family and medical leave above other—arguably more pressing—social issues, such as improving the healthcare system, funding social security, or addressing climate change, among others. Until that time, states with leavefavorable political conditions will likely be the site of family policy innovation in the United States.

# **Conclusions**

Examining cross-state variation in U.S. leave policy, this chapter discusses 72 leave laws adopted by states between 1942 and 2017. This chapter thus makes several contributions. First, I identify seven main types of state leave policy and argue that these policies can be distinguished by whether the leave addresses either medical or caregiving needs, which is in contrast to leave policies in other countries that can be distinguished by who provides and who receives care. Thus, unlike leave provisions in other countries, an employee's sex or gender does not determine leave entitlements under most state or federal laws (or any laws adopted after 1993). Accordingly, leave for pregnancy and childbirth fits within a gender-neutral need to address a temporary disability, and leave to bond with a child is a separate, genderneutral, individual entitlement. Second, an increasing number of states are creating paid leave insurance programs, and the more active states are more generous, suggesting efficacy for incremental reforms in some states. Third, for leave to care for non-child relatives (i.e., family or carers leave), I argue that provisions in some states approach or exceed provisions in other affluent countries where carers' leave is a more recent development (see Bouget et al., 2017). Finally, I point to the prevalence of gender-neutral individual entitlements to job-protected leave to suggest potential for U.S. family policy to catch up to family policies in other affluent countries. Leave proponents can and have layered wage-replacement benefits on top of already-existing, gender-neutral individual entitlements. As states take this layering approach to create family leave insurance programs, they establish provisions that are more comparable to gender-egalitarian family models in other countries.

Cross-national comparative analysis of family policy should be attentive to subnational policy developments in the United States for at least three reasons. First, cross-national comparative studies that include the United States in analysis must be careful in treating the United States as a country that lacks leave rights entirely. The six states that had paid family leave insurance programs as of 2017 are—aside from Rhode Island—among the nation's most populous. In 2018, workers in these states constituted over one-fourth (26%) of the U.S. workforce (U.S. Bureau of Labor Statistics, 2018). While some workers in these states may not meet eligibility requirements for leave, a sizeable and growing proportion of the workforce has access to paid leave provisions under state laws with others possibly covered by employers or collective bargaining agreements. Future cross-national comparative analysis could exclude data from states with paid leave insurance or examine these states separately.

Second, subnational family policy variation offers a rich area for future research. In the following Chapter 18 (in this volume), Zachary Parolin and Rosa Daiger von Gleichen examine the implications of a broad set of family policies for poverty rates, household expenditures, and gender equality in the labor market. Though they find little evidence that state-level policy variation results in a variation in outcomes, recent increases in leave benefit levels under existing state programs and the introduction of new insurance programs in more states, warrant a return to this question of subnational policy variation and outcomes in the future. Additionally, cross-national studies of parental and family leave policy outcomes could potentially include some states in analysis. For example, California, the world's fifth largest economy (Associated Press, 2018), would offer an interesting comparative case, given that its paid leave program has been in effect since 2004. Future research could also examine relationships between family/carers leave provisions and women's labor force participation and earnings at later career stages.

Finally, state-level legislative action may eventually influence federal policy. Researchers have noted the potential for policy diffusion across states and different levels of government (e.g., see Chapter 4 by White in this volume). The FAMILY Act, currently introduced in the House (HR 1185), would establish an insurance program to pay benefits during medical, parental, and family caregiving leave, similar to legislation adopted by states. Benefits would be paid as gender-neutral, individual entitlements. As with other social policies in the United States, states can serve as laboratories where new policy ideas are tested before they spread (see Volden, 2006), and states are the likely sites for leave policy innovation as well.

**Acknowledgements** Preparation of this manuscript was supported by the European Commission, Marie Skłodowska-Curie Actions (742339-ERA).

## **References**

Associated Press. (2018, May 4). California is now the world's fifth-largest economy, surpassing United Kingdom. *Los Angeles Times*.


*International review of leave policies and research 2019*. Available at: http://www. leavenetwork.org/lp\_and\_r\_reports/.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **18**

# **Family Policy in the United States: State-Level Variation in Policy and Poverty Outcomes from 1980 to 2015**

**Zachary Parolin and Rosa Daiger von Gleichen**

The 50 United States have long featured vast differences in their levels of support for low-income families. As far back as 1939, the generosity of social assistance transfers as part of the Aid to Dependent Children program ranged from an average monthly allowance of \$2.50 for a mother in Arkansas to \$24.50 for a mother in New York (Cauthen & Amenta, 1996). Recent evidence suggests that state-level family policy divergence has continued, and perhaps intensified, in more recent decades (Bruch, Meyers, & Gornick, 2016). In the mid-1990s, the federal government granted increasing administrative authority to the states with respect to cash assistance for vulnerable families (Page & Larner, 1997). In the two decades following, many states would increasingly exert control over social and labor market policies, ratifying state-level increases in the statutory minimum wage, supplements to federally administered refundable tax credits, work-preparation programs, childcare subsidies, and even paid sick and family leave policies.

This chapter investigates the diversity of states' family policy packages from two perspectives. First, we detail the variance of states' policies as they relate to three key dimensions of family policy packages: money, services, and time. For each dimension, we document the extent of state-level variance in recent

Z. Parolin (B)

Columbia University, New York, NY, USA e-mail: Zachary.parolin@columbia.edu

R. Daiger von Gleichen

Department of Social Policy and Intervention, University of Oxford, Oxford, UK e-mail: rosa.vongleichen@stx.ox.ac.uk

<sup>©</sup> The Author(s) 2020

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_18

years and, when possible, highlight convergence or divergence in the family policy indicators over time. If states are increasingly diversifying with respect to family policy packages, we might similarly expect that social outcomes for families with children are also diverging. This is the second focus of our chapter. Specifically, we provide descriptive evidence on trends in state-level variance in six social outcomes: (1) child poverty rates after accounting for taxes and transfers, (2) child poverty rates while only accounting for market income, (3) the male-to-female employment ratio, (4) the male-to-female earnings ratio, (5) out-of-pocket medical expenditures among households with children, and (6) expenditures on childcare among households with children. While this is of course not an exhaustive list of family-related outcomes, it should nonetheless provide an idea of whether states are moving farther apart not only in terms of family policy packages, but also in terms of family economic outcomes.

Our primary findings confirm that family policy packages vary widely across the 50 states, and that this is increasingly true with respect to the coverage of cash assistance for low-income families, statutory minimum wage levels, and healthcare coverage. However, this divergence in policies has not always translated into divergence in social outcomes. We find that state-level variation in post-tax/transfer child poverty rates, for example, has remained stable over time despite growing variance in many of states' family-oriented policies. On the other hand, we find clear evidence that some states' decisions to expand Medicaid coverage has contributed to state-level variation in outof-pocket medical spending. We conclude this chapter with a discussion on why state-level family policy variation often does not translate into variation in social outcomes.

# **Background**

## **Family Policy Across the 50 United States**

Family policy encompasses a vast scope of policies, which some scholars have argued should include all measures that affect "the family" (Kamerman & Kahn, 1979; Kaufmann, 1993). Kamerman and Kahn distinguish between "explicit" and "implicit" family policies. Whether or not they aim at a specified goal, explicit family policies "deliberately do things to and for the family" whereas implicit family policies are not "specifically or primarily addressed to the family, but […] have indirect consequences" for families. Given these (in)direct effects, Kamerman and Kahn (1981) highlight the necessity to examine the family policy "package", i.e., the combination of policies that create the policy context for families. Their framework to capture a holistic family policy package distinguishes between family policies providing (1) money, (2) services, and/or (3) time (Kamerman & Kahn, 1994). Money ostensibly provides the choice as to whether family members with care obligations can provide care themselves or purchase care-services instead, although the practical reality of this depends greatly on the generosity level of the monetary benefit. Services provide care-enabling family members to continue active employment. Time policies give family members the option of performing care themselves.

Compared to other advanced democracies, the United States invests very little in "money, time, and services" for families (Daiger von Gleichen & Seeleib-Kaiser, 2018). As O'Connor, Orloff, and Shaver (1999, p. 223) highlight, the state "is clearly subordinate to the market and the family" in the United States when it comes to ensuring the financial security of families with children. Indeed, the primacy of the market is easily identified across all three dimensions of money, services, and time. However, the extent to which this is true varies across the 50 states. As noted in the Introduction, of this Volume, recent studies suggest that states have increasingly diverged with respect to the provision of family-oriented policies. Before getting into detail of trends in state-level variation, we first highlight the core elements of the family policy package across the 50 states. In Table 18.1, we break down the primary family-oriented policies into their money, services, and time dimensions. We focus primarily on specific policies or programs over which state governments, rather than the federal government, have some authority.

Importantly, this is not an exhaustive list of policies that might affect families with children. The United States features many in-kind or near-cash benefits, for example, that are important to the wellbeing of families, but that tend to be funded on the federal level. The Supplemental Nutrition Assistance Program (SNAP, or "food stamps") is a good example: though technically administered by the states, its benefit calculation procedures are set at the federal level, and all benefit allocations are funded by the federal government. Moreover, the benefits are not exclusively available to families with children. As such, we do not feature SNAP prominently in our discussion of state-level family policy variation.

Table 18.1 does include medical services, which in many countries do not qualify as explicit family policies. Here the United States is an exception. The two healthcare services are explicitly geared toward the family, as the rhetoric that circumscribes them is of aiding families in meeting the substantial healthcare costs in the United States (Centers for Medicare and Medicaid Services, 2018).

#### **462 Z. Parolin and R. Daiger von Gleichen**


**Table 18.1** Family policy package in the United States

### **Money, Services, and Time Across the 50 United States**

With the conceptual structure of the family policy package in place, we now provide an overview of the policies relating to "money, services, and time" across the 50 United States. Specifically, we provide details relating to eight programs or services that state governments in the United States have some authority to control and that relate to the financial security of families with children. Later, we will demonstrate trends in state-level variation in the eight programs and services to detail how states have diverged (or converged) over time with respect to each dimension.

*Money*: Temporary Assistance for Needy Families (TANF), the Earned Income Tax Credit (EITC), and statutory minimum wages form the core of the "money" dimension of the American family policy package.

TANF is the country's primary cash-based social assistance program and is primarily targeted at single-mother families. It was introduced in the mid-1990s as a replacement for the less restrictive Aid to Families with Dependent Children (AFDC) program. The legislation that introduced TANF transformed three core components of state-administered social assistance. First, it strengthened the conditionality requirements attached to the receipt of cash assistance. Under AFDC, families under a certain income threshold were entitled to cash support. With the introduction of TANF, however, that entitlement was ended and recipients would be required to engage in "work participation activities" or employment to continue receiving cash support beyond a certain duration (Falk, 2014). Second, the legislation expanded the scope of the program's objectives. While AFDC was primarily a cash assistance program, states can allocate TANF funds toward any of the program's four statutory purposes: "to provide assistance to needy families," "to end the dependence of needy parents on government benefits by promoting job preparation, work, and marriage," "to prevent and reduce the incidence of out-of-wedlock pregnancies," and "to encourage the formation and maintenance of two-parent families" (Parolin, 2019). Third, TANF introduced a "block grant" funding scheme, effectively providing states with a fixed sum of money to manage their own TANF programs. Today, all states spend less on cash assistance under TANF than they did in the late 1990s. Ten states, in fact, spend less than 10% of their TANF block grants on cash assistance, instead allocating the funds toward efforts to encourage work, promote two-parent families, or other miscellaneous expenditures.

The federal EITC is effectively a subsidy for low-wage earners. Though it does not exclusively target families with children, the value of the wage subsidy increases with the number of children in the household. Though the federal EITC was introduced in 1975, it was strengthened considerably throughout the 1990s. Beginning with Wisconsin in the mid-1980s, some states have introduced their own supplements to the federal EITC. By 2015, 23 states offered state-level supplements to the federal EITC, compared to just eight states in 1994 (Internal Revenue Service, 2015).

Finally, statutory minimum wages simply act as the wage floor for employed workers. The federal government sets a minimum wage level that applies across all states, but state governments have increasingly acted to raise the wage floor beyond the federal minimum. In 2014, 22 states offered statutory minimum wage levels higher than the federal minimum, compared to only five such states in 1994 (U.S. Department of Labor, 2015). Though not an explicit family-oriented policy, the level of the minimum wage certainly has an effect on the financial security of low-income families. Different from SNAP, minimum wage levels are set by state governments and vary considerably across the 50 states. Thus, we include minimum wages in our analysis of state-level variation over time.

*Services*: Early childhood education, childcare, and healthcare programs are the core of our analysis of family-oriented services. Pre-Kindergarten ("Pre-K") programs provide care and education support for 3- and 4-year olds. These are mostly supported by state funds, with the explicit intent to support working families with young children (National Institute for Early Education Research, 2015). Greater investment into pre-K increases the odds that a parent with young children can, if desired, secure childcare support and enter the labor market. Pre-K programs exist in addition to (and usually completely separately from) the federal, targeted Head-Start programs that are nearly entirely federally funded. Forty-one states had such pre-K programs in 2014.

The specific health insurance policies examined are Medicaid and the (State) Children's Health Insurance Program (commonly abbreviated to (S)CHIP). Unlike the general federal health insurance Medicare which is funded and administered federally, Medicaid and (S)CHIP are administered by the states under federal guidelines and funded by states with "matching" federal funds of at least 50%. While both Medicaid and (S)CHIP are meanstested, the difference is that (S)CHIP is intended to insure those children whose parents' income exceeds the maximum eligibility for Medicaid but who are still unable to afford private healthcare insurance for their children (Finegold, 2005; Meyers, Gornick, & Peck, 2001). In this sense, both Medicaid and (S)CHIP are addressed to the entire family, aiding its members to meet the financial burden of providing healthcare for children. After passage of the Patient Protection and Affordable Care Act in 2010, states were granted the option to expand Medicaid eligibility to individuals with annual incomes below 138% of the federal poverty level (rather than at a lower income threshold). States expanding Medicaid receive additional financial support from the federal government for doing so. A state's prioritization of Medicaid and CHIP protects low-income families against income shocks in the event of health concerns, and allows families to receive medical treatment without relying on an employer or family member for health insurance coverage.

*Time*: The United States is an outlier among advanced democracies in not providing a paid parental leave program at the federal level. Instead, the country features the Family and Medical Leave Act (FMLA), which mandates 12 weeks of unpaid leave for family or medical purposes. In Chapter 17 in this volume, Cassandra Engeman explores state-level unpaid leave policies that expand on the framework set by the FMLA. In recent years, several states have further introduced their own paid leave programs. Engeman's chapter describes these and temporary disability insurance programs that can provide wage replacement benefits during leave in the previous chapter. Here, we focus on paid leave programs and discuss them in greater detail later in this chapter during our analysis of cross-state divergence.

#### **Divergence in Policies, Divergence in Outcomes?**

The mid-1990s marked a turning point in the governance of family-oriented policies in the United States The introduction of TANF was at the core of a broader movement to shift authority over social programs from the federal to the state government. This era of decentralization premised that each state, rather than the federal government, ought to be able to more effectively serve the interests of its constituents and thus, should have greater discretion over the allocation of public resources (Meyers et al., 2001, p. 459; Obinger, Leibfied, & Castles, 2005). Some scholars referred to this era as a "devolution revolution" (Nathan & Gais, 2001), or a turn from centralized social programs to more devolved, state-led programs. One potential benefit of greater devolution was that state diversity in programs would create natural experiments, building an evidence base of which policy changes to a given program might be most beneficial for improving families' economic security.

In the two decades following the onset of this "devolution revolution", several comparative analyses have sought to understand the extent of statelevel variation in social policies. Studies have demonstrated, for example, that state governments vary widely with respect to labor relations (Hays, 1954), healthcare (Collins, Felderhoff, Kim, Mengo, & Pillai, 2014) and education policy (McLendon & Perna, 2014). With respect to family-oriented policies, Meyers et al. (2001) similarly found that the 50 states feature much diversity in the generosity of their policy packages for families with young children.

Other studies have sought to understand whether states have become *more* diverse over time with respect to social and family policies. Bruch et al. (2016) find the diversity of states' social policies has generally widened from 1994 to 2014, in line with claims of a "devolution revolution" after welfare reform. Conducting a latent-variable analysis of 148 state-level policies across eight decades, for example, Caughey and Warshaw (2016) find that variation in states' "policy liberalism"—a measure of leftward ideological orientation—has steadily increased from 1936 to 2014.

Despite evidence of state-level policy divergence, however, less is known about whether state-level social outcomes for families with children have similarly diverged. Given that state-level policies affect the economic wellbeing of households within the state, and that family-oriented policies have appeared to diverge at the state level, then we might similarly expect to observe a divergence in social outcomes among families living in different states. In this chapter, we look at state-level variance in six indicators of family economic wellbeing that family policies typically aim to address. These include indicators of child poverty (both before and after transfer are included), male–female gaps in employment and earnings, and average family spending on medical and childcare expenses. We do not attempt to isolate the effect of individual policies on the outcomes. First, we expect that each of the family policies affect our outcomes of interest in some way. For example, greater provision of pre-K should have consequences for each of the indicators: insofar as pre-K allows a mother to enter the labor force, it can have an effect on the male-to-female employment and earnings gap, and may similarly affect levels of child poverty. If a working mother can more easily access healthcare coverage, then out-of-pocket medical spending may decline, as well. Second, our interest in this chapter is documenting descriptive trends rather than identifying the causal effects of state-level policies. Thus, we will observe whether state-level trends in the variation of family policies align with trends in the variation of family-related social outcomes.

# **Data and Methods**

Our primary analysis is concerned with measuring the relative variation of states' family policy inputs and social outcomes over time. We measure convergence or divergence using the coefficient of variation (CV), a simple measure of dispersion commonly applied in descriptive studies of policy convergence. The CV is calculated as the standard deviation of a set of indicators relative to its mean in a given year. For example, to measure the CV of states' maximum TANF benefits in a given year, we simply divide the standard deviation of states' maximum TANF levels relative to the unweighted mean of states' maximum TANF levels. When the CV declines for a given indicator over time, this indicates that states are becoming more similar with respect to the indicator (less variance), and a rising CV over time indicates divergence in the set of indicators (more variance). In sensitivity checks, we also estimate convergence measuring the 90th percentile of a state's value for a set of indicators relative to the 10th percentile, but we find no substantive difference in the trends of the two measures. For simplicity, then, we only show changes over time in the CV.

In Table 18.2, we present details on how we construct each of our policy indicators and social outcome variables. For each of the three indicators in the benefits and wages dimension, we have data that span many years (1980– 2015) and thus can produce a reliable estimate of convergence or divergence over time. For other indicators, such as state spending on pre-Kindergarten


**Table 18.2** Measurement of family policy indicators and social outcomes

*Note* Data sources are listed in appendix Table 18.8

programs, we only have cross-sectional data in a recent year. For such indicators, we describe differences in states' levels of investment, but cannot say much about changes in variance over time.

For our five policies of interest, we include indicators of both generosity and coverage. Generosity reflects the level of the given benefit in real terms or, in the case of the two services, state spending on the given service relative to the number of children in the states. Coverage, meanwhile, represents the number of potentially eligible individuals who actually utilize the given benefit or service. For TANF, for example, we measure coverage as the share of families in poverty in the given state and year who actually receive TANF benefits. Lower coverage thus not necessarily indicates less demand for the given policy or service; instead, it may indicate that the benefit or service is simply less accessible for families in the given state and year. With respect to TANF, for example, many states have imposed strict work requirement, drug tests, and short lifetime time limits that inhibit many low-income families from accessing the benefits. In such states, coverage of TANF will be lower.

Data on our six social indicators come from the U.S. Current Population Survey (CPS ASEC). For our two measures of child poverty (pre-tax/transfer and post-tax/transfer), we follow the common practice in international poverty literature in setting the poverty threshold at 50% of the national household median income.<sup>1</sup> Our post-tax/transfer measure also follows the common practice of including all taxes and transfers (even near-cash benefits such as "food stamps") into the income definition.

We measure the male-to-female employment ratio as the average employment rate of men in a given state-year relative to the average employment rate of women in the state-year. With respect to the male-to-female earnings ratio, we measure the median annual market earnings of employed men relative to the median earnings of employed women. Finally, we measure outof-pocket medical expenditures and work and childcare expenses as the mean level of expenditure among households with children in a given state-year. Before measuring the mean, we top-code levels of expenditure at the 99th percentile in each year.

## **Diversity in State-Level Family Policy Packages**

We first present results on the diversity and divergence of the money, services, and time dimensions of states' family policy packages. We then present the descriptive trends in our outcomes of family economic wellbeing.

*Money*—Table 18.3 shows the cross-state diversity in the generosity and coverage of TANF, state EITC supplements, and statutory minimum wages in 2015. The bottom rows of Table 18.3 also show the change in average values of the indicators from 1980 to 2015.

As described in the prior section, the coefficient of variation, presented in the middle row of Table 18.3, describes the relative disparity of states' family policy indicators for 2015. With respect to the generosity of TANF, EITC, and the minimum wage, variation is clearly greatest in the state EITCs (CV of 1.38). This is primarily because nearly half the states offer no state-level supplement to the EITC. Variation is smaller for the maximum TANF benefit

<sup>1</sup>An alternative approach would have been to set the poverty threshold at 60% of national household median income, as is standard in the European Union and much of the social policy literature. In this approach, levels of poverty increase slightly, but trends in variation across states are not meaningfully affected.


**Table 18.3** State-level diversity in generosity and coverage of 'money' dimension of family policy in 2015

*Note* Unweighted state-level statistics. Generosity of EITC refers to share of federal EITC that state government matches. Coverage of TANF refers to share of households in poverty receiving TANF benefits in the state in 2015. Coverage of EITC refers to share of eligible households collecting federal or state EITC benefits in the year. Mean dollar values in 1980 presented in real 2015 USD using PCE deflator

values (CV of .384), but the gap between the 10th and 90th percentiles remains notable: a difference in \$380 per month in benefit levels. From 1980 to 2015, the mean level of maximum TANF benefits declined more than \$300 in real terms. The minimum wage shows the smallest variance of the three indicators (CV of .099), primarily because the federal minimum wage floor ensures that nearly half the states have the same minimum wage of \$7.25 per hour. From 1980 to 2015, there was virtually no change in the mean value of states' wage floors (in real terms).

In the right half of the table, coverage rates for TANF and the EITC are presented. Recall that the EITC coverage rates refer to participation in either the federal or state EITC in 2015 (among eligible households). Thus, even states with no state-level EITC supplement will have positive values of EITC participation. Here, we see that variation in coverage of TANF (CV of .662) far exceeds that of the EITC (CV of .034%). Note that variation in TANF coverage is nearly double the variation observed in maximum benefit levels in the left half of Table 18.3.

We now turn toward an assessment of variation over time for the generosity of AFDC/TANF, state supplements to the EITC, and the minimum wage. In Fig. 18.1, we display trends in the CV. State EITC supplements are unique among the three in seeing strong convergence over time (the downward sloping gray line in Fig. 18.1). Recall that in the mid-1980s, only one state (Wisconsin) featured a supplement to the federal EITC. Today, however, nearly half the states offer a bonus to low-income workers collecting

**Fig. 18.1** Change in variation of state-level wage and benefit policies

the federal EITC. The state additions to the federal EITC vary: in New Jersey, low-wage workers can claim an extra 35% of the total value of their federal EITC. Like most states offering an EITC supplement, New Jersey offers a refundable credit, meaning that if a household's tax credits exceed their tax liabilities, they can collect the negative balance as cash. In a few states (Delaware and Virginia, for example), the state EITC supplements are non-refundable.

Unlike the EITC, states' statutory minimum wages have diverged over time (the solid black line shifting upward in Fig. 18.1). They have further seen ebbs and flows in terms of their similarity: when the federal government orders a minimum wage increase, such as in 2009, state similarity naturally increases. From 2009 onward, however, the upward sloping line in Fig. 18.1 represents decisions from many state governments to lift their wage floor above the federal minimum. Though the time series presented in this paper ends in 2015, the divergence has only continued in recent years. In April 2016, two of the largest United States—California and New York—ratified legislation that would, over time, increase the statutory minimum wage for non-tipped workers to \$15 per hour. Pending an increase in the federal minimum wage (set at \$7.25 per hour since 2009) prior to the full phase-in of the two states' new laws, the shift will mark the first time in modern American history when statutory minimum wage levels in a particular state will more than double the federal standard (U.S. Department of Labor, 2015).

Variation in maximum benefit values of AFDC/TANF, however, have remained relatively stable from 1980 onward. Despite "welfare reform" in the mid-1990s, states have not diverged in terms of the benefit levels offered to low-income families. Of course, this is not to say that states are not diverse in terms of their benefit levels. We saw in Table 18.3 that states vary widely, and that mean benefit values are declining over time. As the figure shows, however, these differences have been relatively stable throughout the "devolution revolution." Later, we return to the question of how variation in *coverage* of AFDC/TANF over time might also shape state-level social outcomes.

*Services*—Unlike for our "money" measures, we only have data on (S)CHIP and pre-K spending and coverage for recent years. Thus, we do not present the trends in the CV, as we do for the prior three policies. Table 18.4 presents state-level variation in the indicators in 2015. As discussed before, generosity of pre-K refers to spending per enrolled child, and generosity of Medicaid/CHIP refers to federal and state spending on CHIP and Medicaid per enrolled child. We measure coverage, respectively, as the share of all 3 and 4-year-olds enrolled in a state pre-K program and the share of eligible children enrolled in Medicaid/CHIP.

As Table 18.4 suggests, state-level variation in pre-Kindergarten is greater, both in generosity and coverage, than state-level variation in Medicaid/CHIP in 2015. Variation in state-spending on pre-K (CV of .819) more than doubles that of spending on Medicaid/CHP (CV of .344). This is in part because only 41 states offered pre-K programs in 2014, whereas Medicaid exists in all states. Though we do not have longitudinal data on pre-K and Medicaid/CHIP spending, we can point to recent state-level expansions of Medicaid as one potential source of state-level divergence. After the passage of


**Table 18.4** State-level diversity in generosity and coverage of services dimension of family policy package in 2015

*Note* Generosity of Pre-K refers to spending per enrolled child. Generosity of Medicaid/CHIP refers to federal and state spending on CHIP and Medicaid per enrolled child. Coverage of Pre-K refers to share of 3- and 4-year-olds enrolled. Coverage of Medicaid/CHIP refers to share of eligible children enrolled

the Affordable Care Act in 2010, states governments were granted the option to expand Medicaid access to people with annual incomes below 138% of the federal poverty level. Most of the additional funding for the program would be covered by the federal government. At time of writing, 36 states and Washington, DC, have expanded Medicaid access, while 14 states have held out. We see from Table 18.1 that there is much diversity in state-level Medicaid spending per child, as well as in coverage rates of eligible children. Given the expansion of Medicaid in the 36 states that have accepted it, we may see an increase in state-level diversity in terms of how much the average family spends out-of-pocket on medical expenses. We return to this possibility when evaluating trends in social indicators of family wellbeing in the next section.

*Time*—The unavailability of paid family or parental leave is a well-noted shortcoming of U.S. family policy. As noted before, the federal government ensures only 12 weeks of unpaid family or medical leave for employed adults (with additional tenure requirements, employer criteria and contract type restrictions). As of 2015, only three states offered their own paid family leave plans beyond disability compensation for the birth mother: California, New Jersey, and Rhode Island. New York implemented a paid family leave program in 2018, and Washington will join the fold in 2020 (for details on the development of these and other leave policies, see Chapter 17 by Engeman, in this volume). Further, the generosity and duration of the benefit among these early adopting states is quite modest. In 2015, California paid family leave was capped at 55% income replacement, and available for six weeks. In New Jersey, the cap was set at two-thirds of wages, also for six weeks. Rhode Island provided paid family leave for four weeks with a cap of \$752 per week (National Conference of State Legislatures, 2016). As three states had implemented paid family leave by 2015, divergence in the "time" dimension of family policy has technically occurred since 1980. However, given the small number of adopting states combined with the modest duration of coverage, this divergence has perhaps not been large enough to make a meaningful difference in variance of state-level social outcomes.

# **Diversity in Social Outcomes**

We now examine levels and trends in state-level variation of the six familyrelated social outcomes. We first look at our two measures of child poverty: market income and post-tax/transfer poverty. Table 18.5 presents the diversity of states' child poverty rates in 2015. The average state had a market income child poverty rate of 35.3% in 2015, but with a range from 27.1 to 44.3%


**Table 18.5** State-level diversity in child poverty outcomes in 2015

*Note* Market income poverty refers to poverty rates before taxes and transfers are included

when jumping from the 10th percentile to 90th percentile to state. From 1980 to 2015, states' mean value of market income poverty rates increased by more than 5 percentage points.

Interestingly, the diversity between states grows when bringing taxes and transfers into measures of household income. As the right half of Table 18.5 shows, the mean state featured a post-tax/transfer poverty rate of 22.4% in 2015. The poverty rate in the 90th percentile state (30.6%) nearly doubled that of the 10th percentile state (16.2%). From 1980 to 2015, states' mean value of post-tax, post-transfer poverty rates declined by 2 percentage points. That average post-tax, post-transfer poverty rates declined while marketincome poverty rates increased suggests that taxes and transfers have played an increasingly important role in reducing levels of poverty.

The coefficient of variation is also larger for post-tax/transfer poverty rates, indicating that states are more diverse in their poverty rates after taxes and transfers are taken into account. How has this diversity changed over time? Figure 18.2 displays trends in the coefficient variation for both measures of child poverty: post-tax/transfer (solid line) and pre-tax/transfer (dashed line). We highlight three observations from this figure. First, state-level variance in market income poverty tends to be consistently lower than post-tax/transfer poverty—not just in 2015, as we saw in the prior table. Second, state-level variance in market income poverty has declined over time (as have overall rates of market poverty). Third, despite the convergence in market income poverty, we see relative stability in state-level variation in post-tax/transfer poverty. Insofar as the "devolution revolution" has contributed to more diversity in state-level family policies, it has not appeared to translate into rising variance in post-tax/transfer poverty rates.

**Fig. 18.2** Change in variation of states' pre- and post-tax/transfer poverty rates (*Note* Poverty threshold set at 50% of national equivalized median household income. *Data* U.S. current population survey. CV = coefficient of variation)

One potential reason for a decline in variation in market income poverty rates is a decline in variation of male-to-female employment and earnings gaps. Table 18.6 presents the diversity of the two measures in 2015. Looking at the male-to-female employment ratio, we see that male adult employment is larger than female adult employment in all states. However, the extent of the gap varies meaningfully: in the 10th percentile state, the gap is a mere 1.12, meaning that male employment is 12% higher than female employment. In the 90th percentile state, the gap is 1.37. With respect to median earnings of males compared to the median earnings of females


**Table 18.6** State-level diversity in male-to-female employment and earnings outcomes in 2015

(right side of the figure), we see even more variance. The coefficient of variation is .054 higher for the earnings gap compared to the employment gap. For the average state, the median earnings for employed men was around 57% higher than the median earning for employed women. However, the male-to-female employment and earnings gaps have declined over time, as the bottom rows of Table 18.6 point out.

Figure 18.3 shows how variation among states in these two measures has changed over time. The solid black line represents trends in the CV of statelevel male-to-female earnings gaps, while the dashed line represents the same for the employment gaps. We see that states have converged over time with respect to both indicators, but much more so for the earnings gap relative to the employment gap. Between 1980 and 2015, the CV for the earnings gap fell from around 0.25 to below 0.15—a notable increase in convergence. This convergence in the earnings gap ratio largely reflects the rise in female educational attainment and women's larger-scale entry into managerial positions since the 1980s. However, from 1996 onward, it is also a product of the change that has taken place in U.S. family policy. Across the 50 states, the introduction of TANF has necessitated mothers with caring obligations, who under AFDC could have refrained from active labor market participation, to take low-wage employment. These earnings likely also contributed to the reduction in variance visible in the earnings gap ratio.

**Fig. 18.3** Change in variation of states' male–female employment and earnings gaps


**Table 18.7** State-level diversity in average medical out-of-pocket and childcare expenditures in 2015

*Note* Mean dollar values in 1980 presented in real 2015 USD using PCE deflator

Finally, we can look at diversity in divergence of out-of-pocket medical spending and work and childcare expenditures for families with children. Table 18.7 presents the diversity across states in 2015. In the average state, the average family with children spent around \$4870 out of pocket on medical expenditures in 2015. Between the 10th percentile and 90th percentile states exhibit a more than \$2000 annual difference (\$3867–\$5818). For childcare expenditures, the cross-state variance is slightly smaller (10.2%) compared to the variance for medical expenditures (14.6%). In the average state, the average family with children spent \$3559 on childcare expenditures in 2015. This is a steep increase from the mean values of inflation-adjusted childcare spending in 1980.

As Fig. 18.4 shows, we have seen divergence over time across states. In other words, states vary more with respect to average medical and childcare expenditures for the average family. This is particularly true for medical spending: as Fig. 18.4 demonstrates, the CV increases from .06 to .15 between 1980 and 2015. Most of this increase occurs around the passage of the Affordable Care Act and subsequent Medicaid expansion across many states. This pattern might be explained by the fact that states vary more in terms of the accessibility of health insurance after recent Medicaid expansion drives (more generous states have expanded Medicaid, but less generous states have not), and that Medicaid expansion should drive down out-of-pocket medical costs for low-income families, as more of these families will now have health insurance. Thus, more variance in the accessibility of Medicaid leads to more variance in out-of-pocket medical spending. From 2009 onward, cross-state variance in medical expenditures has been greater than variance in

**Fig. 18.4** Change in variation of states' average levels of medical and childcare expenditures among households with children

childcare expenditures. For childcare spending, we observe a rise in the CV from .07 to .10 between 1980 and 2015.

# **Discussion**

Trends in state-level family policies and social outcomes do not offer a straightforward narrative of divergence. With respect to our money dimension of the family policy package, we observed convergence in state supplements to the federal EITC, stability in the variance of maximum TANF cash assistance benefits, and divergence in statutory minimum wages. With respect to the services and time dimensions, we observed large variation in states' approaches to childcare support, healthcare coverage, and paid family leave time. With respect to our family-oriented social outcomes, we have not seen meaningful increases in variance in child poverty rates or female employment and earnings outcomes from 1980 onward. In fact, the male-to-female earnings gap and market income poverty rates both declined in variance over time—probably both due to the related phenomenon of rising women's educational attainment and earnings across all states over time. Conversely, state-level variance out-of-pocket medical expenditures and childcare spending has risen over time.

These findings point to two primary conclusions and points of discussion. First, the evidence suggests that *diversity*, more so than *divergence*, is the story of state-level variation in family policy packages. As noted in the opening sentence of this chapter, states have long varied with respect to the generosity of family-oriented social assistance benefits. This remains true today. As we saw in Table 18.3, coverage and generosity of states' TANF benefits in 2015 varies widely, even if we have not seen large shifts in variance over time. Consider also the relatively stable diversity in states' child poverty outcomes. Despite large variance in child poverty rates across states (Table 18.5), we have not seen a lot of change in the cross-state variance over time.

Second, federally funded policies that also target families with children likely mitigate some of the social consequences of cross-state variance in family policy packages. Put differently, federally funded redistribution programs such as SNAP, the EITC, or Supplemental Security Income (SSI) likely "offset" some of states' differences in TANF, minimum wages, and so on. There are two primary reasons for this. First, as state-level redistributive programs (TANF, in particular) have declined in value in recent decades, federally funded transfers such as SNAP, SSI, and the EITC (for which benefit levels tend to be common across the states) have increased greatly in recent decades. Second, prior evidence suggests that when a state cuts back on the generosity of its TANF benefits, federally funded SSI and SNAP spending largely fill in the gap (Parolin & Luigjes, 2019). We can visualize this by displaying changes in state-level variation of observed levels of household benefit receipt for the state-administered TANF benefits versus the federally funded SNAP, SSI, and EITC benefits. Figure 18.5 displays the results.

AFDC/TANF clearly stands out in Fig. 18.5: it is the only program in which state-level variance in *observed benefit receipt* has increased from 1980 onward. Thus, despite the stagnation in variance of maximum benefit levels, states have certainly become more diverse in terms of the level of benefits actually allocated to families with children. The three federally funded programs have seen convergence across states over time in terms of mean levels of observed benefit receipt. Put differently, states are becoming more similar in the amount of benefits received from the federally funded SNAP, SSI, and the EITC, but less similar in terms of benefits received from AFDC/TANF. Given that SNAP, SSI, and the EITC now allocate more total benefits each year than the TANF program, the convergence in states' receipt of benefits from these three programs might help explain the stability in state-level variance in poverty outcomes, as observed in Fig. 18.2.

**Fig. 18.5** Change in variation of states' average levels of benefit receipt among households with children

## **Conclusion**

This chapter set out to investigate the diversity and divergence of three sets of family policy indicators across the 50 United States: money, services, and time. We first provided an overview of family policy packages in the United States. Second, we provided descriptive evidence of the diversity of states' family policy packages in cross-sectional and longitudinal perspective. Finally, we provided a descriptive portrait of trends in state-level diversity of several indicators of family wellbeing.

Our findings demonstrate that the 50 United States vary considerably in their family policy packages, particularly with respect to cash assistance for low-income families. States have become more dissimilar over time with respect to social assistance transfers and statutory minimum wages, but have become more similar in their supplements to the EITC. Moreover, states vary greatly in their levels of support for early childhood education and healthcare funding.

Do state-level trends in the variation of family policies align with trends in the variation of family-related social outcomes? We find mixed evidence that this is the case. Despite large diversity and some divergence in states' family policy packages, post-tax/transfer poverty rates have remained relatively stable over time. This is partially due to an increase in federally funded transfer programs mitigating the social consequences of state-level diversity. States *have* converged, however, in levels of market income poverty rates, a trend that appears to be related to declining variation in male-to-female employment and earnings gaps. Finally, we observed that state-level variation in out-of-pocket medical spending has more than doubled from 1980 to 2015, in large part due to some states deciding to expand Medicaid access from 2009 onward.

**Acknowledgements** Rosa Daiger von Gleichen acknowledges support from the United Kingdom Economic and Social Research Council and the Heinrich Böll Foundation.

# **Appendix**

See Table 18.8.

**Table 18.8** Data source and timeframe for family policy inputs and social outcomes


## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **19**

# **Going Regional: Local Childcare Provision and Parental Work–Care Choices in Germany**

**Pia S. Schober**

This chapter argues that taking into account different levels of regional and local institutional variations within a county may have great potential for generating new insights on the drivers of family policy reforms as well as on mechanisms how family policies affect families' choices and well-being. Cash or tax benefits and leave entitlements involving monetary transfers are frequently regulated at the federal level and not heavily dependent on implementation by municipalities and on local infrastructure. Yet, the effects of monetary transfers on families might still vary by regional economic circumstances. The provision of services for families, such as early childhood education and care (ECEC), however, frequently shows even greater regional variation depending on local politics and economic capacities. Drawing on the case of Germany and focusing on early childhood education and care services, this chapter describes the existing regional variations across federal states and counties. Germany is an interesting case due to its federal structure of education and welfare policies. Furthermore, it provides an illustrative example of large and persistent economic and cultural differences in work and care ideals between East and West Germany and also more nuanced cultural and economic differences across other regions.

The chapter first describes the institutional context of ECEC provision in Germany and the existing variation in regional provision and take-up of services. It then reviews different theoretical perspectives on potential drivers

© The Author(s) 2020

P. S. Schober (B)

University of Tübingen, Tübingen, Germany e-mail: pia.schober@uni-tuebingen.de

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_19

of policy variations at the municipality and federal state level and connects these to existing empirical evidence. The second part of the chapter proposes a framework for investigating socially stratified parental work–care choices at subnational levels by connecting a macro–micro rational choice perspective with the capability approach and the accommodation model of childcare choices. After reviewing existing evidence on the effects of regional variations in childcare provision on social inequalities in take-up, maternal employment, and work–family balance and on some of the mechanisms, the chapter concludes by pointing to research gaps and new frontiers of regional family policy analysis. It outlines the current challenges and new demands for data collection and linkages necessary to realize the full potential of regional family policy analysis.

# **Institutional Context of Early Childhood Education and Care in Germany**

# **Legislative and Regulatory Structure**

Early childhood education and care (ECEC) services are understood to include all forms of state-subsidized and regulated forms of group care for children under school age, including mostly childcare centers and to a lesser extent childminders. In Germany, a childminder is a licensed caregiver who looks after up to five children in their own home or third-party rooms. In Germany, ECEC services are not located within the public education system but within the child and youth welfare sector. Political responsibility for ECEC provision is shared between the federal government, the regional governments, and the municipalities (for more details, see Oberhuemer, Schreyer, & Neuman, 2010). Legislative responsibility for all education and care services for children from birth to school entry (usually at age six) is mostly located at the federal government level under the auspices of the Federal Ministry for Family Affairs, Senior citizens, Women, and Youth. Federal legislation then needs to be transformed into independent laws at the regional level of the states ("*Bundesländer* "). Furthermore, municipalities and states are responsible for organizing and financing ECEC provision. In Germany, a guiding policy principle is subsidiarity. This principle states that "the 'lower-level' actor or institution should be given preference over the 'higher-level' actor or institution, on the condition that they can fulfil their task" (Richter, 2009, p. 238). With respect to childcare provision, this can be roughly interpreted as family before state, welfare organizations before state, local communities before state, and member state before federal state (Richter, 2009). It is originally a concept of Catholic social doctrine and not a general principle of German constitutional law, but the German Bundestag decided in the Children and Youth Act 1990 to apply this principle as guidance for the organization of early childhood education and care (Richter, 2009). In the course of Germany's welfare state enlargement, the principle of subsidiarity was interpreted in favor of (often church-related) nonprofit associations and led to a situation where government was obliged to fund the services provided by these nonprofit associations, while at the same time guaranteeing their independence from state interference (Zimmer et al., 2005). This led to administrative structures and political dynamics of childcare provision, in which voluntary sector associations, many of them church-affiliated, play an important role, especially in West Germany (Bode, 2003). Direct public provision by municipalities is more prevalent in the East German states than in West Germany. For-profit providers play a very limited role. Between 2014 and 2018, only about three percent of ECEC centers were run by for-profit providers with little change over time (Bertelsmann Stiftung, 2017; Statistisches Bundesamt, 2018). Only in some large cities for-profit providers make up a larger share, such as 19% of ECEC centers in Hamburg (Bertelsmann Stiftung, 2017).

The financing costs of ECEC centers are largely covered by municipalities (about 47%) and by the state (about 31%), while parents pay on average about 14% (Spiess, 2008). Municipalities are free to design the parental fee structure to cover the running expenses of ECEC centers under consideration of several criteria stipulated under the Code of Social Law. As a result, fees need to be dependent on the number of children in the family and the daily care hours (Naßmacher & Naßmacher, 2007; Tepe & Goerres, 2013). In addition, municipalities usually design fees to be income-dependent with high-earning households paying more (Schröder, Spiess, & Storck, 2015). Fees in Germany are below the OECD average (OECD, 2015). In 2012, they amounted on average to 144 Euros per month and family (Schröder et al., 2015). In the past, around nine percent of parents paid no fees at all (Fuchs-Rechlin, 2008). Some German federal states and municipalities have provided prioritized access to ECEC slots for single parents who receive welfare support (Spiess, Berger, & Groh-Samberg, 2008).

#### **Historical Developments in East and West Germany**

Before 1990, the institutional setting in West Germany was characterized by low levels of publicly subsidized ECEC provision for children under the age of three and by long but low-paid parental leave entitlements. Together with the joint taxation for couples and a family health insurance, still existing, this has frequently been classified as supported familialism (Hook, 2015) and is considered to suppress employment of second earners and, thus, reinforce gender inequality. By contrast, in East Germany, in the former German Democratic Republic, shorter parental leave and extensive provision of formal care for very young children encouraged a faster return to the labor market of mothers (Rosenfeld, Trappe, & Gornick, 2004). Since reunification, employment trends among mothers with young children have converged somewhat and part-time employment has become the most prevalent arrangement to combine employment and family care in both regions (Konietzka & Kreyenfeld, 2010). Since 1996, all children from age three years have been entitled to a slot in a German ECEC center ('*Kindertageseinrichtung*') for at least four hours a day. From age three, over 93 and 95% of children attended daycare in West and East Germany, respectively, in 2017 (Statistische Ämter des Bundes und der Länder, 2018). Yet, significant East–West variations in daily hours spent in care remain with 39% of children attending full-day care (over seven hours a day) in West Germany in contrast to 74% in East Germany (Statistische Ämter des Bundes und der Länder, 2018).

Recent policy developments have promoted a dramatic expansion of ECEC services for children under three years. Since 2005, a federal law ("Tagesbetreuungsausbaugesetz," Deutscher Bundestag, 2004) stipulated that at a minimum, children under the age of three be offered the chance to enroll in an ECEC center if a lone parent or both parents are employed or in education or want to take up employment, or if no other support program promoting the child's welfare is available. A second law in 2008 ("Kinderförderungsgesetz," Deutscher Bundestag, 2008b) outlined that from August 2013 all children aged one or older are entitled to an ECEC slot for four–five hours per weekday. At first, the ECEC expansion was subject to considerable controversy among the parties in government. Therefore, the introduction of the legal entitlement to a childcare place from age one starting in August 2013 was complemented by the introduction of a small flat-rate cash-forcare benefit for parents who do not use formal childcare until the third birthday of the child (Leitner, 2010). This legislation was very contested, as it was predicted to reduce maternal employment and ECEC use among lowincome families (Haan & Wrohlich, 2011). It was then abolished again in 2015. The "*Kinderförderungsgesetz"* in 2008 (Deutscher Bundestag, 2008c) also stipulated that youth welfare office districts should aim at providing a need-oriented supply of full-day places in ECEC institutions for all children from the age of three.

In parallel to the expansion of childcare services, in 2007 the German government reformed the parental leave policy to resemble the Nordic-style parental leave more closely (Leitner, 2010). The main aim was to counteract the low fertility rates and associated welfare state financing challenges by reducing the income loss of middle- and high-income new parents, by facilitating maternal labor market return as well as paternal involvement in child care (Spiess & Wrohlich, 2008). The policy consists of an incomerelated reimbursement at 67% of net earnings (capped at 1800 Euros per month) or a minimum of 300 Euros for twelve months. The reform also includes an individual leave entitlement of two months, which is reserved for each parent to incentivize paternal leave take-up (Deutscher Bundestag, 2008a). However, the parental leave policy still continues to offer a low flatrate benefit for mothers who were not employed before the birth and the possibility to take leave for three years, even if part of it is now unpaid (Leitner, 2010). Overall, despite a dramatic shift in German family and childcare policy, on closer inspection, current policies include elements, which support both the traditional breadwinner and the dual-earner/carer model. In combination, Germany, therefore, has moved toward a model of optional familialism (Stahl & Schober, 2018). As will be shown next, this provides a lot of room for regional variations in provision and social inequalities in take-up.

#### **Regional Variation in Childcare Take-up**

In Germany, the differences in the supply of ECEC slots for children under the age of three are particularly large between East and West Germany. However, even within these two regions, there are substantial variations across states, counties, and municipalities (Statistisches Bundesamt, 2011), depending on the financial situation of these regions and the political priority given to the provision of ECEC services. Some German states, such as Saxony, Thuringia, and Rhineland-Palatine, have entitled specific groups of children under three years of age earlier to a slot in a childcare center before the childcare entitlement at the federal level came into effect in 2013. As shown in Fig. 19.1, in 2017 coverage rates in ECEC for children under three years varied from about 27% in North Rhine-Westphalia, Bavaria, Bremen, and Baden-Wuerttemberg to about 56% in Saxony, Saxony-Anhalt, and Brandenburg and Mecklenburg-West Pomerania. At the municipality level, variations are even larger with less than 20% of children attending ECEC in several Bavarian counties and only five cities in West Germany over 40% of children attended ECEC institutions. By contrast, the attendance rates across

**Fig. 19.1** ECEC attendance rates across Germany states and regions in 2017, in percent (*Source* Statistische Ämter des Bundes und der Länder [2018]. Kindertagesbetreuung Regional 2017. Wiesbaden: Statistisches Bundesamt)

East German cities and counties were mostly above 50% (Statistische Ämter des Bundes und der Länder, 2018). The variation in hours of care is equally large. In 2017, the percentages of children under the age of three who were granted a full-day care slot, defined as more than seven hours per weekday, varied from about 10% in Bavaria and Baden-Wuerttemberg to close to 50% in Thuringia and Saxony-Anhalt. Among children aged three to six years, only 24% of children attended an ECEC institution for more than seven hours a day in Baden-Wuerttemberg compared to over 90% in Thuringia (Statistische Ämter des Bundes und der Länder, 2018).

### **Regulation and Regional Variation of ECEC Quality**

Due to decentralization, German states and municipalities also vary greatly with respect to minimum quality standards for ECEC institutions. Minimum child–teacher ratios are regulated across all German states but the levels required for different ages vary between states (Viernickel et al., 2015). Minimum requirements for most other indicators of structural quality, such as group size, teacher qualifications, and further training, range from precise to very general or none at all. Minimum quality standards and actual conditions often fall short of evidence-based recommendations (Bock-Famulla, Strunz, & Löhle, 2017; NAEYC, 2014).

**Fig. 19.2** Child–teacher ratios across German states in 2016 (*Source* Bock-Famulla, Strunz, & Löhle, [2017]. Länderreport Frühkindliche Bildungssysteme 2017. Bertelsmann Stiftung)

As shown in Fig. 19.2, observed mean child–teacher ratios in ECEC institutions varied significantly across German states in 2016. Only Baden-Wuerttemberg and Bremen match the standards recommended by pedagogic experts of three children aged under three years per educator and 7.5 children per educator for older preschool children (Bock-Famulla et al., 2017; NAEYC, 2014). Actual child–teacher ratios in the majority of states, especially in East Germany, significantly exceed these recommendations. The stricter regulations in several West German states compared to most East German states have been found to be important drivers of the regional variations in the quality conditions of ECEC institutions (Stahl, 2017). Some German states provide additional funding to socially disadvantaged areas or ECEC centers serving disadvantaged children, but the specific regulations again vary across states (Hogrebe, 2014).

## **Drivers of Regional Variation in Childcare Policy Provision**

Recently, a growing number of studies in political science have advocated for more attention to be paid to the subnational level of policy provision (Andronescu & Carnes, 2015; Greer, Elliott, & Oliver, 2015; Oliver & Mätzke, 2014; Tepe & Goerres, 2013). They have suggested that this may be particularly important for policies aimed at preventing or reducing new social risks, e.g., linked to the decline of the male breadwinner family model, or social investment policies, as these policies align less clearly with traditional party-political logics and are often subject to decentralization of some of the decision-making and most of the implementation.

As described above, one important driver of regional variation in childcare policy provision includes the extent to which regulatory powers, financing structures, and responsibilities for the implementation of childcare services are decentralized (Greer et al., 2015). While national governments have frequently engaged in large-scale campaigns of childcare expansion, the principles underlying decentralized implementation and the implementation structure for the deployment of funds affects the timing, pace, and nature of any expansion.

The political economy perspective has argued that partisan politics also play a crucial role in the design and implementation of childcare policy. Parties respond to social trends, such as the rise of female employment, but they also remain affine to their historical bases and ideological programs (Bonoli, 2005; Esping-Andersen, 1999; Huber & Stephens, 2000; Williamson & Carnes, 2013). In particular, leftist parties have long been associated with the expansion of social welfare spending and services, as well as with defamilialist service provision such as childcare. Such policies free parents, and especially women, from family care obligations and allow them to engage in the labor market. Conservative parties have tended to resist such expansions. Debates whether partisan politics are less influential at the municipal level compared to the nation-state level are still ongoing. Yet evidence for Germany generally points to some influence of political party representations on childcare expansions (Andronescu & Carnes, 2015; Busemeyer & Seitzl, 2018). By contrast, the "partisan convergence" approach has argued that parties with divergent ideological orientations have adopted overlapping policy goals regarding issues such as family policy due to the emergence of an increasingly de-aligned female vote (Morgan, 2013) and the presence of female political leaders supporting the expansion of childcare irrespective of their political party affiliation (Leitner, 2010).

In addition, several structural factors have been considered of importance for expanding public childcare institutions. First of all, demographic factors such as the prevalence of families with young children compared to older citizens and women's employment rates are likely to drive up the demand for childcare places for young children (Andronescu & Carnes, 2015; Tepe & Goerres, 2013). Secondly, varying care ideals and acceptance of formal childcare for young children have been attributed to longer term ideological influences by religious institutions, such as the Catholic church in many European countries (Andronescu & Carnes, 2015; Kremer, 2007). Thirdly, varying economic resources of states and municipalities are likely to provide different opportunities for additional spending on childcare. Next, I will consider each of these potential drivers more in detail for the case of German regions followed by a discussion of to-date evidence on their importance for subnational variation in childcare provision, quality, and fees.

#### **Evidence on the Importance of Partisan Politics, Female Representation, and Structural Variations Across German Regions**

Two studies have provided evidence for the relevance of partisan politics in the course of the German childcare expansion (Andronescu & Carnes, 2015; Busemeyer & Seitzl, 2018). The government participation of left-wing parties (SPD, Die Linke or the Greens) at the state level was positively and significantly associated with regional variations in the level of childcare availability for children under the age of three years with the expansion of such childcare slots since 2006 (Andronescu & Carnes, 2015) as well as with changes in public spending on early childhood education between 1992 and 2010 (Busemeyer & Seitzl, 2018). Tepe and Goerres (2013) also found that leftwing party majorities at the municipal level were associated with higher fees for groups of parents with medium and high incomes.

Qualitative studies have suggested that leading female politicians at the federal level played a key role in pushing through the childcare expansion in Germany (Leitner, 2010). At the municipal level, Tepe and Goerres (2013) found that fees were lower for high-income parents in city councils with a higher percentage of female representatives, irrespective of their political party affiliation. This suggests that they disproportionately favor affordable childcare for groups of women who are similar to themselves.

Economic and structural factors were frequently found to be important in driving regional variations in childcare availability, quality, and costs. These include economic resources of municipalities with wealthier municipalities spending more on ECEC services (Busemeyer & Seitzl, 2018) and creating more childcare slots for under-threes (Andronescu & Carnes, 2015). Structural ECEC quality conditions, such as child–teacher ratios, were also found to be more favorable in more affluent neighborhoods (Stahl, 2017). Furthermore, the higher the level of state spending on ECEC, the lower the fees for parents at the municipality level (Tepe & Goerres, 2013). Also, Mätzke (2019) argued that East–West differences in the speed of the childcare expansion were influenced by the institutional legacy of long-standing differences in the childcare policies between the two regions. Due to more widespread childcare provision to start with and population decline, overcapacities for children over age three in East German regions were transformed into services for the younger age group. Population structure seems to matter with the percentage of children under the age of three being negatively associated with ECEC availability for this age group, whereas more widespread female employment shows a positive association (Andronescu & Carnes, 2015). The population structure seems less important for the fee structure (Tepe & Goerres, 2013). Whereas varying cultural legacies and childcare ideals seem crucial to understand the persistent differences in childcare provision between East and West Germany (Mätzke, 2019), greater prevalence of Catholicism among the population is not significantly associated with variations across states within East and West Germany, respectively (Andronescu & Carnes, 2015).

# **Conceptual Framework for Analyzing Effects of Regional ECEC Provision on Families**

This section will argue that taking into account local variations in childcare provisions is also promising for a better understanding of direct and indirect effects these might have on families and children. To develop a theoretical framework of how different aspects of childcare provision at the local level affect parents' opportunities and capabilities, we combine three theoretical perspectives: (1) a rational choice macro–micro perspective frequently applied in mechanism-based research and neighborhood studies (Coleman, 1990; Esser, 2009; Friedrichs & Nonnenmacher, 2010), (2) the capability approach (Nussbaum, 2011; Sen, 1999), and (3) the accommodation model of childcare choices (Meyers & Jordan, 2006).

All three approaches suggest that childcare policy and the characteristics of the local provision directly affect parents' opportunities and frame the choices they can make following childbirth. The first relevant perspective leads us back to the macro–micro model, originally developed by Coleman (1990) and extended by more recent proponents of wide versions of rational action theory (e.g., Esser, 2009; Hedström & Ylikoski, 2010), especially in the context of analyzing neighborhood effects on micro-level actions (Friedrichs & Nonnenmacher, 2010). One strength of this perspective is that it makes the distinction of direct and indirect effects and the need to specify chains of theoretical mechanisms very explicit. Friedrichs and Nonnenmacher (2010) distinguish between two mechanisms of direct effects of social contexts on individual outcomes: (i) local institutions that provide or hinder various activities, and (ii) collective socialization effects which provide normative settings for choice sets and the adoption of behaviors. The latter norm-setting processes have frequently been found to be nonlinear with stronger contagion effects above a certain threshold. With respect to regional and local childcare policies, variations in childcare services, such as availability of slots or opening hours are likely to directly affect parents' choices of employment hours. Indirect effects are equally conceivable, as reaching a certain threshold of childcare availability and take-up will change acceptability of this type of care arrangement and perceptions of entitlement to different combinations of employment and childcare. Indirect effects of childcare institutions may include effects on more intensified or reduced contact between specific groups of parents, e.g., by meeting during pick-up and drop-off. If characteristics of childcare institutions vary, e.g., in terms of quality and flexibility, in addition selection effects may occur with parents with similar preferences choosing the same institutions resulting in intensified social network and norm-setting effects.

The capability approach, originally developed by Sen (1999) and extended by Nussbaum (2011), offers the most comprehensive, yet also the least tangible, conceptualization of family outcomes and processes. In their application to childcare policy, Yerkes and Javornik (2018; see also Chapter 7 by Javornik and Yerkes in this Volume) refer to childcare policy as a resource for parents to arrange childcare in a way that allows for pluralist interpretations of what individuals have reason to value in life. Yerkes and Javornik (2018) suggest that five salient characteristics of childcare services comprise availability, accessibility, affordability, quality, and flexibility. According to their understanding, availability, accessibility, and affordability provide the foundation for childcare capabilities, whereas quality and flexibility only become more important once childcare is available, accessible, and affordable.

When considering parents' decision-making and resulting implications for outcomes, such as employment, work–family balance, and well-being, the capability approach assumes that even when individuals have equal access to means in terms of financial resources or childcare policy, they may not achieve the same outcomes or functionings (Sen, 1992). Key mechanisms here are variation in an individual's situated agency (Hobson, 2014, 2018) depending on conversion factors which reflect their individual resources and relational embeddedness in the community and social contexts. In combination, these may result in different capability sets representing different options and perceived alternatives from which individuals choose.

Similar to the capability approach, the accommodation model of childcare choices has been developed to complement or even substitute a previously dominant rational choice perspective with a rather narrow economic framing (Becker & Tomes, 1986; Blau, Ferber, & Winkler, 2002). The latter has been frequently criticized for assuming that (i) parents are perfectly informed about all childcare options, and (ii) parents have homogeneous and relatively fixed ex-ante preferences for childcare characteristics (Chaudry, Henly, & Meyers, 2010; Meyers & Jordan, 2006). The accommodation model seeks to combine a rational action perspective of parents with insights on varying information on the childcare system, and the role of social networks in processing information and making decisions. It suggests that parents adapt their childcare preferences based on context-specific care availability and easily accessible information. From a capability perspective, one might argue that the accommodation model of childcare choices focuses on specific conversion factors—knowledge and preferences and their social embeddedness—that shape parents' agency and may, for instance, result in substantial social disparities in childcare capabilities and family functionings.

The three perspectives can be fruitfully combined, as the capability approach leaves a lot of scope for specifications of more detailed processes of how individual resources and conversion factors impact the extent to which institutional provisions as resources actually result in individual capability sets and achieved functionings. The combination with a more mechanism-based perspective on causal chains might enhance our understanding of direct and indirect effects of local childcare policy institutions and help to shed light on some of the key conversion factors, such as how individuals come to perceive different choice sets and develop a sense of entitlement to policy support. Childcare policy evaluation studies have frequently sought to analyze the effects of different aspects of childcare provision on social inequalities in take-up of childcare and on gender inequalities in employment and work– family balance. This chapter therefore outlines how a combined framework may contribute new insights on these relationships. Social disparities are likely arise in part due to variations in individual means or resources, including gender, education, income, and time variations. Conversion factors are likely to vary at the community level depending on the presence of a partner and other social networks and their resources as well as at the societal level in terms of cultural norms which tend to vary by gender and ethnicity (for a graphical illustration, see Fig. 19.3).

**Fig. 19.3** Conceptual framework for analyzing effects of regional childcare provision (*Source* Adapted from Hobson [2018] and Javornik and Yerkes [Chapter 7 in this volume])

Depending on the national and local childcare institutions, personal and community resources will vary in their importance for shaping achieved functionings. Higher childcare costs, for instance, will increase income disparities, whereas prioritized access rights for single parents are likely to reduce the impact of partner resources. Cultural childcare norms and variations across gender and ethnic groups are reinforced by socially stratified networks, which also shape the accessible information. These moderate the process how national and local childcare provisions convert into perceived childcare alternatives and result in a varying sense of entitlement to childcare arrangements with valued characteristics. The resulting capability set is likely to show substantial variation which is then also reflected in different achieved functionings in terms of actual childcare experiences, parental employment, and work–family balance. These achieved functionings at the individual level in turn affect inequalities in childcare and education experiences as well as longer term labor market outcomes including income inequalities.

#### **Evidence on Effects of Local Childcare Provisions and Potential Mediators**

In Germany, parents can generally choose freely between ECEC centers, as there are no designated catchment areas. A recent analysis (Stahl, 2017) found that three-quarters of parents choose centers within 2.1 kilometers of their home (for similar analysis of parents' proximity to childcare institutions in the Netherlands, see Chapter 20 by Emery in this Volume). Given parents' strong preference for centers close to their home, families in densely populated areas likely enjoy much greater choice than families in rural areas. For instance, an analysis based on geocodes from Berlin suggests that, in the inner city, families often have 20 and more centers at close range (Franke, Pieper, Kürten, & Schweikart, 2015). In the representative K2ID survey from 2013 to 2014, however, 91% of parents reported that they had some choice between different centers (Stahl, Schober, & Spiess, 2018). As described above, childcare resources in terms of availability and opening hours are more restrictive in the West compared to East Germany, whereas structural quality characteristics tend to be more advantageous in some West German regions. With respect to all these characteristics as well as costs, German counties and municipalities, however, show substantial variations.

Existing studies provide consistent evidence of pronounced, and even increasing, social inequalities by education and migration background in take-up of childcare institutions, especially for children under three years of age and with respect to full-day care also for older children (Jessen, Schmitz, Spiess, & Waights, 2018; Schober & Spiess, 2013; Schober & Stahl, 2014; Stahl & Schober, 2018). Interestingly, the increase in social disparities in childcare take-up and maternal employment unfolded in similar ways in East and West Germany. Drawing on data from the International Social Survey Program from 1994, 2002, and 2012, Schober & Stahl, 2014 also found that the change toward less traditional gender ideologies was more pronounced among medium and highly educated respondents compared to the low educated both in East and West Germany. However, the drivers of increasing social disparities in childcare take-up and attitudes toward maternal employment are likely to differ between East and West Germany. In West Germany, the new incentives of shorter but income-related parental leave benefits in combination with wider availability of formal childcare corresponded better with labor market opportunities and work orientations of higher-educated mothers and enabled them to achieve their desired combination of employment with formal childcare. In East Germany, despite previously dominant cultural norms of full-time employment, the constantly lower job prospects may have kept low educated mothers from returning to the labor market faster and from using the increasingly available childcare resources from early on (Stahl & Schober, 2018).

Notably, social disparities between education groups so far have not narrowed since the introduction of the legal entitlement to a childcare place (Jessen et al., 2018). Variation in financial means in terms of household income seems to be of lesser importance in the German context (Schober & Spiess, 2013), probably due to the income-dependent fee structure. Single mothers have been found to disproportionately increase their childcare takeup in the process of the recent expansion of childcare availability, possibly because they have been entitled to prioritized access across many municipalities (Stahl & Schober, 2018). Recent studies also provide some evidence of modest differences in quality characteristics of ECEC institutions attended by children whose parents vary in their individual means of education and social normative conversion factors related to ethnic backgrounds (Becker & Schober, 2017; Stahl et al., 2018). Again, household income and single motherhood was not systematically associated with lower quality of ECEC institutions as part of the achieved childcare functionings (Stahl et al., 2018).

By connecting regional data of the ECEC expansion for under-three-yearold children with individual-level survey data, Zoch and Hondralis (2017) found that mothers in regional counties with greater childcare availability returned earlier to the labor market after a second birth. They also found positive effects on the probability of mothers returning to part-time or fulltime employment as opposed to marginal employment. This might be partly explained by altered conversion factors in the form of increasing social acceptance of maternal employment and formal care for young children among mothers in West German regions where ECEC availability increased substantially from a relatively low level (Zoch & Schober, 2018). To answer the question whether the greater availability of ECEC institutions contributed toward better reconciliation of employment and family care, two studies explored how the ECEC expansion for under-three-year-old children and the extension of childcare hours across counties was associated with changes in parents' subjective well-being. They found positive effects of the expansion of full-day childcare for all children up to school age on satisfaction with family life and with life overall for mothers in East Germany (Schober & Stahl, 2016). In West Germany, the results only pointed to positive associations of full-day take-up of childcare resources for single mothers and full-time employed mothers. These results suggested that conversion factors seemed crucial, as mothers in East Germany may feel a greater sense of entitlement to using full-day childcare services also during spare time beyond their employment hours, which may be partly due to the greater availability as well as the greater cultural acceptance of formal childcare for young children. In West Germany, priority access was given to single mothers and to mothers who needed full-day care to cover their full-time work hours. As a result, probably these groups felt most entitled to using this type of care were able to convert the expanded childcare resources into greater subjective well-being. For the expansion of ECEC slots for under-threes, the results also pointed to improved functionings in terms of subjective wellbeing for single mothers (Schober & Schmitt, 2017). Transitions to full-time employment were associated with reductions in subjective well-being irrespective of local availability of childcare resources among partnered mothers in West Germany but not in East Germany. These results suggested that the greater acceptance of maternal employment and formal childcare use in East compared to West Germany was a more important moderator of the relationship between maternal employment and satisfaction than short-term regional expansions of childcare resources. This points yet again to cultural conversion factors—possibly as a result of longer-term policy institutions shaping perceived alternatives and the sense of entitlement. Fathers' subjective well-being seemed unaffected by variations in childcare resources, cultural conversion factors, and functionings in terms of different combinations of employment and care (Schober & Schmitt, 2017).

## **Evidence on Possible Mechanisms**

The observed social group differences in childcare capabilities might partly be owing to differences in conversion factors, such as preferences for specific childcare aspects (Klein, Biedinger, & Kolb, 2016; Stahl et al., 2018) or in their information behaviors (Vandenbroeck, Visscher, & Nuffel, 2008), which are likely to go hand in hand with a varying sense of entitlement to a high-quality childcare place.

Research from Germany as well as other countries suggests that parents' understanding of the childcare available, how to obtain a place, and subsidies, remains limited—especially among ethnically and linguistically more isolated groups (Becker, 2010; Burghardt & Kluczniok, 2016; Vorsanger, 2005). Also, in terms of knowledge and perceptions of ECEC quality, studies have found significant discrepancies between parents and the care providers (Camehl, Schober, & Spiess, 2018; Cryer & Burchinal, 1997; Helburn & Bergmann, 2002; Mocan, 2007) because it is difficult for parents to observe many qualitative features of childcare. Some studies from the US and several European countries found that parents with higher educational attainment rate the quality of their children's classrooms slightly lower and more accurately than less educated parents (Cryer, Tietze, & Wessels, 2002; Mocan, 2007), whereas a recent German study found no systematic differences across education groups (Camehl et al., 2018). Less-educated as well as ethnic-minority and foreign-language families in Belgium (Vandenbroeck et al., 2008) were found to spend less time on the search and to register later than other groups of parents. In Germany, Turkish parents visited on average fewer centers prior to registration than German parents (Klein et al., 2016).

A few US studies found that higher parental education or financial resources correlated with stronger preferences for ECEC quality criteria and less importance attached to practical concerns apart from costs (Johansen, Leibowitz, & Waite, 1996; Peyton, Jacobs, O'Brien, & Roy, 2001), whereas others did not (Cryer et al., 2002; Kensinger Rose & Elicker, 2008). In Belgium, ethnic minority parents attached less importance to the quality of the ECEC infrastructure as well as the opening hours of a center than Belgian parents (Vandenbroeck et al., 2008). In Germany, parents with college education were nearly ten percentage points more likely to report characteristics of pedagogic quality (e.g., child–teacher ratios, pedagogical concept) as opposed to practical considerations (e.g., proximity, opening hours) as the most important criterion than those with lower levels of education (Stahl et al., 2018). Parents with migration background were more likely and native-German parents were less likely to use the nearest ECEC center if it had a large proportion of children with migration background (Klein et al., 2016).

Given that social networks tend to be stratified by location, race/ethnicity, and other sociodemographic characteristics (Chaudry, 2004), these factors will shape the conversion factors, including recommendations parents receive, their awareness of different alternatives and subsequently their childcare preferences and sense of entitlement. To better understand the mechanisms of childcare capabilities, future studies need to go down to the regional and local levels to actually consider local provisions jointly with parents' personal means and conversion factors which drive their subsequent childcare experiences and family functionings. To-date, there is no systematic evidence as to how key aspects of local childcare provision, such as availability and quality, might moderate the importance of different conversion factors for explaining social and gender inequalities.

## **Discussion**

The first part of this chapter has described the institutional context of ECEC provision and the pronounced regional variations. It then reviewed different theoretical perspectives on potential drivers of childcare policy variations at the municipality and federal state level. In line with the international literature (Greer et al., 2015), one important driver of substantial regional variation in German childcare policy provision is the decentralization of regulatory powers, financing structures, and responsibilities for the implementation of childcare services. Whereas in other countries, debates whether partisan politics are less influential at the municipal level compared to the nation-state level are still ongoing, evidence for Germany generally points to some influence of political party representations and female political leaders at federal state and municipality level during the recent childcare expansion (Andronescu & Carnes, 2015; Busemeyer & Seitzl, 2018; Tepe & Goerres, 2013). In addition to varying economic resources of states and municipalities, the supply of childcare services is also affected by the demand of voter groups in terms of the share of families with children and female employment rates. Furthermore, differing cultural care ideals seem crucial to understand the persistent East–West differences (Mätzke, 2019).

The second part of the chapter proposed the cornerstones of a conceptual framework for investigating socially stratified parental work–care functionings at subnational levels. It has connected a macro–micro rational choice perspective of neighborhood effects with the capability approach and the accommodation model of childcare choices. It outlined how the study of social inequalities in childcare experiences and gender inequalities in employment and work–family balance might benefit from combining the conceptualization of the capability approach with a more mechanism-based perspective on direct as well as indirect effects of local childcare policy institutions. This would help to shed light on specific conversion factors, such as how individuals come to perceive different choice sets and develop a sense of entitlement to policy support. In particular, the framework has aimed to give specific examples how variations in individual means, such as education and income, and conversion factors, for instance, in terms of social networks, are likely to lead to variations in perceived and preferred alternatives and a substantially different sense of entitlements and subsequent access to high-quality childcare. Depending on the local childcare institutions, these resources and conversion factors will vary in their importance in shaping achieved functionings. Furthermore, cultural childcare norms, which are frequently socially stratified e.g., with respect to gender and ethnicity, moderate these processes. The resulting capability set is likely to show substantial variation which is then also reflected in different achieved functionings in terms of actual childcare experiences, maternal employment, and work–family balance. The review of existing empirical studies on the effects of regional and local childcare provision in Germany pointed to substantial social inequalities in quantity and quality of childcare take-up (Stahl & Schober, 2018; Stahl et al., 2018) and to heterogeneous effects on maternal employment (Zoch, 2018; Zoch & Hondralis, 2017) and work–family balance (Schober & Schmitt, 2017; Schober & Stahl, 2016). It provides examples of the moderating role of work–care culture in explaining not only persistent differences in childcare take-up and maternal employment but also differences in the effects of the recent childcare expansion between the East and West Germany (Schober & Schmitt, 2017; Schober & Stahl, 2016). It also presented existing evidence of some key conversion processes through socially stratified information on available alternatives, sense of entitlements, and preferences for varying childcare arrangements (Mocan, 2007; Stahl et al., 2018; Vandenbroeck et al., 2008).

Yet, it also became obvious that we are still missing empirical evidence on a number of pieces in the puzzle. In particular, the indirect effects of local childcare institutions through forming and reinforcing social networks and norm-setting processes would hold great promise for future research on social stratification of childcare experiences as well as parental employment and work–family balance. Another open question concerns the most appropriate level of analysis. Depending on the outcomes of interest and the specific political and legislative context, a close link to boundaries marked by political institutions or even smaller local neighborhoods aligning more closely with social network might be favorable. To dive deeper into these issues, statistical offices in Germany as well as other countries would need to collect and provide easier access to detailed data at different regional levels and allow linking these with large-scale survey data on social networks. Another interesting extension would be to extend a few existing studies (Neimanns, 2017; Zoch & Schober, 2018) on policy feedback processes by examining how the expansion of childcare shapes voters' work–care attitudes and ideals over the short- and long-term and how this in turn affects positions of political parties and subsequent reforms. Subnational analyses of countries with decentralized structures, such as Germany, are particularly promising to investigate how substantial regional and social inequalities and shorter term variations therein influence future policy development within similar long-term cultural and institutional contexts.

The theoretical frameworks applied for exploring key drivers of childcare provision and consequences for families are equally applicable to other social policies with decentralized responsibilities of legislation and implementation, which frequently is the case for child welfare and protection policies, education policies, or labor market activation programs. Accordingly, some of the findings regarding important conversion processes which shape social inequalities and family functionings may be transferable to these other policy domains. On the whole, the aim of this chapter was to show that subnational analyses exploiting regional variations have great potential. Given data availability at local levels, subnational analyses are likely to be advantageous over cross-national comparisons or historical single-country studies, for instance, by allowing a closer identification of direct and indirect effects of local policies and institutions and for analyzing the determinants and consequences of reforms separately from longer term cultural and institutional context factors.

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **20**

# **Private Childcare and Employment Options: The Geography of the Return to Work for Mothers in the Netherlands**

**Tom Emery**

In this chapter we examine the extent to which the return to work for mothers is affected by their proximity to various resources, facilities, and opportunities at the sub-regional level. Geocoded data from the Netherlands Kinship Panel Study (NKPS, waves 1–4, 2005–2014) are matched with geocoded data on childcare provision from administrative data, neighborhood level data on unemployment levels, and the proximity of individual's parents and other relatives who might act as childcare alternatives. The rich spatial data combined with a longitudinal approach enables the testing of a geographical perspective to women's employment. Crucially, the analysis looks at whether nurseries, childminders, or a mix of childcare institutions are most effective in supporting female employment. Given developments in Dutch childcare policy over the period under study, the analysis examines important policy considerations, especially the need to identify and address childcare deserts.

The relationship between childcare supply and female employment is a pressing policy issue in many developed countries. Central to this discussion is disentangling causal relationships between the form, quality, affordability, and accessibility of childcare and the quantity and quality of female employment. Elements of this discussion are further elaborated in, for instance, Chapter 11 by Hook and Li, Chapter 7 by Javornik and Yerkes, Chapter 8 by Vandenbroeck, and Chapter 24 by Sirén, Doctrinal, Van Lancker, and

T. Emery (B)

Erasmus University Rotterdam, Rotterdam, The Netherlands e-mail: tom@odissei-data.nl

Nieuwenhuis, all in this volume. There have been numerous studies examining the geographical distribution of childcare provision that have illustrated that it is almost never uniform, with clusters of high provision in urban, high income areas (Hank & Kreyenfeld, 2003; Van Ham & Mulder, 2005). Provision tends to be greatest in areas where the female employment rate could be said to be high for other reasons such as greater general prosperity, more job opportunities, and families selecting into areas with good childcare provision when they want to work (Dickens, Wollny, & Ireland, 2012; Holloway, 1998). Most analyses are also conducted on an aggregate small area basis given that individual level data is not available and this therefore risks conclusions based on ecological fallacies (Asai, Kambayashi, & Yamaguchi, 2015).

In this chapter we aim to utilize the unique features of the Netherlands Kinship Panel Study (NKPS) and Dutch Administrative Data on childcare facilities to examine whether the availability and type of childcare facilities in an individual's immediate vicinity are associated with mothers' labor market activity (employment type and hours worked). The Netherlands is a relatively homogenous country with high population density. Even still, the level of childcare provision is variable between locations as childcare centers are operated by private companies and the density of childcare providers reflects this. Furthermore, Dutch childcare is dominated by two forms of childcare: *kinderdagverblijf* (childcare centers) and *gastouders* (childminders). These options can be used in combination with each other or in isolation. Both are subsidized by the Dutch Government at a rate that is dependent on the parents' income. Childminders are widespread and pervasive as they are small and relatively easy to establish. Childcare centers are, however, larger operations and tend to be more variable in their provision. The options open to parents are therefore shaped by the availability of both forms of care.

This market-driven approach to provision risks geospatial and structural inequalities in access to childcare that potentially have consequences for female employment and other policy aims. These geospatial and structural inequalities in access to childcare are, however, tightly intertwined with other structural and geospatial inequalities such as poverty, a lack of employment opportunities, housing affordability, or many other factors. This makes any causal assessment exceptionally difficult and is a common feature of geospatial, sub-regional level analyses.

The longitudinal nature of the NKPS data, which has been collected over four waves between 2005 and 2014, allows us to take many of these selection effects into account and employ more robust within-individual estimates of maternal labor market activity. Through a series of models, we examine a variety of factors which may shape and determine the decision of a mother to work and the level of work she chooses to undertake. These include the geographic proximity of family and support networks, the local availability of jobs ascertained through low level employment data, and local small area demographics (i.e. age and ethnicity profile of the neighborhood).

The results provide a rich insight into the labor market decisions of women in the Netherlands. The findings illustrate the strong association between the geographical availability of childcare and the ability of women to work longer hours. However, more informal forms of childcare support such as childminders or family networks are not directly associated with a return to employment. These conclusions illustrate the vital role that childcare provision plays in enabling women to work a greater number of hours on the labor market.

## **Existing Research**

Existing research on the association between childcare and female employment has centered on two dimensions of childcare availability; price and the number of places available. The ability of childcare to support maternal employment is of course heavily dependent upon the net cost of formal childcare. Studies have focused on quasi-experimental methods such as difference in difference models. For example, Havnes and Mogstad (2011) utilized an expansion of subsidies in Norwegian childcare to examine the impact of individuals' employment levels. Their findings appear to show that there is little to no impact on employment suggesting that previous research using aggregated data failed to taken into account potential ecological fallacies, omitted variables, and selection effects (Jaumotte, 2005). Using a similar design in Sweden, it was also found that the effect of a drop-in price was zero (Lundin, Mörk, & Öckbert, 2008). In the Netherlands, the Dutch government's assessment of the introduction of subsidies in the Netherlands in 2005 using a difference in difference analysis, also indicated only a modest increase in female employment of around 3 percentage points (Bettendorf, Jongen, & Muller, 2011). Studies from Australia (Rammohan & Whelan, 2006), Finland (Viitanen, 2011) and the UK (Viitanen, 2005) have also shown at best modest impacts of the effect of price on female employment.

This provides a pessimistic picture regarding the role of childcare in supporting female employment. Yet, basic rational choice models of labor market participation suggest that affordable childcare is a key part of a broader family friendly policy package to support female employment (Blau & Kahn, 2013). To better understand the modest findings of existing research on childcare and female employment, it might be necessary to consider additional dimensions of childcare provision that may affect the ability of women to participate on the labor market. There are a number of studies that have attempted to do this, but data limitations are a considerable problem for such questions.

Van Ham and Mulder (2005) found that in the Netherlands, the probability of a woman being employed increased with the number of slots available in childcare facilities in their immediate area. The findings were able to utilize a very local level of data but are still limited by potential endogeneity surrounding individuals housing decisions and the potential for omitted variable bias at the small area level. In Germany, Hank and Kreyenfeld (2003) use regional data on child care provision but this showed no association with the probability for women to be engaged on the labor market. The study was however limited by the detail of data available to them and size of the geographical units being employed. In Japan, Asai et al. (2015) use small area estimates for municipalities and employ fixed effects using longitudinal data from 1990 to 2010 but show that there is again no effect of childcare provision at the municipal level. They reason that this merely reflects that informal childcare is a perfect substitute for formal childcare and thus the impact of public childcare provision is minimal.

In addition to studies of childcare provision, there is also an extensive literature on childcare substitutes, especially the availability of grandparental childcare. Posadas and Vidal-Fernandez (2013) found that in the United States the availability of grandparental childcare increased the female employment rate by up to 9 percentage points. This is also supported by findings in Europe which have suggested that grandparental childcare support can greatly increase the likelihood of maternal employment (Hank & Buber, 2009). These empirical findings regularly argue that these effects are attributable to the flexibility and support offered by grandparental care provision, intermating that flexibility is a key element in resolving work and family conflicts which prevent women from working for longer hours during their child's early years (Sarti, 2010; Wheelock & Jones, 2002). Also of interest within the context of this chapter is the persistent finding that childcare provision by grandparents is itself highly dependent on the physical proximity of the grandparent, representing a geographical constraint on employment of the mother (Zamberletti, Cavrini, & Tomassini, 2018).

Existing studies are of course hamstrung by data limitations, and the increasing availability of geographic data and longitudinal survey data is greatly improving the capacity of social scientists to provide more acute insights into individuals' circumstances. This is especially the case when we consider that finding childcare arrangements that enable higher levels of employment involves a complex set of factors including the availability of auxiliary support such as grandparents and the availability of work. In this paper, we go down beyond small area estimates to provide an individual's unique position within a broader childcare context. By utilizing geocoded data on childcare facilities in the Netherlands it is possible to estimate the precise proximity of individuals to various childcare facilities and examine what specific dimensions are most important in determining the level of women's engagement on the labor market.

## **The Context in the Netherlands**

During the 1980s and early 1990s the employment rate in the Netherlands grew considerably above the pace of other OECD countries in what came to be known as the 'Dutch Miracle' (Visser & Hemerijck, 1999). However, the expansion in employment and particularly in female employment was largely driven by an increase in part-time work (see Fig. 20.1).

**Fig. 20.1** NKPS respondents in Den Haag (*N* = 241)

This trend has continued and the Netherlands now has the highest rates of part-time employment in the OECD. The large number of part-time workers has been recognized as a 'One and a half worker' model in which a large number of couples operate half way between the traditional breadwinner model and a full adult worker model (Lewis, Knijn, Martin, & Ostner, 2008). This is usually represented by men working 1–0.8fte and women working 0.4–0.6fte. Key to this arrangement is flexibility on the part of employee, employer, and the state in ensuring that the balance is carefully managed and facilitated, an approach characteristic of the Dutch model (Fokkema, de Valk, de Beer, & van Duin, 2008; Woldendorp, 2005).

The parental leave system in the Netherlands is an example of this flexibility as leave entitlements are specified in hours rather than days and can be used in any configuration as long as it is agreed upon by the employer and employee (Den Dulk, 2016). This flexibility enables both men and women to take parental leave as an afternoon per week, or a day per week over a much longer period of time rather than as either a single chunk or series of blocks. The replacement rate for such parental leave varies depending on which collective agreement the employee is subject to, and although a significant minority have no paid parental leave provided for within their collective agreement, the replacement rate generally varies between 50 and 75% of an individuals' average salary. The result of this is that women choosing to only work 3–4 days per week can do so for the first few years of a child's life with only limited financial impact. However, despite this policy being gender neutral, parental leave usage among men is much lower (Keizer, 2015).

With regards to childcare, the emphasis in the Netherlands is also on supporting employment while simultaneously allowing for flexible arrangements. The first Childcare act was introduced in on 1 January 2005, coinciding with wave 1 of the NKPS. This has led to a significant shift in childcare provision over the period of study (Fokkema et al., 2008). The 2005 act prohibited municipalities from providing childcare directly and eliminated subsidies from municipalities and employers. Instead, a subsidy would be provided directly by the central government to parents which was to be used to pay for childcare from approved commercial or non-profit providers (Knijn & Lewis, 2017). In 2004, before the act was introduced, just 30% of all childcare facilities were run by private organizations, with the majority of the remaining facilities operated by local authorities, but this had risen to 70% by 2010 (Bettendorf et al., 2011).

There are two primary forms of formal childcare for children aged 0–4 provisioned for in the Childcare act which are childcare centers and childminders. Childminders often provide more specific needs but are commonly

single individuals who are willing to provide childcare from their own home and are unattached to specific institutions. Childcare centers vary in size but generally have around 40–70 places. They have fixed opening hours which generally last from 7:30 (a.m.) through to 18:30 (6:30 p.m.). Unlike with parental leave, childcare centers charge only with regards to days and do not allow for reductions in costs if children do not spend the full day in the childcare center. This limits the flexibility in combining childcare with parttime work. According to data from the Ministry of Social Affairs, the number of childcare centers has risen from approximately 3464 in 2010 to 6025 in 2015.1 As of 2015, there are approximately 35,000 registered childminders. Therefore, while childminders are greater in number, childcare centers serve approximately 10 times the number of children.

The costs of childcare are subsidized regardless of whether an individual uses a childcare center or child minder. The average cost of childcare is around e1800 per month for care of five days a week. The subsidies are highly dependent upon the joint income of both parents with the average subsidy covering around 40–60% of childcare costs. Net Childcare costs represent around 10.1% of family net income compared with the OECD average of 11.8% and contrasting with 26.6 and 23.1% found in the UK and US respectively. Given this, the childcare facilities in the Netherlands are relatively cheap and flexible in supporting working parents. However, see Chapter 13 by Maldonado and Nieuwenhuis in this Volume to have similar numbers differentiated by twoparent and single-parent families. An additional feature of the Netherlands childcare market however is the range of choice available to parents in the local area. The average distance to a childcare center is now less than 3 km and the average distance to a child minder is just 250 m. This means that the average Dutch household is just a 4-minute drive from their nearest childcare center or about a 10-minute cycle.

It is possible that this diverse range of options open to Dutch Parents facilitates their work-life balance and enables greater female employment. It should be noted that the childcare provision referred to here is purely with reference to children aged 0–4. At the age of 4, children are eligible to attend the public school system. However, given that public school opening hours are more restricted than those of childcare facilities, there is a need for childcare beyond the age of 4. This is addressed through the provision of after school care (*Buitenschoolse Opvang*, BSO). BSO facilities are sometimes integrated within childcare centers but can also be provided by the local school or a

<sup>1</sup>www.landelijkregisterkinderopvang.nl.

further organization. Therefore, if a parent has children aged 0–4 and children aged 4+, it is necessary to ensure that all children have care provision for the whole day.

## **Hypotheses**

In this paper we seek to move beyond a cost-benefit analysis of childcare and female employment and instead adopt a geospatial perspective focused on the access to childcare options. Rather than the decision to work being guided by a pocket book analysis of returns to employment versus the costs of childcare, we posit that the decision as to whether to work is largely driven by practical constraints associated with the proximity and flexibility of childcare. The subsidies within Dutch childcare do not eliminate the costs of childcare but they are designed such that work is always fiscally positive (i.e. the cost of childcare per hour is less than the hourly wage). Under this system, marginal returns on working an additional hour are still variable and particularly across different income groups. However, this does suggest that cost-benefit assessments are not the primary constraint on work decisions.

Instead we choose to focus on the availability, proximity, and diversity of childcare options available. This approach assumes that, given the persistent positive returns to working additional hours, women will seek to work 0.4– 08fte. The decision to not work full-time or something approaching full-time is due to constraints on childcare availability to enable this, as well as the availability of parental leave. Women faced with the decision of how much to work will first seek to make use of their parental leave entitlements. Spread over four years, these entitlements roughly translate to one day per week in which a parent can stay home and care for the child without incurring a significant financial loss. Beyond this, we assume that women will seek to work a 32-hour week and that the primary constraint on this is availability of childcare. Regarding the availability of childcare, there are two ways to view this availability; in proximity to their work or in proximity to their home. We opt for proximity to their home for several reasons. Firstly, while women take on a disproportionate responsibility for childcare, most couples seek a childcare placement near their home rather than place of employment as it at least makes practical the possibility that the male partner can also drop off and pick up the child (Doan & Quadlin, 2019). Secondly, we assume that couples tend to opt for childcare close to their home than work as a move of address is less likely than a move in employment and couples will

seek to limit disruption to their childcare arrangements (Langford, Higgs, & Dallimore, 2019). Finally, the assumption of a fixed place of work is often misplaced, making it hard to determine an appropriate childhood location. This is borne out by the NKPS data which provides vague or missing data on the location of many places of employment.

This paper tests the theory that geospatial availability of childcare relative to a women's residence determines their employment hours by using the following three hypotheses:

*H1 The closer a woman is to a childcare facility (either childcare center or childminder), the more hours she will work each week.*

This hypothesis examines the straight forward idea that physical proximity enables parents and women specifically to better coordinate work and family life and ensure that they are able to work more.

*H2 The more childcare facilities a woman has in 1 km radius, the more hours she will work each week.*

This hypothesis examines whether it is density and not merely proximity which aides employment by increasing the number of options available to a woman in negotiating her work-life balance.

*H3 The more types of childcare facilities (either childcare center or childminder) a woman has in a 1 km radius, the more hours she will work each week.*

This expands on the second hypothesis by suggesting it is not the density of childcare facilities but also the range. If true, an individual with many childcare centers and childminders in their immediate proximity would be the most likely to work longer hours as they can utilize both forms of childcare to best match their needs. Throughout the analysis we also test for the types of childcare independently of each other, recognizing that they support maternal employment in distinct ways. This analysis therefore seeks to gage the impact of the expansion in Dutch Childcare since 2005 at the sub-regional level and explore the geospatial determinants of female employment that it potentially implies.

# **Data & Methods**

### **The Netherlands Kinship Panel Study (NKPS)**

The data in this analysis are provided by the Netherlands Kinship Panel Study (www.nkps.nl). The NKPS is a panel study on family dynamics in the Netherlands that as of 2014 has collected four waves of data on individuals and their kin. The NKPS covers the population aged 18–79 in the baseline year of 2005 and who were resident in the Netherlands. The research questions revolve around the theme of solidarity, which is defined as 'feelings of mutual affinity in family relationships', and how these are expressed in behavioral terms. The NKPS has four special features that make it highly innovative: it is large (*N* = 9500 at Wave 1), it is a panel (prospective longitudinal design), it is multi-method (the data collection involves both structured interviews and in-depth open interviews), and it is multi-actor (the data are from individual respondents as well as from family members).

The sample for this analysis was limited to women of childbearing age (18–40 in 2005) who were then followed over the next nine years and reinterviewed every three years (2008, 2011, 2014). From these interviews, it is possible to follow employment and fertility histories of these women over the 9 years covered by the study. Only women who responded in all four waves of the study were included in the analysis yielding a total analytic sample of 604 women and 2416 observations. The attrition in the study was 68.36% between wave 1 and wave 4 which indicates high potential for selectivity bias within the analysis. For example, in the target sample at wave 1 the percentage of individuals with higher education is 36% but once those who do not participate in all four waves are excluded, this rises to 44% indicating higher attrition among lower educated groups.

The main dependent variable within the study is the number of hours worked per week in the last 6 months. This is reported in each wave of the NKPS and refers to the 6 months prior to interview. The average is 23 hours per week, inclusive of those not working. This rises to 27 hours per week when those who do not work are excluded. In addition to the independent variables which are detailed as part of the contextual data below, several controls are included below based on previous research. These include the number of children the respondent has, the marital status of the respondent at the time of interview (i.e. single, cohabiting or married), the educational level of the respondent at interview (Low: ISCED 0-2; Medium: ISCED 3- 4; High: ISCED 5-7) and the distance in kilometers to the nearest parent or


**Table 20.1** Descriptive statistics

parent-in-law at the time of interview. Descriptive statistics are presented in Table 20.1.

The NKPS provides highly detailed data on the respondents' geography including geo-coding of their address, their parents' address, their work address, their partners work address, and their closest friends address. This ensures that respondents can be placed within a highly accurate geographical context in relation to other geographic sources of data in the Netherlands. To illustrate the degree of detail this provides, Fig. 20.1 shows respondents in the city of Den Haag (population: 510,909). The gray lines represent the demarcations of neighborhoods (*buurten*).

#### **Contextual Data**

To enrich the data provided by the NKPS, the *Landelijk Register Kinderopvang en Peuterspeelzalen* (Register for Childcare and Toddler Groups) was used. This register is the official register for all childcare providers in the Netherlands and is publicly accessible at (www.landelijkregisterkinderopva ng.nl). The postal codes of all facilities are provided as part of the register and can therefore be geolocated to their exact location. Upon request this dataset can also be accessed in historic form by which dates of registering and deregistering of childcare facilities can be accessed. This therefore provides a record of all childcare facilities and their geographic location for the period of the study between 2005 and 2015, including facilities that open and close between these dates and the months in which they do so. Using this information, it is possible to calculate childcare availability for each respondent in the NKPS via a number of metrics. The register also demarcates childcare facilities by types. This allows for type specific metrics to also be calculated. To demonstrate the richness of this data, Fig. 20.2 shows the childcare centers registered in Den Haag in 2010.

Figure 20.3 also illustrates the shift in the mean distance between a member of our analytical sample and a private day care center over the four waves of the NKPS. The results illustrate that in 2005, when the legislation had only recently come into force and private sector providers were scarce, the average distance to a private childcare provider was around 13 km. Over the years the number of childcare providers has increased considerably, and the expansion has reduced not only the median but also eradicated large outliers resulting in average distance of just 750 m and an observed maximum of just 7 km. These figures show the changing nature of childcare availability over the period of study.

**Fig. 20.2** Nurseries in Den Haag in 2010 (*Source* www.landelijkregisterkinderopva ng.nl)

**Fig. 20.3** Distance to nearest private childcare provider, by wave

In addition to data from the NKPS and childcare registries, we use data from CBS2 on small area measures of economic prosperity and growth that will inevitably shape a woman's labor market activity. This includes both measures of unemployment, average house prices and demographic composition of the local area. These measures are available at the level of *gemeente* (community) which normally consist of between 5 and 10 neighborhoods (*buurten*).

#### **Analytical Framework**

The analysis is multilevel in nature with 2416 observations nested in 604 respondents. These are clustered in communities to produce robust standard errors. The dependent variables were derived from the respondents in the NKPS at the person-wave level and reflected the number of hours worked. The number of hours worked is then estimated using a multilevel tobit model with two levels (i.e. person-month, person). The multilevel tobit structures the analysis such that whether the respondent works is first estimated

<sup>2</sup>Centraal Bureau voor Statistiek/Statistics Netherlands—The Dutch Statistical Office.

(i.e. hours worked > 0) and then the number of hours are estimated using a weighted OLS (Amemiya, 1984).

The independent variables are those associated with the proximity and density of childcare facilities that are derived from the contextual data. These are variable at the level of person-wave given that the facilities in someone's proximity changes over time and individuals move. It is also important to note that the proximity of childcare facilities is independent of the neighborhood and municipality boundaries with all values calculated for each respondent, providing a unique value on all independent variables. Three models are presented as part of the analysis. First the proximity of the closest childcare facility is included in order to ascertain whether it is crude proximity that is associated with female employment. Second, the number of each specific facility within 1 km is used as an independent variable with the proximity variable in the first analysis omitted. This is to measure the impact of density of each form on female employment. Finally, an interaction term is introduced to extend the second model between types of childcare facilities in order to test the possibility that diversity in facilities influences female employment.

# **Results**

The results are presented in Table 20.2, and are broadly in line with findings from the existing literature on female employment patterns (Hank & Kreyenfeld, 2003; Van Ham & Mulder, 2005). Education is positively associated with working more hours per week, with those with higher education working a full 10 hours longer than those with lower education. In terms of marital status, we see that divorced individuals work longer hours than those who have never been married or those who are married. Children have a steep negative impact on the number of hours worked per week. If they do work, those with one child will work around 8 hours (one working day) less per week than those who are childless. Those with two or three children work between 11 and 12 hours less and those with four or more children work around 18 hours less. Controls for wave and age were also included but they did not exhibit significant effects or affect the results of the other variables reported.


**Table 20.2** Results of tobit models 1–3

(continued)



\* = p<0.05' \*\* = p<0.01' \*\*\* = p<0.001

Regarding the first model the results show that the relationship between distance and the number of working hours is negative as hypothesized. For each additional kilometer that the childcare is from the respondents' address, an employed respondent works around 15 minutes less per week. This relationship is strongly significant. The 5 and 95% ranges of the distance to a childcare facility are 0 km and 20 km, respectively, which suggests a maximal effect size of around 5 hours per week. Another way to consider this given the trends observed across the Netherlands during this period is to consider the difference in the average distance to childcare in 2005 and 2014. This fell from 12 km to just 750 m which, according to these results would have resulted in an increase of around 3 hours per week in work for those in employment. Ordinal measures of distance were tested (not shown) but similarly showed a linear relationship between proximity and employment. Given this, the parsimonious linear model was preferred.

The distance to the nearest childminder is not significant. The distance to parents or parents-in-law who might offer an alternative form of childcare was significant and positive, but the effect-size was very small and represents just 30 minutes more work every month for every additional kilometer they are further away. It disappears entirely when the density and variety of childcare facilities are added to the model, emphasizing the limited role that the proximity of grandparental care has in shaping female employment in the Netherlands. The positive association could be related to a selection effect by which individuals who live further away from their parents, do so because they moved for improved employment prospects. Moves made within the period of study are controlled for through the fixed effects design but this doesn't fully capture the potential long-term decisions of individuals to move away to university or urban centers in pursuit of work, and simultaneously areas with better childcare provision.

In model 2 we tested the hypothesis that density was a better estimator of the number of hours worked than simply recording the proximity. The results indicate that the number of hours worked are around 4 hours more when there are 2 or more childcare facilities within 500 m of an individual's home. This may seem like an unrealistic density for childcare facilities, yet in wave 4 of the NKPS 25% of respondents in our analytical sample benefited from such density. This finding holds true even though we are controlling for population density and employment rates at the municipality level and clustering the standard errors.

The number of childminders within 500 m is negatively associated with the number of hours worked, though the standard errors are large and only significant at the 95% level. The reason for this is because 85% of the sample have at least 2 childminders within 500 m and 93% have at least one. This means that the reference category in this group is relatively atypical and small and this may lead to the strange result. The difference between 1 and 2+ is, however, not statistically significant. We recoded and recast this variable in a number of ways with various intervals to reflect the distinct distribution, but none showed a significant and consistent effect in any direction of the density of childminders on working hours.

The third model presents an interaction term between the density of the childminders and the day care centers, but the results were not significant. We also ran an interaction term to see if the effect of day care centers varied across educational groups or parity but there were no significant effects. When looking at the model fit statistics, it appears as though density of day care centers are a marginally better model fit, despite a restricted categorization that is used. This supports the second hypothesis that it is density rather than strict proximity that better explains the number of hours worked by women.

# **Conclusion**

This chapter sought to examine the way in which the geographic availability of childcare at a sub-regional and highly localized level. Since 2005, the Netherlands has seen a rapid expansion in the number and geographic spread of childcare centers and this offered a natural experiment in the shaping of women's decisions to work longer hours. This contrasts strongly with much of the existing literature which has focused on an economic and cost focused analysis of childcares effect on female employment (Rammohan & Whelan, 2006; Viitanen, 2005, 2011). Research on family support and grandparenting has, however, illustrated that availability, proximity, and flexibility are crucial factors in determining childcare efficacy with regards to female employment.

The results provide strong evidence that in the Netherlands, the increased geographic availability of childcare increased the number of hours a woman works. This effect is consistent across educational groups and the results suggest that increased geographic availability could increase a woman's working week by up to 4 hours. This may appear a small effect but would represent 2 additional days' work per month, a 15% increase on the average number of hours worked by women aged 18–35 in 2015. Whether assessing the proximity of childcare or the density of childcare providers, there was a clear positive effect on working hours. From our results, it is also clear that childminders (i.e. small, independent childcare providers) are not positively associated with employed women working longer hours. They were also not seen to affect the efficacy of day care centers by providing diverse choice.

The results here were possible due to highly innovative data collection within the NKPS that allows for analysis of individuals' geospatial context. This perspective is highly valuable in providing greater detail on people's immediate surroundings and differentiating between their lived experiences of policies and the world in which they are embedded. However, the analysis here can be extended in many different directions to offer greater detail and insight into these policy issues at a sub-regional basis. Many social surveys and data facilities are making geospatial analysis possible and this opens up a new range of possibilities for the analysis of policy at a sub-regional level. Secure remote access environments, distributed computing, and the general data protection regulation provide an increasingly practical framework for collecting analyzing this kind of geospatial data.

In this instance, disclosure issues were managed by restricting access to the original data collecting institutes and thus ensuring that it was only used for research purposes (Article 9.2(j) of the GDPR). Outputs such as Fig. 20.1 were then assessed for their disclosure risk which was deemed minimal given that (a) respondents' location was masked by a random error term extending up to 100 m and (b) by not displaying covariates in a geospatial distribution. Social scientists and policy analysts must be increasingly open and engage with geospatial sciences to best exploit the potential of such methodologies and their ability to provide key and direct policy insights. The data used in the analysis here was used to identify several neighborhoods in the Netherlands where childcare accessibility was below that expected of an area with given density and income levels. Identifying such childcare deserts is a potentially important tool for policy makers.

This chapter struggled to fully capture potential confounding and selection effects at a geo-spatial level, and this could be much improved not only through further analysis in this area but also through the development of research in the field of geospatial policy analysis. Having said that, shortterm selection effects during the twelve-year study period are captured and controlled for and only long-term selection effects remain. Furthermore, the main shift in individuals' exposure to childcare facilities came from rapid expansion in childcare provision rather than from personal moves into areas with better childcare. In wave 1, those living in an area within the top 20% of childcare provision would have been in the bottom 20% by wave 4. The variance in the main independent variable is therefore temporal and not associated with any personal decisions or characteristics. It is unlikely that the effects presented here are attributable to a further geographical feature that is correlated with both female working hours and day care centers or that it is a selection effect by which those who want to work longer move to the places with better childcare. To further address longer term selection effects however, it would be necessary to apply an instrumental variable approach or construct a natural experiment but these lay outside the scope of this chapter.

The potential for such analysis with regards to policy is considerable. The more fine-grained results within this analysis allow for the identification of childcare deserts and precise and explicit policy advice to policy makers at a highly localized level. Precise estimates of geo-spatial contexts translate directly into more precise policy interventions and a shorter line between research and policy makers.

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Part V**

**Organizational**

# **21**

# **Company-Level Family Policies: Who Has Access to It and What Are Some of Its Outcomes?**

**Heejung Chung**

Most industrial societies have seen a rise in women taking part in the labor market in the past two decades (Chung & Van der Horst, 2018b). With it, we also observe changes in gender norms in whose role it is to care for children and elderly/disabled family member, and whose role it is to do the breadwinning (Knight & Brinton, 2017). Increasing numbers of men and fathers are voicing their interest in taking a larger part in childcare (Working Families, 2017), and with it, there is a rise in the demand from workers for a better work–life balance and demand for more family-friendly policies at the company level. For example, studies have shown that there are more workers who consider work–life balance as (very) important when considering their next job compared to those who believe other more traditional factors such as higher income is important (Chung, 2017). More recent studies have also shown that many workers place flexible working, one of the most common types of family-friendly arrangement currently used across Europe and the US, as the top benefit they would like in the workplace exceeding in many cases other more financial types of benefits (Franklin, 2019; Scott, 2018). This demand is more prevalent among millennials—i.e., those born between 1983 and 1995 (Deloitte, 2018). Some studies have shown that 4 out of 10 millennials have said they have refused a job due to the lack of flexibility (Franklin, 2019), compared to a quarter for all workers.

H. Chung (B)

University of Kent, Canterbury, UK e-mail: h.chung@kent.ac.uk

This chapter aims to closely examine who has access to company-level family policies and what its outcomes are for the company, the individual, their family, and possibly society as whole. This is done through synthesizing a range of studies that have examined company-level family policies focusing mostly on studies in Europe, but with references to studies that have been carried out in the US and other countries. Although it aims to capture a wide range of family policies, many of the studies examined focus on workingtime flexibility arrangements—a frequented topic of study given the increased demands for such arrangements.

The next section defines what we mean by family-friendly policies at the company level, followed by an examination of some of the trends of family policies using some secondary quantitative data. The next section provides some key summaries of the outcomes of family-friendly policies at the company level. This includes outcomes for the worker, their family, and the company. The final section sums the chapter up, with some final thoughts on the frontiers of research in the field and what needs to be done to develop the field in the future.

# **Defining Family-Friendly Policies and Flexible Working**

Family-friendly policies can be defined as policies that directly support the combination of professional, family, and private life (Plantenga & Remery, 2005). Company-level policies are those introduced or implemented by firms to enhance work–life balance of workers. This does not have to be the firm's independent policy and could involve the implementation of national or sectoral collective agreements or legal regulations. Companies can restrict access to policies that are implemented at the national and sectoral level so that de facto workers are not able to take them up. On the other hand, companies can provide additional policies that do not exist at the national or sectoral levels to help workers balance work with other aspects of life to meet a range of different needs that companies themselves face (Chung, 2012; Chung & Tijdens, 2013). This is why many studies find a discrepancy between the national and company-level practices in relation to family-friendly policies (Den Dulk, 2001; Ollier-Malaterre, 2009), and why sometimes scholars define company-level policies as the "final availability" workers actually have toward various arrangements (Chung & Tijdens, 2013; Lambert & Haley-Lock, 2004). Having said this, sometimes this availability written in company policies is also not a guarantee that workers feel comfortable taking them up (Cooper & Baird, 2015). This is especially the case when there is a culture within the organization and the profession which stigmatizes workers who take up family-friendly working arrangements—i.e., the so-called "flexibility stigma" (Chung, 2018b; Williams, Blair-Loy, & Berdahl, 2013), which can hinder the take-up of arrangements even when they are available.

There are several different types of family-friendly arrangements that are commonly provided by companies, and examined by work–family scholars.

First, there are family-friendly working-time arrangements, or what others call employee-friendly working-time arrangements (Chung & Tijdens, 2013; Rubery & Grimshaw, 2003). This includes arrangements that allow workers to have more control over when they work—i.e., *flexitime* (flexible starting and ending time of work), condensed or compressed working hours (for example, working full-time over four rather than five days), annualized hours (where working time is calculated not over the course of the week but across a longer period of time—up to the whole calendar year), workingtime autonomy (where workers have almost complete control over when and how much they work, as long as the work gets done), or the ability to take a couple of hours off work to tend to personal issues. Family-friendly working-time arrangements also include arrangements that allow changes in the hours worked by the worker, that is in most cases the ability to work less than full-time, albeit sometimes temporarily, to fit workers' needs to balance family or life with work. This includes part-time working (namely, working less than full-time, in some cases defined as working less than 30 hours a week), term-time working (where workers work only during school term times), temporary reduction of hours (where workers work reduced hours for a short period of time), and phased retirement (where workers gradually reduce the number of hours of work before retirement). Although not directly related, many group these family-friendly working-time arrangements with other types of arrangements such as teleworking/home working to brand them as flexible working arrangements or schedule control (Dex & Scheibl, 2001; Glass & Estes, 1997; Kelly, Moen, & Tranby, 2011; Lewis & Humbert, 2010).

Second, there are arrangements provided by companies in which workers take a longer period of time off work to take care of their responsibilities outside of work, such as maternity/paternity, parental, and carer's leave. Companies can either provide additional time off, or provide top-up of benefits given during this period. For example, a large number of companies in Sweden provide additional parental leave pay which tops up the benefit levels set by the national policies (Duvander & Löfgren, 2019). In a broader perspective of work–life balance, these types of arrangements can also include (paid) leave for education, training, and general sabbaticals.

Third, there are services provided by the company. These can be in the form of facilities, such as kindergarten or other childcare amenities including in-house crèches. Or they can be in the form of financial support, for example, for parents using private childcare facilities or other care support services. This can also include support for other types of household work e.g., laundry facilities.

# **Provision of Family-Friendly Arrangements Across Europe**

There is not a lot of cross-nationally comparative data on the extent to which this wide range of arrangements is provided in companies. One of the few existing surveys that cover a wide range of arrangements and is comparable across Europe is the Establishment Survey for Working Time and Work–Life Balance (ESWT). The ESWT covers establishments of 10 or more employees across 21 European countries and was collected in 2004/5 (for more, see Riedmann, Bielenski, Szczurowska, & Wagner, 2006). The more recent version of this survey, the European Company Survey, does provide information on some working-time arrangements—e.g., flexitime provision—yet does not cover information about family-friendly leaves or services provided by the company. This raises a serious issue about the lack of comparable data sets on company-level provisions of family-friendly policies which needs to be addressed. Chapter 22 by Begall and Van der Lippe in this volume details one such innovation.

According to the Establishment Survey for Working Time and Work–Life Balance, 48% of all companies reported providing some sort of flexitime arrangements for their workers in 2004/5. Note that this number has increased to 57% in 2009, and 65% in 2013 according to the European Company Survey data (Chung, 2014). In the 2004/5 survey, 53% of all companies said they provided some sort of leave options, 37% for care of elderly care, ill or disabled relatives, 36% provided leave for further education, and 26% for other purposes (excluding parental leave). Examining the childcare facilities offered by establishments, approximately 3% of all establishments offered an own company kindergarten or crèche services, while 2% offered other forms of childcare help—with larger companies and companies within the service sector more likely to offer such services (Riedmann et al., 2006). There are large variations across countries in the extent to which services and leaves are provided. Chung (2008b) examines the cross-national variation in the diversity of arrangements provided at the establishment levels using the following operationalization based on the ESWT data:


As shown in Fig. 21.1, leaves and working-time arrangements were the most commonly provided family-friendly arrangements while not many establishments provide services to their workers. Examining the cross-national variation of the company-level policies, we see that the Northern European countries Finland, Sweden, and Denmark, which are typically known to have generous national-level family policies (Bettio & Plantenga, 2004; Ferragina & Seeleib-Kaiser, 2015; Korpi, Ferrarini, & Englund, 2013; Plantenga & Remery, 2009) are the forerunners in the provision of family-friendly arrangements. This is also the case in the Netherlands, well known for its flexible

**Fig. 21.1** Cross-country variation in the provision of family-friendly arrangements for 21 European countries (establishment weighted) (*N* = 17,308) (*Source* Chung, 2008b)

labor market (Wilthagen & Tros, 2004). On the other hand, Southern European countries—namely, Portugal, Greece, Cyprus, and Spain—are those where establishments do not offer much in terms of family-friendly arrangements.

When examining more recent data, focusing on the access to flexible working arrangements using individual-level data, similar patterns are observed. Based on the most recent European Working Conditions Survey of 2015, Figs. 21.2, 21.3 and 21.4 show the extent to which workers have access to a number of family-friendly working-time arrangements. Due to lack of data, we are unable to look at workers' perceived access to other types of family-friendly arrangements. Flexitime is defined in the Working Conditions Survey as a worker being able to "adapt their working hours within certain limits," while working-time autonomy is defined as a job "where your working hours are entirely determined by yourself." Time off work for personal reasons include those who have answered "very easy" or "fairly easy" to the question "Would you say that for you arranging to take an hour or two off during working hours to take care of personal or family matters is…". Those who work from home are defined here as those who have worked in their home at least several times a month in the past 12 months, and this group and those who have worked in public spaces at least several times a month in the past 12 months are considered those who teleworked.

Figure 21.2 shows more than a quarter of all dependent employed workers across the 30 European countries, that is the 28 member states plus Switzerland and Norway, have some sort of schedule control. Approximately 20% have access to flexitime and another 6% full working-time autonomy.

**Fig. 21.2** Proportion of dependent employed with schedule control across 30 European countries in 2015 (*Source* EWCS, 2015; Chung, 2019a)

**Fig. 21.3** Proportion of dependent employed who have worked at home or in public spaces several times a month in the past 12 months across 30 European countries in 2015 (*Source* EWCS, 2015; Chung & van der Lippe, 2018)

**Fig. 21.4** Proportion of dependent employees across 30 European countries with access to time off during working hours for personal reasons in 2015 (*Source* EWCS, 2015, author's calculations)

There is a clear pattern here, again with Northern European countries such as Sweden, Denmark, Finland, and the Netherlands—being the ones where flexitime and working-time autonomy is prevalent. On the other hand, Eastern and Southern European countries, which also have limited family-friendly policies at the national level, are the ones where access to family-friendly working-time arrangements are also limited. Similar patterns emerge when examining the patterns of teleworking across Europe, wherein the Northern European countries workers are more likely to have worked at home or in other public spaces regularly in the past 12 months. On the other hand, Southern and Eastern European countries are those where such patterns of work are not as prevalent. Actually, previous studies have shown that in countries where national family policies are generous, workers are also more likely to have access to company-level family-friendly working-time arrangements as well (Chung, 2018a, 2019a; Den Dulk, Groeneveld, Ollier-Malaterre, & Valcour, 2013). This pattern is not as clear-cut in the case for time off work for personal reasons, as shown in Fig. 21.4. First of all, it is noticeable how widespread this arrangement is in comparison to the other types of arrangements, with 61% of all dependent employees noting that they are able to take a couple of hours off work to tend to personal/family issues. Although the Netherlands and Belgium remain in the top group of countries, where workers note that these arrangements are accessible, other countries with the highest level of provision include Romania, Malta, and Italy, all of which were not forerunners in terms of the provision of other types of familyfriendly working-time arrangements. This could possibly be due to the fact that in the case where workers have greater flexibility in their schedule, they can try to tend to family and other personal issues outside of working hours. On the other hand, those who are more fixed to a 9–5 schedule within the office, may have to resort to taking a couple of hours off work during working hours to manage their personal/family issue. Further investigation is needed.

We expect that there will be an increase in the use of family-friendly working-time arrangements in Europe in the future due to the mass scale home working that was encouraged or enforced during the COVID-19 lockdown periods across all countries (Chung et al., 2020). In addition, there is likely to be a rise in flexible working due to the new European Directive on Work–Life Balance that has been passed by the European parliament in 2019.1 The new directive includes the right to request flexible working for parents of children and workers with care responsibilities, which provides workers across Europe a stronger right to access flexible working arrangements—namely, what we discussed above as family-friendly working-time arrangements and workers' ability to work from home. Although the directive aims to influence national-level legislations, family-friendly working-time arrangements are de facto provided at the company level. In this sense, wide spread of national-level legislation that provides workers the right to request flexible working is expected to shape company-level provisions and accordingly workers' access to these arrangements. Given that many European

<sup>1</sup>See this link: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52017PC0253.

countries—such as the UK, the Netherlands, Finland, and Italy—already have some legal provision similar to this already, the larger changes will occur in other countries where no such rights exist at the moment. However, more research is needed to investigate this further.

# **Who Has Access to Family-Friendly Arrangements?**

## **National-Level Determinants**<sup>2</sup>

In this section, we turn to national-level determinants of access to familyfriendly policies—again focusing more specifically on flexible working arrangements. Table 21.1 reviews 18 existing studies that examine the use of family-friendly policies, or flexible working arrangements in a cross-national perspective. From the table we can see that, industrial relations and power resources of unions, cultural factors including national norms on gender issues and work orientation, the institutional factors, i.e., family policies, national-level demand, measured through women's labor market participation, and economic conditions and structures, i.e., affluence of the country, economic labor market condition, the composition of the economy, are all relevant factors we can consider when examining the national-level factors that explain who provides more family-friendly arrangements.

#### **Family and Social Policy**

Perhaps one of the most widely examined and one of the most interesting factors for scholars is the influence of national-level family- and other social policies in the provision of family-friendly arrangements at the company level. There are two theoretical assumptions held in examining the relationship between national-level policies and provision of (additional) family-friendly policies by the company. Firstly, "crowding out" theory (Etzioni, 1995)—usually used to examine the relationship between welfare states and social capital—argues that generous national-level social policy programs "'crowd out' informal caring relations and social networks, as well as familial, communal and occupational systems of self-help and reciprocity"

<sup>2</sup>This section is adapted from Chung, H. (2014) Explaining the provision of flexitime in companies across Europe (in the pre- and post-crisis Europe): role of national contexts. WAF working paper 1. Canterbury: University of Kent.




**Table 21.1**

(continued)



**Table 21.1**

(continued)



**Table**


(Van Oorschot & Arts, 2005, p. 6). Based on this theory, countries where generous family policies exist at the national level, companies will not be willing to or may not feel a need to provide occupational policies to address similar issues. The counterargument to this comes from the "crowding in" theory (e.g., Künemund & Rein, 1999; Van Oorschot & Arts, 2005), which argues that it is rather the countries with generous family policies that usually have companies that also provide more and better family-friendly policies at the company level. The theoretical argument is similar to that of institutional theorists, who argue that institutions, laws, and policies may put pressure on organizations to become similar to national institutions (DiMaggio & Powell, 1983). Den Dulk et al. (2013) argue that governments put institutional pressure on organizations to develop work-life arrangements through coercive powers. Work–life-balance-related policy regulations that enforce provision and tax incentives for such policies directly influence company behaviors in these matters. The pressure can also take the form of normative isomorphic pressure, i.e., national-level policies changing the norm and subsequent public demand for companies to be more family-friendly (Den Dulk et al., 2013), or mimetic pressure, i.e., where companies imitate or mimic the practices of other (successful) organizations (Been et al., 2017; Davis & Kalleberg, 2006). Institutional theory argues that institutions and bureaucratic systems, laws, and policies put pressure on organizations to become similar through isomorphic processes. Based on this line of reasoning, we can expect company-level family-friendly policies to be more generous and widespread in countries where there are generous family policies.

Previous studies provide evidence for both crowding in and crowding out. There is evidence that show in countries where there aren't many statutory regulations on family policies, companies use family-friendly policies as retention or other strategic goals (Den Dulk, 2001, 2005; Ollier-Malaterre, 2009) and thus can be more generous. Others argue that there is no clear relationship between statutory regulations and (extra) company provision (Kassinis & Stavrou, 2013; Präg & Mills, 2014), and only when there is a very large involvement from the state, a crowding-out impact can be seen (Evans, 2002). However, increasingly there is more evidence that countries with generous family policies at the national level are those where companies also tend to be more active in providing family-friendly arrangements (e.g., Been et al., 2017; Den Dulk, Peters, & Poutsma, 2012; Den Dulk et al., 2013; Lyness, Gornick, Stone, & Grotto, 2012). More recently, Chung (2018a, 2019a) argues that the type of policy in question matters in examining the relationship between national-level family policies and the provision of family-friendly arrangements at the company level. There are different associations between the provision of company-level family policies with nationallevel work-reducing policies (leaves) against "work-facilitating" measures (Misra, Budig, & Boeckmann, 2011). Work-facilitating policies—the extent to which the state encourages women's labor market participation/dual-earner system, for example, through public childcare provisions—are positively associated with ("crowd in") access to family-friendly working-time policies (see also, Chung, 2009; Den Dulk et al., 2013; Lyness et al., 2012). In contrast, work-reducing policies "crowd in" only to a certain degree and then "crowd out," similar to what was found for women's employment patterns (see, Misra et al., 2011). Finally, scholars have shown that the crowding in/out may be different depending on the types of companies examined—e.g., public vs private sectors (Den Dulk et al., 2013) and types of workers examined—e.g., high- vs low-skilled workers. Chung (2018a, 2019a) shows how the crowding in of national-level policies are especially stronger for high-skilled workers.

#### **Industrial Relations**

Industrial relations at the national level have also been seen to have a major influence on the choices managers/companies make in the provision of family policies, and providing workers with control over work. According to the power resource theory, welfare states are shaped by the power that is mobilized by the wage earners, may it be through political parties or through interest organizations such as labor unions (Korpi, 1989). In addition to the direct impact trade unions may have on shaping national policies, when there are strong unions within the company and at the national level, this will lead to a "contagion from the left" (Korpi, 1989, p. 316) influencing the way employers act in providing family-friendly arrangements at the company level. In addition, in the Varieties of Capitalism literature (Hall & Soskice, 2001), it has been argued that different institutional structures—including industrial relations structures—impact the behaviors of employers in choosing their competitive strategy. Thus, centralized negotiating structures and platforms will help employee representatives negotiate family-friendly arrangements with employers, but also change the way employers behave in choosing their strategies for competition—taking more of a high-performance route. In sum, strength of the trade union, as well as the collective bargaining structures are likely to impact the way companies behave in providing workers with flexitime. Studies have also shown that collective bargaining coverage rates and union density is positively correlated to the use/provision of flexible working arrangements (Berg, Appelbaum, Bailey, & Kalleberg, 2004; Chung, 2009, 2018a; Lyness et al., 2012; Präg & Mills, 2014).

## **Demands/Culture**

It can be expected that countries with a higher proportion of women in the labor market will be those where there are larger demands for familyfriendly policies at the company level (Ortega, 2009). This is similar to what is expected at the company level, as will be discussed below. A larger proportion of women in the labor market is expected to change the work culture within organizations to be more family-friendly, because of more demands throughout the labor market regardless of the number of women working in that specific company. Empirical evidence supports this, and use of familyfriendly and flexible working arrangements have been shown to be positively related to female labor market participation rates (Chung, 2009, 2014; Ortega, 2009; Präg & Mills, 2014), although others have shown that there are no significant relationships once affluence of the country is taken into account (Lyness et al., 2012). Similarly, normative views on women's role in the market and household may also influence the way employers provide flexible work arrangements. In countries where gender norms are positive toward women and especially mothers working, there may be more demand from workers toward employers to provide family-friendly arrangements (Kassinis & Stavrou, 2013; Lyness & Judiesch, 2008).

Using Mincer's (1962) theory of the relationship between affluence and people's preference toward leisure over paid work time, Präg and Mills (2014) argue that greater affluence of a country will influence worker's willingness to work fixed hours. In fact, GDP per capita has been positively linked to the use of flexible working arrangements (Chung, 2014; Lyness et al., 2012; Präg & Mills, 2014), although it has been examined only through individual-level data thus far. Similarly, some studies (Chung, 2014; Den Dulk et al., 2013) directly examine the work culture of the country—namely, work centrality, to see how it can change the company's provision of flexible working arrangements. It is assumed that in cultures where work is more central to one's life people are likely to work longer, and companies are not likely to provide various flexible working arrangements. Work centrality cultures have shown to reduce the use of family-friendly arrangements, including flexitime, working-time banking, grouped with right to part-time work and right to reduce working hours. In fact, it has been shown to be one of the most important factors explaining the company's provision of flexible working arrangements when examining company-level data (Chung, 2014).

#### **Economic Conditions/Structures**

When the economy is in a stain and there is greater labor supply than demand, this may decrease workers' negotiation power in use of familyfriendly arrangements. On the other hand, when there is greater demand than supply, employers may use family-friendly arrangements as incentives to help recruit and maintain workers (Aryee, Luk, & Stone, 1998; Batt & Valcour, 2003; Chung, 2009; Den Dulk et al., 2013). Prevalence of service sectors and public sectors have also been examined to see the how the structure of the economy as a whole has an influence on individual companies through the diffusion of practices (Chung, 2009, 2014; Lyness et al., 2012; Präg & Mills, 2014). Service sectors and public sectors are more likely to adapt to flexible working arrangements (see the next section). It is thus hypothesized that when these sectors dominate the economy, this may change the work practices of the whole country—thus diffusion of work practices across sectors. The prevalence of the service sector can also be linked to the theory of deindustrialization. Deindustrialization, that is the increase of service sector employment in the economy, has been linked to changes in labor market regulations, public sector employment, as well as general changes in the market structure (Esping-Andersen, 1999; Iversen & Cusack, 2000). Results are mixed and can be found in Table 21.1.

#### **Company-Level Determinants**

In this section, we will go into greater detail about who has access to familyfriendly arrangements, specifically focusing on flexible working arrangements—namely, family-friendly working-time arrangements and the ability to work from home. Here, I will use the term flexible working arrangements to discuss this. Before moving on, we need to discuss the different factors that can shape the company's capacity or willingness to provide flexible working arrangements. Unlike statutory policies, where worker's access to nationallevel family-friendly policies is guided by law, and limiting access may come with legal consequences, provision of occupational-level family-friendly policies will largely depend on employers. Many academics (e.g., Dex & Smith, 2002; Seeleib-Kaiser & Fleckenstein, 2009; Wiß, 2017) distinguish between structural and agency factors in explaining which companies provide flexible working or broader family-friendly policies. Structural factors are factors that prohibit or enable companies to provide flexible working and other family-friendly arrangements. For example, company size and sector are some key structural factors. Due to the administrative costs that are involved in providing these arrangements, larger companies may find it easier to administer and may have more resources to provide it. Having said that, smalland medium-sized companies may be able to provide more informal or ad hoc arrangements (Dex & Scheibl, 2001). The type of work that is being done has always been noted as one of the biggest constraints to the introduction flexible work arrangements by managers (Van Wanrooy et al., 2013). There are jobs where it is harder to apply flexible working arrangements than in others due to, for example, production structure (machinery, clients demand, etc.) or sensitivities toward certain business cycles. This would mean that certain jobs in sectors such as manufacturing, construction, education, retail, and health and social services may be restricted in their application of flexible working arrangements. Public sector employers, on the other hand, have been seen to be better at providing flexible working and other types of family-friendly arrangements because they are not as sensitive to business cycles (Evans, 2001).

Agency factors pertain more to the willingness of managers and/or the push they get from workers to provide family-friendly/flexible working policies. Agency factors include a range of factors including the composition of workers as well as the existence of (strong) unions, and/or characteristics of managers. For example, scholars have noted that theoretically more women in the company would mean that there will be a higher demand for, and thus higher prevalence of, family-friendly arrangements within that company (Goodstein, 1994). However, empirically, at least in the case of flexible working arrangements, this is not the case (Adler, 1993; Chung, 2019b; Glass & Estes, 1997). This may be because employers are more reluctant to trust women, especially mothers, to privilege work above care/housework (Williams et al., 2013), and believe that women may abuse their ability to work flexibly to essentially do less work. Based on the power resource theory (Korpi, 1989) powerful unions may drive employers to provide schedule control to their workers as a part of their efforts to improve working conditions. What is more, organized labor within the establishment might allow for the introduction of family-friendly policies that managers would not have adopted (Seeleib-Kaiser & Fleckenstein, 2009). In this case, unionized workplaces with employee representatives should be the ones where family-friendly flexible work arrangements will be most prevalent. Empirically, however, the results are rather mixed—some saying that unions matter in the provision and access to flexible working arrangements (e.g., Berg, Kossek, Misra, & Belman, 2014; Seeleib-Kaiser & Fleckenstein, 2009), others noting that there is no significant effect (e.g., Chung, 2018a, 2019a), and some noting that this depends on the country (e.g., Wiß, 2017). On the other hand, some studies argue that rather than unions, managers are important in the introduction of family policies at the company level. For example, companies with supportive managers will be more likely to provide workers with familyfriendly flexible work arrangements (Hammer, Kossek, Yragui, Bodner, & Hanson, 2009; Kossek, Hammer, Kelly, & Moen, 2014; Minnotte, Cook, & Minnotte, 2010) and are places where workers feel like they are more able to take up the arrangements (Cooper & Baird, 2015). Some studies argue that female managers are more likely to provide family-friendly arrangements to their workers (Galinsky & Bond, 1998; Ingram & Simons, 1995), however, recent studies have shown no significant association between having a female manager and workers' access to flexible and other types of family-friendly arrangements (Chung, 2018a, 2019a).

#### **Individual-Level Determinants**

Now we look more closely at individuals' access to family-friendly arrangements, again specifically focusing on flexible working arrangements. To better understand what can explain who has access to flexible working arrangements, we need to understand the dual nature of flexible working arrangements. Flexible working arrangements are not only used to meet the demands of workers—in particular working parents within the company but also used to enhance performance outcomes of the company (Brescoll, Glass, & Sedlovskaya, 2013; Den Dulk et al., 2013; Ortega, 2009; Osterman, 1995). High-performance or high-involvement strategy scholars argue that when workers have more control or discretion over their work, this will increase their performance outcomes (Appelbaum, Bailey, Berg, Kalleberg, & Bailey, 2000; Davis & Kalleberg, 2006). Flexible working can be seen as a part of this high-performance strategy specifically aimed at enhancing the performance and productivity of workers.

When understanding the dual nature of flexible and other family-friendly working arrangements (Rapoport, Bailyn, Fletcher, & Pruitt, 2002), we can think of three distinctive principles employers can use to decide who gets access to family-friendly/flexible working arrangements; namely, principle of need, equity, and equality (see also, Lambert & Haley-Lock, 2004; Swanberg, Pitt-Catsouphes, & Drescher-Burke, 2005). When employers are genuinely interested in addressing the work–family needs of workers, those with the most family demands or most need of family-friendly arrangements are likely to request and use flexible work arrangements (Golden, 2009). In addition, companies with workers with more family responsibilities are likely to face a higher demand to provide family-friendly arrangements (Goodstein, 1994), explaining why some studies—especially looking at company-level data, and manager's perceived provision—have linked the proportion of female workers in a company to the likelihood of the company providing flexible working arrangements (Bardoel, Moss, Smyrnios, & Tharenou, 1999; Dex & Smith, 2002; Kerkhofs, Chung, & Ester, 2008; Wood, De Menezes, & Lasaosa, 2003). However, other studies—especially when looking at individual-level data and workers' perceived access—have shown that unlike expectation, female-dominated workplaces are where workers are less likely to access family-friendly flexible working arrangements (Adler, 1993; Chung, 2019b; Glass, 1990; Glass & Finley, 2002). This may be because rather than responding to the demands for flexible working, employers are more interested in the enhanced performance/outcomes gained from introducing the arrangements—i.e., the so-called principle of equity. When employers' motivation for providing flexible working arrangements are driven by the principles of equity, companies will provide these arrangements only to workers they can reap benefit out of - e.g., workers who managers think will work harder or will increase their productivity when working flexibly. In this case, we can expect it to be used more in knowledge-intensive fields (Brescoll et al., 2013) and provided to workers with higher occupational statuses/skills levels in expectation that it will enhance their productivity. This is why many studies have shown that high-skilled workers and workers in higher occupational groups are more likely to gain access to family-friendly arrangements (Chung, 2019a; Gerstel & Clawson, 2014; Glass, 1990; Golden, 2009, 2001; Kelly & Kalev, 2006; Nagar, 2002; Ortega, 2009; Wiß, 2017). Some scholars (Adler, 1993; Schieman, Milkie, & Galvin, 2009) also argue that especially flexible working, where workers gain more control over when and where they carry out their work, is given to higher status workers—again those who are valued in the organization and most likely higher skilled, and possibly in a better bargaining position. On the other hand, workers in disadvantaged positions—e.g., low wage, low-skilled, lower educated—are least likely to have such access (e.g., Golden, 2009; Swanberg et al., 2005; Wiß, 2017). Chung (2018a) examines the degree of access "outsiders" (Schwander & Häusermann, 2013)—workers in disadvantaged/weak positions within the labor market—have to family-friendly/flexible working-time arrangements across Europe. What she finds is that although fixed-term contract status does not influence one's access to flexible working arrangements (unlike what was found in previous studies, Präg & Mills, 2014), low skilled and those who perceive their jobs to be insecure were significantly less likely to feel that they had access to flexible working arrangements. She concludes that workers' relative bargaining power may be highly relevant in explaining one's access to family-friendly/flexible working arrangements.

Lastly, scholars (Lambert & Haley-Lock, 2004; Swanberg et al., 2005) also argue that some companies may implement the principle of equality when providing family-friendly arrangements. In this case, access to arrangements will be provided to all workers equally, regardless of their care demands or potential performance outcome. There is no evidence of this based on empirical studies.

# **Outcomes of Family-Friendly Arrangements/Flexible Working**

#### **Performance Outcomes**

There is a wealth of studies that have been done around the so-called "business case" for flexible working and family-friendly arrangements and performance outcomes (for an overview, see Beauregard & Henry, 2009; De Menezes & Kelliher, 2011; Kelliher & De Menezes, 2019). To sum these studies up, flexible working arrangements and other types of family-friendly arrangements have been shown to have positive links to increasing workers' organizational commitment, job satisfaction, loyalty, and reduced turnover intention (see also, Masuda et al., 2012; Moen et al., 2017; Ruppanner, Lee, & Huffman, 2018). In turn, the provision of these arrangements is linked to increased worker retention, and reduced worker recruitment problems (Aryee et al., 1998; Kerkhofs et al., 2008; Kossek & Ollier-Malaterre, 2019). In addition, flexible working has been linked to reduced sickness, absenteeism, health, and other undesirable well-being outcomes (see also, Avendano & Panico, 2017; Moen et al., 2016). Some studies have also tried to link worker productivity/organizational performance directly with familyfriendly arrangements, for example, such as profit and return on investment, labor productivity, etc. (Chung, 2009; De Menezes & Kelliher, 2011).

#### **Work–Life Balance**

In relation to work–family conflict and work–life balance the evidence is mixed—especially when examining the relationship between work–family conflict and flexible working arrangements. Although some studies show that flexible working reduces work–family conflict for workers (Kelly et al., 2014), others show that the impact is rather minimal (Allen, Johnson, Kiburz, & Shockley, 2013; Michel, Kotrba, Mitchelson, Clark, & Baltes, 2011). Some argue that, especially working from home, may actually increase work–family conflict (Chung, 2017; Duxbury, Higgins, & Lee, 1994; Golden, Veiga, & Simsek, 2006). This may largely depend on the organizational contexts (Van der Lippe & Lippényi, 2018) and national contexts (Lott, 2015). In other words, in more family-friendly contexts flexible working arrangements are more likely to lead to better work–life balance outcomes.

One main reason why flexible working arrangements do not improve work–life balance of workers, or even increase work–family conflict, is because flexible working can lead to workers working longer overtime in paid work or to work spilling over to family spheres. A number of recent studies have shown how flexible working can result in workers working harder and/or longer hours, in many cases (unpaid) overtime hours (Bathini & Kandathil, 2019; Chung & Van der Horst, 2018a; Glass & Noonan, 2016; Kelliher & Anderson, 2010; Lott & Chung, 2016; Schieman & Young, 2010). Other studies have shown that flexible working can lead to mental spillover of work, of workers worrying or thinking about work when not at work (Lott, 2018).

#### **Flexible Working and Gender Inequality**

The extent to which flexible working leads to increased working hours/work intensity is not the same for men and women. Men are more likely to increase their (unpaid) overtime hours when working flexibly (Chung & Van der Horst, 2018a; Glass & Noonan, 2016; Lott & Chung, 2016). This is largely due to the social normative views about gender roles between heterosexual couples. Although there are some changes, men still do and are expected to take on the breadwinning role especially after childbirth (Knight & Brinton, 2017; Miani & Hoorens, 2014; Scott & Clery, 2013) and women are expected to—and actually do—carry out the bulk of caregiving as well as housework (Bianchi, Sayer, Milkie, & Robinson, 2012; Dotti Sani & Treas, 2016; Hochschild & Machung, 1989; Hook, 2006; Scott & Clery, 2013). Such gendered divisions of labor and social normative views about mothers' and fathers' roles shape the outcomes of flexible working (Chung & Van der Lippe, 2018). While men increase their working hours, women on the other hand increase their childcare/housework hours when working flexibly (Hilbrecht, Shaw, Johnson, & Andrey, 2013; Kim, 2018; Kurowska, 2018; Lott, 2019; Radcliffe & Cassell, 2014; Sullivan & Lewis, 2001). Clawson and Gerstel (2014) argues that, in this way, flexible working allows workers—especially middle-class workers—to "do gender" (West & Zimmerman, 1987). In other words, flexible working may "allow" workers to adhere to the social normative gender roles prescribed within societies, thus traditionalizing gender roles (Chung & Van der Lippe, 2018). However, it should be noted that this gendered outcome of flexible working is not inevitable. Kurowska (2018), in her analysis comparing Sweden and Poland, shows how in countries where more egalitarian gender norms prevail, men and women may have more similar outcomes when working flexibly. Again, the gender context matters.

Such gendered outcomes of flexible working also impact people's perceptions toward flexible workers. For example, qualitative studies have shown that when women take up flexible working, for example, working from home, those around them expect women to carry out domestic work simultaneously while working (Hilbrecht et al., 2013; Shaw, Andrey, & Johnson, 2003; Sullivan & Lewis, 2001). This then feeds into the extent to which workers are likely to gain access to flexible working arrangements. Studies have shown that women, especially mothers, are less likely to gain access to flexible working arrangements, even when not used for care purposes (Brescoll et al., 2013; Munsch, 2016). This can explain why flexible working arrangements that provide workers more control over their work are less likely to be provided in female-dominated workplaces—see sections above. What is more, such preconceived notions of where workers' priorities lie and how they will use flexible working arrangements will naturally shape what the consequences of flexible working for one's career. Lott and Chung (2016) show how women are unlikely to gain financial premiums as their male counterparts do, even when they work similar levels of overtime. Williams et al. (2013) speak of the flexibility stigma. This is the stigma and the negative career consequences workers using family-friendly arrangements face, largely due to the fact that such take-up makes them deviate from the ideal worker image. Here an ideal worker is that of a worker who does not have any other responsibilities outside of work, and privileges work above everything else (Acker, 1990; Williams, 1999). Although there is a dispute on whether men may face a double stigma of "femininity stigma" (Rudman & Mescher, 2013) in that they deviate away from the ideal worker image alongside the image of masculine breadwinner roles, evidence suggests that this is not the case (Chung, 2018b; Coltrane, Miller, DeHaan, & Stewart, 2013).

In this sense, flexible working can potentially increase gender inequalities in the labor market, due to the preconceived notion people will make about women's flexible working. However, the picture is much more complex. Several studies have shown that flexible working may allow women to work longer hours than they would have otherwise after childbirth (Chung & Van der Horst, 2018b). In other words, flexible working—especially workers' ability to control when and where they work—may reduce their need to go into part-time jobs. Part-time jobs in most cases entail occupational downgrading (Connolly & Gregory, 2008, 2009), resulting in career penalties/income loss across the life course and considered one of the key causes of the persistent gender pay gap (Costa Dias, Robert, & Parodi, 2018). Thus, giving workers more control to meet family demands may help workers maintain their careers. Similar results have been found by several scholars. Flexible working—again workers' control over their work—has been shown to help women stay in relatively stressful yet high paying occupations (Fuller & Hirsh, 2018), and workplaces with flexible working arrangements are those where the gender wage gap is smaller (Van der Lippe, Van Breeschoten, & Van Hek, 2018). In this sense, we need more evidence to see how these rather conflicting directions of impact of flexible working on the gender pay gap act in the longer term for workers.

# **Conclusion and Future Research Agenda**

As we have seen in this chapter, there has been a great rise in the demands for more family-friendly and specifically flexible working arrangements by workers. Studies have shown that flexible working, and many other familyfriendly arrangements, are not necessarily provided to address work–life balance demands of workers but also used to enhance performance outcomes. This can explain why many have found that it is mostly the high-skilled workers in higher statuses that gain access to these arrangements—a topic also addressed by Begall and Van der Lippe in Chapter 22 in this volume. The chapter also raised issues around the discrepancies between provision stated at the company or state level, versus workers' access to flexible and other family-friendly policies. Such real access to arrangements is shaped by workers' individual (and collective) negotiation/bargaining power, and prevalence of flexibility stigma or the fear of negative career consequences when taking up family-friendly arrangements. More studies need to be done to find out how to ensure that all workers with the demand for more flexible working and other family-friendly arrangements can get access to them, without fear of repercussion.

The chapter further examined the role of national context factors in shaping companies' provision of and workers' access to flexible and other family-friendly arrangements. There is increasing evidence to show that there is a positive rather than a negative relationship between generous nationallevel policies and generous family-friendly policies at the company level/better access to family-friendly policies from the workers' perceptive, pointing to a "crowding in" effect. However, more needs to be examined in terms of the dynamics in which this effect takes place, as well as whether there are variations across different types of family policies as well as across different groups of the population. Further research is needed to examine whether national family policy contexts shape the outcome of flexible and other family-friendly working practices.

Similarly, the chapter summarized some of the key outcomes of flexible and other family-friendly arrangements. As the review has shown the relationship is not as clear-cut as expected, with flexible working leading to worse rather than better work–life balance outcomes, and gender inequality outcomes in some cases. More research needs to be done to examine these relationships further. Especially of interest for many policy makers and company managers will be the impact of flexible working in the longer run for gender equality. The chapter has shown how there are rather conflicting dynamics at play on one hand enabling better access to labor market participation for women, but on the other, enabling or enforcing traditional divisions of labor between men and women. More could be explored in terms of what can be done to ensure to avoid some of these negative unintended consequences of flexible working.

Finally, we need more data to capture the extent to which family-friendly policies are being provided at the company level. As seen from this chapter, the most recent cross-national comparative data comes from 2004, more than a decade and a half from the publication of this chapter. More recent data are presented in Chapter 22 in this volume by Begall and Van der Lippe. A large number of companies provide family-friendly policies above and beyond the national regulations for a number of reasons including skilled worker recruitment and maintenance, as well as to enhance the corporate social responsibility image. On the other hand, we know from case studies that many companies do not even allow workers access to national-level provisions that are supposed to be protected by law, may it be due to lack of knowledge or lack of any bargaining power. Company-level surveys that can capture both managers' and workers' perspectives on a wide range of familyfriendly company-level arrangements, above and beyond flexible working arrangements, are needed to fully understand workers' true access to family policies.

# **References**


of labour compared. *Social Indicators Research*. Online first. https://doi.org/10. 1007/s11205-018-2034-9.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **22**

# **The Educational Gradient in Company-Level Family Policies**

**Katia Begall and Tanja van der Lippe**

This chapter further explores organizational work-family policies, following up on the literature review in Chapter 21 by Chung in this volume. The entrance of women and in particular mothers into the labor force across industrialized countries in the past decades has prompted governments and employers to accommodate the family demands of their workforce by implementing policies such as flexible working times and parental leaves. Not only for female employees but also for men who no longer rely on a homemaker spouse for care tasks, increasingly face work–life balance challenges. In many organizations, work-family policies are part of competitive employee benefits and serve as a way to attract sought-after employees. Indeed, previous research has shown that highly skilled and/or highly paid workers are more likely to have access to and use work-life policies (Chung, 2018; Golden, 2008; Weeden, 2005). Moreover, while women perform the majority of caring and household tasks, they are not more likely to have discretion over their work hours, nor are employees with small children (Golden, 2001; Ortega, 2009). This suggests that work-family policies, rather than being used for the benefit of the part of the workforce that has the strongest need for support

K. Begall (B)

Radboud University, Nijmegen, The Netherlands e-mail: k.begall@maw.ru.nl

T. van der Lippe Utrecht University, Utrecht, The Netherlands e-mail: t.vanderlippe@uu.nl

in reconciling work and private life, for instance, low-income families or single parents, these policies are offered to high-status workers who possess skills valuable to their employers. This relates to issues of social inequality: in most industrialized countries diverging trends between the higher and lower educated in terms of income have occurred since the 1980s. But in addition, social inequalities in terms of life chances more broadly such as health (Präg & Subramanian, 2017), family relationships (McLanahan, 2009), and life satisfaction (Gerdtham & Johannesson, 2001) are on the rise. An important potential source of such disparities which we focus on in this chapter is the access to organizational work-family policies among different groups of workers.

In this chapter, we examine the access to organizational work-family policy, that is whether managers and employees report that a certain policy exists in their organization. We use the terms availability, access, and provision interchangeably. When referring to the educational gradient we also refer to skills or the skill gap. We aim to answer the following research questions: *(1) Are organizations with a higher proportion of highly skilled employees more likely to provide access to organizational work*-*family policies? (2) Are higher skilled employees more likely to report access to organizational work*-*family policies? (3) How is the combined provision of work*-*family policy at organizational and team level related to employee's perceived access to the policy and does this differ by employee skills?*

We use unique multilevel survey data from the European Sustainable Workforce Survey (ESWS from here on, Van der Lippe, Lippényi, Lössbroek, et al., 2016) to investigate the educational gradient in the organizational provision of a variety of work-family policies at different organizational levels. Specifically, our contribution to the knowledge on access to work-family policies is threefold: first, we assess the availability of a variety of workfamily policies at three different levels, enabling us to examine the consistency between policy provision at the organizational (highest) level as compared to the department manager (middle level) and how this relates to employees' perceived access to work-family support (lowest level). We thereby shed more light on how policies are distributed to various groups of workers within organizations. Exploring how employee's perception of availability relates to organizational provision measured independently also illustrates potential difficulties in validly and reliably assessing availability of work-family policies in employee surveys. Secondly, we model the educational gradient in access to the different work-family policies considered while controlling for the main theoretical explanations at employee and organizational level why higher skilled workers would be expected to have more access to these policies. We argue that this "net educational gap" in the availability of workfamily policies reflects an important dimension of social inequality. Thirdly, our multi-outcome, multilevel design encompasses nine European countries, thereby increasing the scope of previous studies.

# **Background and Expectations**

In this chapter, we examine work-family policies provided by organizations. We define these as policies aimed to facilitate employees in combining paid work and family responsibilities which go above and beyond statutory arrangements mandated by national law.<sup>1</sup> The types of policies considered in this chapter follow the conceptualization of organizational work-family policies provided in Chapter 21 by Chung (in this volume) which distinguished three categories. These represent the full range of work-family policies commonly provided by employers to facilitate the combination of work and (family) life (Been, 2015; Plantenga & Remery, 2005) and are (1) working-time arrangements, (2) leaves, and (3) services. As noted by Chung (this volume) there are not many cross-national data sources that provide information on all categories of organizational work-family policies as most surveys focus on flexibility arrangements. Exploiting the richness of the ESWS (Van der Lippe, Lippényi, Lössbroek, et al., 2016) we are able to include in our analysis six different organizational work-family policies covering all three types. These are working from home, flexible start and finish times, and a reduction in work hours (*working*-*time arrangements*), supplemental family leave and the extension of statutory leave with vacation days (*leave*), and childcare assistance in the form of company-based childcare, financial assistance, or mediation in finding a spot (*services*).

We develop expectations about differences in access to organizational work-family policies by employee's skill level and distinguish between policy provision reported at the organizational level by the human resource manager (thus relating to differences between organizations) and perceived availability reported by employees within organizations. For each of the six organizational work-family policies considered, we ask whether an organization with a higher proportion of highly skilled employees is more likely to provide this policy and whether at the employee level the higher skilled employees are more likely to have access to the policy. Moreover, we exploit the fact that

<sup>1</sup>Organizational work-family policies can be part of collective agreements.

provision of work-family policies was reported in our data at the organizational and team level, enabling us to combine these sources of information to present a more fine-grained examination of the combined availability of work-family policy for different groups of workers.

Drawing on the main theoretical mechanisms for why work-family policies are adopted by organizations and for which groups of employees within organizations, as described in Chapter 21 by Chung (this volume), we develop expectations about whether an educational gradient in access to the different policies exists between and within organizations (see Table 22.1 for an overview).

While decreasing work-family conflict may be the primary aim of workfamily policies, the degree to which these measures also serve the interest of employers—for instance, by directly increasing workers' productivity or by enhancing worker commitment and retention—varies, and so do the costs associated with offering these policies. This implies that certain groups of workers which are perceived by employers as more valuable to the organization and/or more likely to show increases in productivity when supported by work-family policies are expected to be more likely to have access to them (Lambert & Haley-Lock, 2004). This argument has been put forward in particular with regard to flexibility policies (Chung, this volume). These policies not only help employees to achieve better work–life balance but can also enhance worker productivity by enabling workers to carry out their work more flexibly. Workers with a high degree of autonomy in carrying out their tasks likely gain most in terms of productivity. As this applies primarily


**Table 22.1** Expected differences by employee's skill level in access to organizational work-family policies by employee skills

to higher skilled white-collar jobs, we expect that an educational gradient exists in the access to flexibility in time and location of work. Following this argument, a higher proportion of highly skilled workers in the organization should also be associated with a higher likelihood of the organization offering these policies. Thus, organizations with a higher proportion of workers in high-skilled positions will be more likely to report that they provide flexibility policies.

With regard to supplemental leave (i.e., family leave offered by the organization which is longer in duration or with a higher income replacement than the statutory national allotment) and childcare assistance, the implications for enhanced performance are less clear-cut. Yet, these policies can be a powerful means of signaling organizational concern for employee's well-being, resulting in increased commitment, performance, and retention on part of the employee (Begall et al., 2020; Kurtessis et al., 2017). The assumed mechanism is that perceived organizational support elicits the norm of reciprocity, leading to a perceived obligation to help the organization and increased feelings loyalty and commitment (Gouldner, 1960; Wheatley, 2017). Because supplemental leave and childcare assistance are relatively costly policies to implement, we expect that employers would adopt these for the parts of their workforce most valuable to the organization in order to invest in their loyalty and commitment. On the other hand, since the use of these policies is less dependent on job characteristics (contrary to flexibility), concerns about social legitimacy may inhibit employers from offering these policies only to selected groups of workers whom they wish to retain. These arguments lead us to expect no educational gradient in access to these policies within organizations. It is important to note that certain categories of employees may nevertheless be formally excluded from using the policy, for instance, those with temporary contracts (Chung, 2018). Because supplemental leave and childcare assistance are costly programs and employers are more likely to employ these strategies to attract and retain workers with highly valued or scarce skills, we expect that organizations with a higher proportion of workers in highly skilled positions will be more likely to report that they provide supplemental leave and childcare assistance.

The remaining two work-family policies we consider—offering a reduction in work hours and using vacation days to extend family leave—can from the employers' perspective be regarded as low-cost options without performance implications and thus can be expected to be made available for everybody to the same extent. As such, they form an interesting point of reference in our comparison. This means that for these measures we expect the lowest degree of within-organization stratification according to education. The same argument can be applied to between-organization differences, where we expect that the provision of these policies is not predicted by the proportion of workers in higher skilled positions.

## **Combined Availability of Work-Family Policies**

Within organizations, policies are filtered by different levels of hierarchy before reaching individual employees. This can occur in the form of formal regulations as certain jobs may not be suitable for the use of certain policies, but this process may also occur informally. Lower-level management can grant access to work-family reconciliation measures not formally adopted by the organization, for instance, through granting workers flexibility in working times informally by allowing teams to plan their own work-rosters (Been, 2015; Lambert & Haley-Lock, 2004). Similarly, lower-level managers may constrain the access to policies by not communicating the existence of the policy to their employees or by informally obstructing the use, as has been frequently reported for men's use of parental leave (Allard, Haas, & Hwang, 2011). Figure 22.1 summarizes the four categories in the combined availability of work-family policies in case of two hierarchical levels. While our survey data on availability of work-family policies does not entail information about the reason or intention for (not) making a policy available at either level, the labels are useful for structuring our thinking about inconsistencies in reports from different sources. Because the multitude of potential mechanisms to be considered across the four categories and six organizational

**Fig. 22.1** Categories of combined availability of work-family policies at organizational and team level

work-family policies considered here we do not formulate explicit expectations, but test for each policy whether the educational gradient in employees' reported access differs across the combined availability at organizational and team level.

### **Alternative Explanations**

A number of alternative explanations at individual and organizational level which may fully or partially account for any educational gradient in access to organizational work-family are included in the empirical test of our expectations. At the individual level the most important predictors of access to work-family policies identified in the literature (see also Chapter 21 by Chung in this volume) are the autonomy in tasks (specifically for flexibility), characteristics of the workers related to relative bargaining power such as tenure, permanent position, occupational status, and work hours, and the actual need for work-family support indicated by workers' sex and the presence of young children in the household.

Alternative explanations for why organizations provide work-family policies relate to pressures of social legitimacy, to which larger and public sector organizations are more susceptible, and internal demand indicated by a higher proportion of women. Since not all types of work are suitable to implement flexibility policies and to account for sectoral agreements which might exist in some countries, we include information about the industry. Moreover, we take into account that work-family policies may serve as a way to attract sought-after staff by including information about whether organizations reported difficulties finding skilled employees. Because flexibility may be part of a high-performance strategy, we include information about the proportion of employees with performance-based pay.

Extending our perspective from between-organization to between-country differences, organizational polices can either extend or supplement national policies or act as a substitute in the case of a lack of national provisions. Institutional theory maintains that organizations' provision of work-family policies is a response to normative pressures from the environment (Beham, Drobniˇc, & Präg, 2014). Previous findings support this notion as it has been found that national provisions of leave, childcare, and work-hour flexibility correlate positively with organizational provisions of these measures (Den Dulk, Groeneveld, Ollier-Malaterre, & Valcour, 2013). In order to account for the large variation in institutional contexts represented in the nine countries in our study (Begall & Van Doorne-Huiskes, 2019), we include country dummies in all models.

# **Data**

We make use of the ESWS (Van der Lippe, Lippényi, Van Lössbroek, et al., 2016), which contains data on 10,673 employees, 726 departments or teams, and 259 organizations in Bulgaria, Finland, Germany, Hungary, the Netherlands, Portugal, Spain, Sweden, and the UK, collected in 2015/2016. Organizations were sampled based on their representation of six different sectors (manufacturing, health care, higher education, transportation, financial services, and telecommunication) and three different sizes (1–99 employees; 100–249; 250 or bigger), using a combination of stratified random sampling and personal connections. After organizations agreed to participate, employees and department managers were addressed at work and asked to participate in an online or paper-and-pencil questionnaire. The human resource manager filled in the questionnaire on behalf of the organization, as is common in this type of research because they are considered to be well-informed about the entire organization (Haas & Hwang, 2016). The response rate was 61.4% among employees, 80.9% among team or department managers, and 98% among the organizations that had agreed to participate.

## **Motivation and Challenges Collecting the** *European Sustainable Workforce Survey*

We encountered numerous challenges during the collection of data. The first challenge was the choice of countries and sectors. The nine countries constitute different types of welfare regimes (Bäck-Wiklund et al., 2011). Although differences between these types are somewhat fluid, Finland and Sweden are typically categorized as social-democratic regimes, Germany and the Netherlands as conservative regimes, Spain and Portugal as Mediterranean regimes, the UK as a liberal regime, and Hungary and Bulgaria as post-communist regimes. The six industries are selected to reflect variation in the causes and types of investments in a sustainable workforce. These six industries vary therefore in the percentage of women working in the sector, the percentage of older employees, flexibility in contracting, and the extent of technological development.

The second challenge was that the unique depth of the ESWS limited the possibilities of drawing a random representative sample from the whole population of organizations. Instead, we used stratified purposeful sampling. First, we used quantitative sampling to randomly select cases within sector and size categories from lists of business organizations. Second, we complemented the random selection with convenience sampling of cases from alternative sources, mainly web searches and referrals. We used this method when the random sample from the business lists did not yield enough participants.

The third challenge was the participation of organizations, as they often conduct their own surveys of employees and see surveys as an extra burden on employees. We targeted the relevant gatekeepers being the HR managers as the actors who are potentially most interested in this research. A team of native research assistants in each country performed a web search and made calls to the establishment to get the contact details of the HR manager. In most cases, we then visited the HR manager to get his or her cooperation. It was thereby important to provide credentials and offer a tangible product in return for cooperation. We offered a benchmark report which compares the establishment with establishments from the same sector and country with regard to workforce investments and key productivity outcomes. Organizations value information about what their competitors are doing and how they perform. In our fieldwork, organizations mentioned seasonal high workload as a vital constraint not to cooperate.

The fourth challenge was participation of employees and managers, albeit this turned out to be a smaller issue than entering the organization. Once an organization joined our research, the response rate was good. In order to minimize individual variation within establishments, we sampled a limited number of employee groups, mostly teams. This also allowed us to survey managers of teams and match them with employees. Managers are an important employee group to study in their own right, and their reports provide extra information about workplace characteristics. A similar strategy, using occupational groups, has also been successfully employed in the Dutch Time Competition Survey (Van der Lippe & Glebbeek, 2003). At least two teams were chosen that represent the organization's core activity: for example, if the organization is a hospital, we interviewed nurses. Furthermore, we chose at least one team whose tasks do not particularly refer to the core activity of the organization, such as finance, communication, or maintenance departments. The rationale was that support teams perform similar activities regardless of the sector, and this provides an opportunity to compare economic sectors while limiting the variability in job characteristics. We used incentives and offered the choice between a lottery among employees who completed the survey and a monetary gift to the organization to spend on organizing activities for staff. Furthermore, we used other well-established strategies to get cooperation from employees and managers, being a personalized cover letter and invitation email where we had access to names, and where we emphasized anonymity and trustworthiness.

The ESWS is unique among organizational surveys. It has a multilevel design, including a large-scale sample of employees and managers within teams and organizations, and it is also cross-country comparative.

## **Measures**

## **Organizational and Team Level**

Availability of work-family policies was measured at the organizational level (reported by the HR manager) and the team or department level (reported by the team manager). Both respondents could indicate whether their establishment (respectively department) offered employees the possibility to "work at home during normal working hours" and use "flexible starting and finishing times (for example, to start working at 7AM one day and at 9AM another day)." The options "switching to working fewer hours a week" or "taking holiday to extend leave period" were reported separately for male and female employees and asked explicitly in the context of availability and use of leave and childcare arrangements. Both policies were coded as available if they are offered to at least one group (men or women) in the organization (department). Supplemental family leave was coded as available if the organization (department) offered either "a longer period of leave than it is obliged to offer by law" or "better-paid leave arrangements than it is obliged to offer by law" for any of the following categories: Longer maternity or paternity leave or longer parental leave for women and/or men. Again, these items were separately assessed but recoded here as available if at least one of the eight options were available. With regard to childcare assistance, financial assistance, childcare at the workplace, and assistance finding or arranging childcare were reported separately but coded as available in this analysis if at least one of the three types of assistance was available. The questions concerning childcare assistance were only reported by the HR manager (organization level).

The proportion of workers in (highly) skilled positions was reported on both levels by the HR manager on a nine-point scale ranging from none (1) to all (9). Furthermore, we take into account whether according to the HR manager the organization had difficulty finding staff for skilled jobs with answers ranging from 1 (very often) to 5 (never). The HR manager also indicated what proportion of nonmanagerial employees receives wage or salary components that depend on their individual performance on a six-point scale ranging from "none" to "most."

Additional organizational level controls are the industry and sector (private vs public) in which the organization is active, the natural logarithm of organizational size and the proportion of women among the employees (1 = none to 9 = all). Descriptive statistics of all variables at the employee level are presented in Table 22.2.

#### **Employee Level**

From the employee questionnaire we obtain employees' education (coded in seven categories ranging from 1 = primary education to 7 = Ph.D.), occupational status coded as International Socio-Economic Index of occupational status (ISEI, Ganzeboom, De Graaf, & Treiman, 1992) and monthly earnings which we coded in country-specific deciles. Job autonomy is measured by four items assessing the freedom of the respondent in deciding the tasks and their order as well as when and how to do the work. Factor and reliability analysis (available upon request) confirmed that these items could be collapsed in one scale of job autonomy on a five-point scale with higher values indicating more autonomy. Moreover, we also include weekly work hours, job tenure in years, whether the respondent has a permanent contract, the presence of children under the age of 18 in the household and the age of the youngest child, the presence of a partner in the household, and respondents' age in years and sex.

#### **Analytical Strategy**

To test our expectations presented in Table 22.1 we estimate two series of models, at the organizational level (*n* = 259) and at the employee level (*n* = 10,673) level. At the organizational level we predict the availability of each work-family policy by the proportion of workers in highly skilled positions (Model 1), and controlling for industry, sector, country, organizational size, the proportion of female workers, the degree of difficulty in finding skilled staff, and the proportion of employees with performance-based pay (Model 2). We use logistic regression models to predict the availability of policies since they are all measured as dichotomous variables (not available = 0, available = 1). Because of the small sample size models were estimated using a penalized maximum likelihood estimator (Rainey & McKaskey, 2015). These models provide an estimate of the skill gap in access to work-family policies *between organizations.*

#### **586 K. Begall and T. van der Lippe**


**Table 22.2** Descriptive statistics of all variables used at employee level (*<sup>N</sup>* <sup>=</sup> 10,673a)

*Note* aValid *N* per item may vary due to missing values *Source* ESWS data

At the employee level we use logistic organization-fixed effects regression models to predict the perceived availability of each work-family policy to estimate the educational gradient *within organizations* (Model 1), and controlling for all individual-level controls (Model 2).

In the second series of models we predict the perceived availability of each policy at the employee level by the combined availability at organizational and department level (see Fig. 22.1) and education as well as individual and organizational controls. Because employees are clustered in teams within organizations, we estimate three-level random intercept models and include a random slope for education. We estimate two models for each policy, a first model with the combined availability at organizational and departmental level, employee education, occupational status, income, and all other individual and organizational controls (see Table 22.2). In the second model we allow for an interaction between the combined availability and employee education to assess the educational gradient separately for each category.

## **Results**

Before turning to discussing the multivariate results, we present the reported availability of policies at organization, team, and employee levels in Fig. 22.2. The availability of policies varies substantially according to type and the level

**Fig. 22.2** Availability of polices at organization, department and employee level (*Note* Childcare assistance is excluded because access was not reported at team level. Taking vacation days to extend leave is excluded because access was not reported at employee level)

of reporting. The most common policies offered by organizations are a reduction of work hours and the use of vacation days to extend leave. More than 80% of the organizations in our sample offer these measures. Since most employees have a legal right to accessing both measures and they are costfree it is not surprising that these are most frequently offered, but since the questionnaire of the ESWS asked about providing these measures specifically in the context of childcare and leave arrangements, we regard them as part of the range of organizational work-family reconciliation measures. Interestingly, employees appear not to be well-informed about the possibility to reduce their work hours, as less than 30% report this as available in their organization. The availability of extending leave with vacation days was not measured at employee level.

Flexibility in the time and location of work is available at the majority of organizations as work-time flexibility is the third most frequently provided measure with 75% of organizations reporting to offer it, and more than half report that their employees can work from home during office hours. It is noteworthy that employees' perceived availability is substantially lower compared to the report by managers. Only 46% report access to flexible work times and 29% to working from home.

The least frequently provided measures are supplemental leave and childcare assistance, less than one-third of organizations provide these options. These policies are most costly and directly benefit the employee more than the employer. For both policies, the provision at the team level (reported by employees' direct supervisor for his/her own team) is approximately 10 percentage points lower than at the organizational level. Among employees, just over 10% state that supplemental leave and childcare assistance are available at their company.

Because the sample of the ESWS is not a random sample, it is not clear in how far these numbers can be interpreted as reliable estimates of the availability of work-family policies in these countries and large-scale data sources assessing different work-family policies to which these numbers could be compared to are scarce. The fact that our numbers are consistent with the findings reported in Chapter 21 by Chung (this volume) is reassuring in this regard.

How the two levels come together in the combined availability (see Fig. 22.1) is presented in Fig. 22.3 as the percentage of employees in each category for the different policies. The bars are sorted by "consistent access," which refers to the situation in which both HR manager and team manager report that access to the policy is provided in their organization.

**Fig. 22.3** Distribution of employees over categories of combined availability of work-family policies at organizational and team level (*Note* Childcare assistance is excluded because it was not reported at team level, *N* = 10,673)

The figure shows that inconsistent categories, which we refer to as restricted and informal access, are relatively most common for working from home and supplemental leave. For both measures, the category of restricted access, i.e., where the policy is reported as available at the organizational but not at the team level, is relatively large compared to the situation of consistent access.

Turning to the multivariate results, Fig. 22.4 and Table 22.3 present the results of the regression model predicting the provision of each policy at the organizational level. The predictor of interest is the proportion of workers in (highly) skilled positions, which we expected to be positively associated with the provision of flexibility policies, supplemental leave, and childcare assistance but not with the possibility to reduce hours or extending leave with vacation days (see Table 22.1). Figure 22.4 shows the exponentiated coefficients (odd's ratios) associated with a one-unit increase in the proportion of staff in (highly) skilled positions on the likelihood that each policy is provided according to the HR manager. We report a model with only the proportion of (highly) skilled staff (Model 1) and a model which includes all controls at the organizational level (Model 2). The full estimates of all control variables are presented in Table 22.3. The results confirm our expectation with regard to flexibility policies: both measures are more likely to be provided when there are more workers in highly skilled positions (Model 1). The inclusion of the control variables accounts partially for the effect of a higher proportion of highly skilled workers on flexible work times, which is no longer statistically significant in Model 2. For the other four measures no association between more high-skilled positions and the likelihood of provision is found. However, for supplemental leave, the effect of more highly

**Fig. 22.4** Effect of the proportion of employees in highly skilled position on likelihood of policy provision at organizational level (*n* = 259), reported by HR manager (odd's ratios) (*Note* See Table 22.3 for full model [model 2])

skilled positions falls short of conventional statistical significance only after control variables are introduced and given the low power at this sample size, we might just not be able to detect this effect with sufficient precision. The effects of control variables included to account for alternative explanations of the skill-policy provision relationship mainly pertain to country differences. In addition, larger companies are more likely to provide childcare assistance and compared to organizations in manufacturing, companies in telecommunication are more likely to provide access to working from home.

The results pertaining to the within-organization educational gradient in access to work-family policies are presented in Fig. 22.4, which shows the exponentiated coefficients (odd's ratios) of a one-unit increase in employee education predicting the perceived availability of each policy by employee's education only (Model 1) and with all individual controls added (Model 2). Because the model included organization-fixed effects, these estimates pertain to the effect of employee education within organizations. We find support for our expectation that higher skilled employees would report higher access to flexibility policies, as Fig. 22.5 shows significant positive effects of a oneunit increase in employee education on the perceived availability of flexible work times and working from home. As Fig. 22.5 shows, the inclusion of the



**Table 22.3**

(continued)


**22 The Educational Gradient in Company-Level Family Policies 593**

*Source* ESWS wave 1

**Fig. 22.5** Effect of employee education on perceived availability of organizational work-family policies within organizations, reported by employee (odd's ratios from organization fixed effects model)

control variables accounts for a large part of the effect of education on access to working from home, implying that individual characteristics other than education play an important role in explaining who can work from home. Also for the third item in the category work-time arrangements, work-hour reductions, higher employee education is associated with higher perceived availability. This goes against our expectation that access to policies which are not costly for the employer nor performance enhancing would not be stratified by employee skill.

Turning to the combined availability at organizational and team level, we predicted employees' perceived availability by the four categories of combined availability, employee education and all individual and organizational control variables. The results of the multivariate models with the individual control variables are presented in Table 22.4, organizational control variables were included in the estimation but excluded from the table.


**Table 22.4** Multilevel logistic regression results (odds ratios) predicting perceived availability of work-family policies at employee level

(continued)



*Note* aProvision of childcare assistance was only reported by the HR manager (org level). Use of vacation days to extend leave is excluded was only reported as actual use at employee level. Coefficients of organization-level control variables excluded from table: size(log), industry, sector, public sector, prop. women in organization, degree of difficulty finding skilled staff, country. Variance components excluded from table. Estimation included a random slope for employee education at organizational level. *t* statistics in parentheses. <sup>+</sup>*p* < 0.10, \**p* < 0.05, \*\**p* < 0.01 *Source* ESWS wave 1

We discuss the main effects of the combined availability on employees' perceived access first before turning to the interactions with education. The first conclusion based on the results in Table 22.4 is that managers' reports of availability are strongly related to employees' perception of availability and this is particularly the case when comparing the "consistent access" category in which managers at both levels report a policy as available to the "no access" category which served as the reference. With regard to the two inconsistent categories, restricted and informal access, the picture which emerges from Table 22.4 is more mixed: Employees' perceived availability of working from home is predicted by informal access, but not restricted access, which suggests that the team (manager) is a more important determinant of perceived availability than the provision at organizational level. Contrary to this, employees' access to a reduction in work hours and supplemental leave is predicted by restricted access, but not informal access, pointing to the organizational level as the decisive layer. Perceived access to flexible work times is higher across all categories when compared to the reference category "no access."

We also tested whether the educational gradient in the perceived availability of work-family policies differed across the combined availability, or in other words, whether higher skilled employees would profit more from any category than lower skilled workers or vice versa. We evaluated all interaction effects between employee education and the combined availability at organizational and team level compared to the reference category of "no access" as well as against the observation-weighted grand mean. As can be seen in Table 22.4, three out of five policies, flexible work time, working from home, and reducing hours showed a positive significant main effect of employee education, implying that higher educated employees are more likely to report these policies as available.

The results from the interactions (results not shown, available upon request) show that for work-time flexibility in fact this positive educational gradient is only present in the "consistent access" category, in all other categories no difference by employee education in the likelihood of reporting access to flexible work times as available is found. For working from home and reducing work hours, no significant interaction effects are found, implying the we find a similar educational gradient across all categories of combined availability. With regard to supplemental leave, interaction effects point to a positive educational gradient in the category of "restricted access" and (slightly) negative educational gradients in the other categories, which explains the absence of a main effect of education. This points to higher skilled workers being better informed about policies when their direct supervisor is not aware of a policy. This could be related to higher educated workers being more likely to work in white-collar positions in which access to information about employee benefits is easier accessible.

For childcare assistance, we do not have information about provision at the team level, but we nevertheless estimate an interaction between access reported by the HR manager and employee education to predict perceived availability of childcare assistance. Similar to the results obtained for supplemental leave, the estimates point to a positive gradient when childcare assistance is provided at the organizational level, but no differences by education when no provision is reported. Again, this explains the absence of a significant main effect of employee education. Like in the case of hour reductions, it also points to an information advantage among higher skilled workers.

The individual-level control variables show effects related to two of the alternative explanations offered in the literature, need for support, and relative bargaining power. The first applies to access a reduction in work hours, supplemental leave, and childcare assistance, which are all significantly more likely to be reported as available by respondents with young children in the household. Women are more likely to report that work-hour reductions are available at their organization. In line with previous research, access to working time arrangements is predicted by indicators of higher relative bargaining power: income and occupational status. It is interesting to note that a higher amount of autonomy predicts higher access to all five policies considered. This indicates that this item captures not only practical aspects being able to work flexibly and thus make use of arrangement.

Finally, with regard to country differences we examined how far the availability of policies differed from the weighted mean of all countries (results not shown), but could not discern a clear pattern related to the statutory policies and most countries did not differ from the mean significantly.

# **Conclusion**

In this chapter, we aimed to provide a comprehensive overview of the educational gradient in the provision of and access to organizational work-family policies while taking a multilevel perspective and examining various types of policies. More specifically we asked whether organizations with a larger proportion of highly skilled employees are more likely to provide access to organizational work-family policies and whether higher skilled employees are more likely to report access to organizational work-family policies. We find that at the organizational as well as the employee level, higher skilled employees have more access to working time arrangements which grant flexibility in the time and location of work. While at the organizational level only access to working from home shows a "net educational gradient" which persists after controlling alternative explanations, all three indicators of worktime arrangements at the employee level show "net education gaps" in access. This is in line with previous research and shows that neither the type of work nor the personal need for work-family support can explain the advantages and benefits higher skilled workers enjoy in their conditions of employment (Chung, 2018; Golden, 2008; Weeden, 2005).

Our uniquely rich data source enabled us to assess also how the provision of work-family policies at organizational and team level, independently reported by HR and team manager, relates to employee's perceived access to the policy and examine in how far this differed by employee skills. We distinguished four categories in this multilevel perspective on the combined availability at two hierarchical levels. The two inconsistent categories of restricted (only access at organizational level) and informal (only access at team level) cover approximately one-fifth of the sample across all policies, of which two-third in the category restricted access and one-third in the informal access category. These inconsistencies which may stem from a variety of processes clearly show the importance of taking into account information from different sources for understanding workers' access to policies.

That the overlap in reported access between the various levels of organizations is not perfect is also apparent in the finding that for all five organizational work-family policies reported access by employees is substantially lower than provision reported by the team managers, which in turn is lower than the provision reported by the HR managers. Nevertheless, we find strong relationships between access to work-family policies reported by both managers (consistent access) and employee's perception. Although we believe that this speaks to the validity of measuring provision of policies at employee level, as is frequently done in survey research, the more mixed results for the two inconsistent categories point to complex processes in the distribution of information in organizations. While for supplemental leave and a reduction in work hours the provision at the organizational level appears to be the more important determinant of employees perceived availability of these measures than the team level availability, for working from home the team manager appears to be more decisive. The fact that a manager reports that a policy is available in the organization clearly does not imply that this is known across the organization or that employees feel they have access to a policy.

We cannot discern between employees not knowing about a policy and perceiving it as inaccessible, which unfortunately limits our deeper understanding of the processes involved and poses a limitation to our research. Moreover, there exists a trade-off between collecting detailed information which allows for deeper insights but limits the scope of a study, and covering many areas with a more superficial inquiry. While the ESWS has an exceptional scope in terms of topics covered, its multilevel design and cross-national nature, it is not always clear how reliable information from different sources in the organization is. For instance, what does it mean if an HR manager reports that a certain policy is generally available in their organization? To whom exactly does this apply and under which conditions?

As pointed out by Chung in Chapter 21 in this volume, few studies have examined the full range of work-family policies and even fewer do so in a cross-national design. To be sure, the collection of a dataset such as the ESWS which allows for such an examination is a challenging and costly endeavor. Nevertheless, it is desirable that future research invests in comparable projects that enable a comparison of different policies across countries. In order to gain deeper insights in the educational gradient in access to policies, more fine-grained questions about the information employees possess as well as their perceptions regarding access to policies for themselves and other workers should be separately assessed.

We believe that the skill-gaps in the access to organizational work-family policies identified in this chapter form an important dimension of social inequality in the labor market. A sustainable European workforce has become increasingly important in the present day. Organizational workforces are displaying growing diversity with respect to age, gender, ethnicity, and family status. Now more than ever, organizations need to consider investing in all workers to improve their performance and levels of satisfaction. These investments can take many forms, including flexible work arrangements, training plans, child-related policies, and health programs (Van der Lippe & Lippényi, 2019), but they should be distributed fairly.

**Acknowledgment** This work was supported by the European Research Council under the European Union's Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement n. 340045.

# **References**


of workplace work-life arrangements in Europe. *European Management Journal, 31*(5), 478–494. https://doi.org/10.1016/j.emj.2013.04.010.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **23**

# **Managing Work-Life Tensions: The Challenges for Multinational Enterprises (MNEs)**

### **E. Anne Bardoel**

Leading scholars in International Human Resource Management (IHRM) have raised awareness of the constraints and challenges related to managing a global workforce (e.g., Scullion, Collings, & Gunnigle, 2007; Sumelius, Björkman, & Smale, 2008), but discussion specific to work-life management in a global context is limited (e.g., Allen, Shockley, & Biga, 2010; De Cieri & Bardoel, 2011; Lewis, Gambles, & Rapoport, 2007). Somewhat earlier, tension-centered theory emerged within the management literature (Poole & Van de Ven, 1989) and, as I argue here, provides a promising framework for understanding the constraints, challenges, and opportunities associated with work-life management in multi-national enterprises (MNEs).

MNEs are central to understanding the global economy. According to the United Nations (2019), as of 2018, flows of foreign direct investment (a measure of MNE activity) were below levels found in 2007. Nonetheless, the value of merger and acquisition expenditures within the total has basically been on an upward trajectory from 2009 to 2018 (from less than \$300b to \$816b), suggesting that individual MNEs are becoming larger over time. Further, most of that expansion was driven by mergers and acquisitions with investment flowing from developed to developing nations, such that cultural differences may be increasingly relevant to MNE operations.

Department of Management and Marketing, Swinburne University of Technology, Hawthorn, VIC, Australia e-mail: abardoel@swin.edu.au

© The Author(s) 2020

E. Anne Bardoel (B)

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_23

Work-life management per se is increasingly acknowledged as an important aspect of HRM (De Cieri & Bardoel, 2015; McCarthy, Darcy, & Grady, 2010; Ryan & Kossek, 2008). Work-life management practices refer to those practices in organizations that are introduced by firms to facilitate employees' ability to meet work and non-work demands (McCarthy et al., 2010). In line with the approach taken by Chung (Chapter 21 in this volume) and Begall and Van der Lippe (Chapter 22 in this volume), work-life policies include flexible working conditions, leave options (e.g., parental, adult care, bereavement, etc.), and child and dependent care (e.g., childcare centers, afterschool care support, etc.) (Smeaton, Ray, & Knight, 2014). Various studies have also linked work-life practices to enhanced employee health and well-being (Grzywacz, Carlson, & Shulkin, 2008), job satisfaction (Forsyth & Polzer-Debruyne, 2007; Muse, Harris, Giles, & Field, 2008), improved employee commitment (Allen, 2001; Muse, 2008; Richman, Civian, Shannon, Hill, & Brennan, 2008) and trust in the organization (Scholarios & Marks, 2004), improved staff morale (McCampbell, 1996), employee performance and organizational citizenship behavior (Lambert, 2000; Muse et al., 2008), and a reduction in employee turnover intentions (Batt & Valcour, 2003; Forsyth & Polzer-Debruyne, 2007).

HRM researchers typically interpret and promote work-life management as an important issue to senior management, develop work-life policies and practices, and work with line managers on work-life policy implementation (McCarthy et al., 2010). However, there are inherent and distinct challenges for work-life management in MNEs because there are different understandings of work-life in different societies and the effectiveness of a MNE's work-life policies will be affected by local factors such as national culture and institutional frameworks (Bardoel & De Cieri, 2006: Lewis & Beauregard, 2015). The concept of work-life balance (WLB) has western origins and it is important to recognize how the different meanings and experiences of worklife issues across an MNE lead to a focus on different work-life policies and programs (Lewis & Beauregard 2015). For example, Coca-Cola recognizes that areas being more severely impacted by the HIV/AIDS epidemic require a different local response than other regions (Coca-Cola, 2019). In Africa, their work-life programs combine prevention, awareness, and treatment, including free condoms, confidential voluntary counseling, and testing for all of their employees and their families. They also make antiretroviral drugs freely available to all employees who need them.

As Morris et al. (2009, p. 987) note in their investigation of the replication of HR practices in global firms, "there is a constant tension between HQ (a.k.a. headquarters) and subsidiaries" in relation to the dilemmas surrounding responsiveness to local culture and legal demands. This logic suggests that substantive tensions will arise in MNEs attempting to implement global work-life programs. For example, work-life balance remains an unfamiliar concept for HR managers in China (Xiao & Cooke, 2012), and language barriers/nuances can add to difficulties in communicating worklife policy in a global organizational structure (Van den Born & Peltokorpi, 2010).

Although tension-centered theory might be used to analyze a broad range of organizational issues, for present purposes, it is here applied to work-life issues in MNEs. Technology change, cultural differences within and between nations, relevant local and national laws, regulations, and political circumstances, and diverse understandings of gender and sexual orientation each impinge on the tensions that emerge around work-life in MNEs. Nonetheless, these issues are addressed only to the extent they are relevant to the emergence or resolution of tensions around work-life in MNEs.

The research question addressed here is whether tension-centered theory can be used to develop a conceptual framework for analyzing and enhancing work-life initiatives in MNEs. The analysis is centered upon relevant bodies of research, with background information from the results of qualitative research on this topic involving managers at 27 MNEs (Bardoel, 2016).

### **Tensions in Global Work-Life Management**

In developed western countries a combination of demographic, labor market, generational changes, and the emergence of a knowledge-based economy and subsequent increasingly competitive landscape for recruiting and retaining talented employees have all been factors driving the growth in work-life initiatives in organizations (Bardoel & Grigg, 2010). More recently, worklife management is also becoming a key concern for employees in many Asian regions because the rapid growth of developing local economies often requires long hours and overtime. A Catalyst study of work-life perspectives in Asia found that employees working in China, India, and Singapore reported high levels of job focus, career ambition, and interest in work-life fit and workplace flexibility (Sabattini & Carter, 2012). Consequently, the concept of work-life effectiveness is coming to the attention of global organizations because, in managing a globally dispersed workforce, they also need to remain sensitive to the role of national, social, and institutional contexts. Evidence of this comes from Chandra's study (2012) of WLB in Anglo-Saxon and Western European countries which showed that WLB practices and family-friendly policies are distinctly different between western and eastern countries, and managing for these differences presents a challenge for MNEs. For lessdeveloped nations, Heymann (2006) finds that culture, gender relations, and economic circumstances each impinge on work-life balance in different ways across nations.

However, although there are variations of work-life issues based on national and cultural differences, work-life consultants Shapiro and Noble (2001) identified three consistent work-life themes that employees from around the world identify as being important barriers to reconciling their work and personal lives (see also Heymann & Earle, 2009). These included a lack of flexible work policies and practices, the availability and affordability of dependent care, and the negative impact of work overload and long working hours. Hence, there are cross-national commonalities in terms of tensions between work and life for employees that support calls for global work-life initiatives.

Tensions might be understood as existing at the level of the individual addressing conflicts between work and family; related research typically finds that the work role often interferes with the family role in developed nations, particularly for women (Grandey, Cordeiro, & Crouter, 2005; Losoncz & Graham, 2010), and that work-life initiatives can respond to these conflicts (e.g., Powell & Greenhaus, 2006). Several recent studies have explored tensions in organizational life, by identifying tensions at both micro- and macro-levels (Ashcraft, 2000; Dallimore & Mickel, 2006). Although tensions at the individual level are crucial to understanding work-life management, applying that term is here restricted to the organizational and cross-national levels. So, for example, conflicts between work and family facing women in one society are relevant to this research only to the extent those conflicts are similar to or diverge from those in other societies, but are not labeled as tensions per se.

# **Tension-Centered Theory**

Stohl and Cheney (2001, pp. 353–354) define organizational tensions, as a "…clash of ideas or principles or actions and the discomfort that may arise as a result of organizational conflict." Following this definition, Tracy argues these organizational contradictions are "inescapable, normal and, in some cases, to be embraced" (2004, p. 121). Lewis (2000) argues that organizational tensions and paradoxes may be used in two ways to develop insights into organizational phenomena and change. First, tension and paradox may be viewed from a structural perspective and understood as a characteristic of

**Fig. 23.1** Tensions in global work-life management

organizational structure that shapes organizational life. Second, tension and paradox may be viewed from an individual/agency perspective, as a social process through which individuals make sense of their organizational life. As Trethewey and Ashcraft (2004, p. 82) explain, "a tension-centered approach begins with the premise that organizations are conflicted sites of human activity" where conflicting demands shape corporate decision making and individual experience.

Poole and Van de Ven (1989, p. 564) argue that tension-centered approaches are an important adjunct for theory building because they expand the scope of inquiry and allow us to "…divulge inconsistencies in our logic or assumptions." At a more practical level, Tracy and Trethewey (2005) contend that tension-centered approaches offer opportunities for organizational change and transformation through, for example, constructing alternative perspectives, or new forms of language and concepts that contribute to change by capturing the tensions inherent in the politicized nature of organizational structures and practices.

Tension-centered theory offers a framework for understanding and responding to conflicts that can otherwise render organizational initiatives ineffective. Is this also true in the case of work-life management in MNEs?

Organizational tensions are placed into the three categories depicted in Fig. 23.1. Drawing on the principles of grounded theory (Glaser & Strauss, 1967; Holton, 2008), these particular categories were selected *ex post*, after themes had emerged from the earlier qualitative analysis (Bardoel, 2016), and were selected in light of both the results and consistency with prior research on global HRM and work-life management.<sup>1</sup> They are described and discussed here to help clarify what follows, and while the categories

<sup>1</sup>The categories were not developed in a theoretical vacuum, and a lengthier list (with six mainly but not entirely overlapping categories) was developed from prior research to generate the interview instrument.

fit the data presented later, other categories could emerge in future studies. Following discussion of the categories, the issue of tension resolution is addressed; that material is drawn directly from tension-centered theory.

## **Strategic/Policy Versus Operational**

Lobel and Faught (1996) identified a strategic approach to work-life initiatives as involving an integrated set of policies, programs, and cultural change efforts that reinforce business strategies (e.g., lower cost producer or providing superior customer service). Implicitly, they recognized that immediate, operational needs of the organization may need to be sacrificed in favor of a strategic, long term approach toward work-life management and business needs. In practice, front-line managers may experience tension between existing, proven operational methods and strategic work-life initiatives. For example, where inflexible and long work hours are the norm, introducing flexible scheduling may disrupt production and raise costs in the short-run.

It is possible that tensions between strategic and operational concerns have heightened in recent decades for some MNEs. While advances in cheap and fast computing and communication may help facilitate global operations, global sourcing has increased dramatically in the last few decades (Hausman, Lee, & Subramaniam, 2005), simultaneous with substantial reductions in inventories (Chen, Frank, & Wu, 2005, 2007; Rajagopalan & Malhotra, 2001). Taking a just-in-time approach to global production makes a MNE subject to high costs from disruptions, which could be associated with adverse weather or other natural events, man-made catastrophes or political unrest. In turn, while strategic work-life concerns might justify shutting a plant for religious celebrations, this might occur precisely when increased production becomes necessary.

### **Centralization Versus Decentralization**

One of the most well-documented issues for MNEs is the tensions caused by the degrees of centralization and decentralization of decision making that result from managing employees across multiple geographic locations (Brewster, Sparrow, & Harris, 2005; Lucio, 2008). Dowling (2009) observed that 20 years ago most MNEs ran their headquarters' HR as silos with little day to day oversight of workplace level HR. Things are different now. The focus of national/regional HR is often to generate routine employment contracts and policies, whereas headquarters' HR will often transmit a whole new layer of global HR initiatives to national/regional HR and employees worldwide. The different foci of the different levels of HR in an MNE can potentially lead to tensions in work-life management (Dowling, 2009).

Pivotal to the centralization versus decentralization debate is that, although work-life policies might be designed and developed at the central level by HR managers and directors, they are generally implemented and managed at the unit level by line managers and supervisors (McCarthy et al., 2010; Ryan & Kossek, 2008). Wise (2005) found in a study of line managers and human resource managers in two financial services organizations that there were tensions between implementing formal work-life policies versus informal and discretionary practice. In this case, although HR had taken a strategic approach to developing a policy related to employees taking time off for dependents, there was considerable variability across the two organizations operating under the same corporate policy.

Somewhat differently, the fact that ultimate authority resides in global headquarters may lead to situations where a mixed policy in terms of centralization/decentralization is perceived as pure centralization by local managers. For example, a global diversity manager claimed to have an 80/20 rule, whereby 80% of work-life policies are standard, with 20% being tied to local business needs, culture, and laws, however a local manager believed that cultural considerations were ignored in the process of policy setting (Bardoel, 2016).

Many studies have concluded that line manager support is crucial in terms of employee work-life outcomes (e.g., Crain et al., 2014; McCarthy et al., 2010; Sánchez-Vidal & Cegarra-Navarro, 2012; Thomas & Ganster, 1995). However, line managers can also be instrumental in undermining formal work-life policies if they believe their performance is strictly tied to costs and productivity (Bardoel, 2016), or by only allowing some subordinates to access policies or in some instances none at all (Hammer, Kossek, Yragui, Bodner, & Hanson, 2009; Kossek & Friede, 2006). Similarly, managers might put subtle pressure on employees not to use work-life policies for fear of being seen as poor workers (Waters & Bardoel, 2006), particularly if the employees are women (Williams, 1999). More subtly, Budig, Misra, and Boeckmann (2012) find that the motherhood wage penalty varies substantially cross-nationally, depending in large measure upon social attitudes regarding mother's employment. It is plausible to suggest those attitudes are often held by supervisors in a position to provide or deny effective access to work-life policies within a MNE.

### **Contextual/Institutional Versus Organizational**

The organizational context for MNEs is characterized by the constant need to respond to global competition and technological innovations which often generate pressure on employees in terms of the patterns and demands of work. At the same time, the contextual and institutional environment in which employees live has undergone dramatic changes in the recent decades (Hein, 2005). The development of work-life initiatives can be traced to the entry of women and particularly mothers into the workforce and professional careers in developed nations, such as the US (Moen & Roehling, 2005; Williams, 1999); work-life initiatives facilitated the retention of these women. Internationally, with the exception of Africa, it is also true that women's labor force participation grew faster than men's in recent decades (Lim, 2002). However, rates of women's labor force participation vary substantially across individual nations, altering the context for MNE work-life initiatives across nations.

For examples of the complexities involved here, De Cieri and Bardoel (2009) find that MNE efforts to attract and retain women diverge depending on local gender norms and economic conditions. MNE managers claim that Korean women tend to stop working after they marry and have children; women in Singapore and Hong Kong face an abundant supply of cheap immigrant, domestic labor, making childcare less of an issue, while women in China and Vietnam may have access to childcare performed by other (women) family members (p. 190).

Simultaneous with the increase in women's labor force participation were decreases in fertility, particularly in developed nations (Lim, 2002), motivating work-life initiatives to accommodate non-caregivers. More recently, although women continue to provide most childcare, men's participation has expanded in developed nations (Bianchi, Robinson, & Milkie, 2006; Craig & Mullan, 2011), increasing the value of work-life initiatives that respond to the caregiving commitments of men. Very differently, workforces are aging in many developed nations (Bardoel & De Cieri, 2006), which expands the value of work-life initiatives that help retain older workers while facilitating caregiving commitments to the elderly.

A rapidly developing trend in recent years has been the expansion of corporate rights and benefits, particularly in the US, for gay or lesbian employees, and for transgender employees (reported by 91 and 83% of Fortune 500 companies, respectively; HRC, 2019). Almost simultaneously, the legalization of same-sex marriage has occurred in many developed nations (Procon.org, n.d.). Work-life initiatives among MNEs may be well-suited to respond to these shifts, but may create tension between organizational consistency and the local context if they operate in nations, such as Nigeria, where same-sex marriage is illegal (ibid.). Luiz and Spicer (2019) found MNEs based in developed countries with operations in Africa found it difficult to export LGBT inclusive policies and, while some made the attempt, ultimately local laws and mores tended to be controlling. Broadly, the specific groups covered by diversity and inclusion initiatives varies substantially, as found in a sample of corporate websites in four European nations and the U.S. (Jonsen, Point, Kelan, & Grieble, 2019), making it less than obvious which groups should be targeted in work-life initiatives across an MNE.

Legislation related to work-life has been passed in most countries where MNEs might operate (Heymann, Earle, & Hayes, 2009). Legislation might in some cases effectively replace or supplant MNE work-life policies, as in the case of paid family leave. However, the diversity of coverage in terms of length of maternity leave, payment mechanisms, whether paternity leave is covered, and leave policies for other purposes vary substantially (ibid.). This diversity makes the application of generic work-life policies less efficient because, for example, employees in one nation might be largely satisfied with state-funded maternity or paternity leave, and instead place a high value on child care supports. Elsewhere, leave provisions might be more highly valued, enhancing the value of flexible work-life policies that can be applied appropriately under local conditions. More specifically, Ollier-Malaterre (2009) finds in a comparison of corporate work-life policies in the U.S., U.K., and France that there is far lesser coverage in France, in part because of state support for early childhood education, but also because French employers tend to view work-life programs as a social benefit, rather than a competitive advantage.

Legislative differences may in part reflect cultural differences, and those differences could similarly drive a wedge between the effectiveness of any specific work-life policy or practice across nations. For this reason, Von Glinow, Drost, and Teagarden (2002) encouraged future researchers to use broad research lenses with multiple embedded contexts when conducting globally distributed HR research.2 Similarly, work-life scholars have called for future research to incorporate a focus on how global organizations can be inclusive of work-life issues in multiple cultural contexts (De Cieri & Bardoel, 2015; Poster & Prasad, 2005).

More broadly, these shifts and differences may imply that what represents an effective work-life policy or practice in one nation may be in tension with a policy or practice that is effective in another. Further, it is not clear that resolving this tension in favor of local context is always ideal, given that

<sup>2</sup>This research is specifically related to IHRM. For simplicity, the term "global HR" is used instead.

employees often work or are transferred cross-nationally, and that knowledge of inconsistent or conflicting policies could become widespread and generate adverse publicity or low employee morale.

Martin, Gollan, and Grigg (2011) provide an example of how these tensions intersect in the context of employer branding of MNEs. They argue that branding creates an inherent tension between differentiating the organization and making it fit in or be similar to others. That tension can be viewed through the lens of local tailoring of work-life initiatives, which could either serve to differentiate the organization in order to attract and retain rare talent (e.g., through working time flexibility) or by mimicking the practices of other local organizations (e.g., through design of locally-specific diversity programs). That is, even if local input is valuable in terms of addressing the tension between centralization and decentralization, there remains a tension between contextual/institutional factors and organizational factors.

# **A Framework for Resolving Tensions in Global Work-Life Management**

Having identified potential tensions related to global work-life management, the next step is to consider ways to address these tensions. Poole and Van de Ven (1989, pp. 566–567) identify four potential responses to tensions, or what they label "paradoxes," that can be applied to global work-life management: opposition, spatial separation, temporal separation, and synthesis. The first, opposition, involves recognition and understanding of tensions, while the next three strategies represent conscious attempts to resolve tensions. An alternative is added to this list of four potential responses: no recognition of tension. This option is not strategic, but is instead added to account for the possibility that work-life initiatives designed without recognition of the potential for organizational tensions may well exacerbate those tensions, or that unrecognized tensions may undercut the effectiveness of any work-life initiative. Each approach is discussed in turn.

# **Opposition**

The *opposition* strategy involves understanding and accepting that tension exists. For example, suppose an MNE created a company-wide policy supporting reduced-hours arrangements as a strategic matter, and that policy created tension with operational needs in some facilities. The opposition strategy involves recognizing, studying and understanding this tension. The key to opposition strategies lies in an explicit recognition of tensions, and not necessarily an explicit resolution. As such, it "…accepts paradox and tries to work out its implications, [and] can serve as a preliminary step to the other three" (Poole & Van de Ven, 1989, p. 567).

#### **Spatial Separation**

The second strategy, *spatial separation*, involves clarifying the distinct dynamics at different levels of the organization and different roles across these levels. It can also or instead be applied in the geographic sense; for instance, when two horns of a paradox differ by "physical or social locus" (Poole & Van de Ven, 1989, p. 566). For example, spatial separation could involve placing the responsibility for implementing work-life initiatives in the hands of the geographic managers at lower levels of the MNE and giving them accountability for measuring their own success (for related examples, see Budhwar, 2000; Currie & Procter, 2001). Differently, flexibility might only be made available to high-level employees or rising stars, which is a form of spatial separation which may respond to tension between strategic and operational concerns, but could cause demoralization among front-line employees.

Spatial separation is particularly relevant to the present analysis. Rozenweig (2006) concluded that HRM is the function in an MNE that is most likely shaped by local responsiveness. Earlier work by Rozenweig and Nohria (1984) found that the forces for local responsiveness were strongest for those HR practices in MNEs which are influenced by well-defined cultural norms and by local labor laws. In MNEs, work-life policies and practices need to be informed by local norms and laws because the relationship between work and non-work activities will be heavily influenced by both cultural norms and labor laws (Bardoel & De Cieri, 2006; Edwards & Kuruvilla, 2005). An organization's ability to respond to work-family issues is also related to institutional forces that vary across nations (see Milliken, Martins, & Morgan, 1998; Wood & De Menezes, 2010), creating tensions that are unique to global work-life initiatives in MNEs.

Note that spatial separation may in practice create tensions. For example, if local work-life initiatives are implemented and gauged using local criteria as a way to attract talent, that talent may be developed in a fashion which creates expectations the enterprise does not meet elsewhere. That tension, in turn, may limit the ability of the enterprise to develop talent around the globe which can be promoted to headquarters level of transferred to manage operations in different cultural contexts.

## **Temporal separation**

The third strategy is *temporal separation*, which involves taking time into account. For example, if a corporation has spent many years developing global work-life policies, and a strong commitment has been developed (e.g., at IBM, Childs, 2005), it will be easier to mandate standardized work-life policies across all geographies of the MNE. If employees are often moving between countries within the MNE, the value of a consistent approach will be enhanced. Very differently, a temporal separation strategy could involve initial experimentation at a limited number of locations, with expansion later, or the development of a time-line for implementing local work-life initiatives, with different locations starting on the time-line over time. Note that IBM explicitly conceptualizes a temporal separation approach, with the intention to expand the reach of policies over time, which contrasts with spatial separation, where expansion of policy coverage is not planned.

A major advantage of these temporal separation policies is that they allow for feedback and learning over time. Bardoel (2016) found an MNE in banking that pushed out flexible work arrangements, encountered resistance in operations in a different nation, and discovered that managers wanted field training in managing flexible work. That training was provided, which allowed the expansion to continue.

An alternative temporal separation policy involves allowing for, and building out from, local initiatives. For example, a regional HR manager for a health insurance company developed work-life policies specifically designed for an aging workforce, which global HR ultimately provided to other MNE locations (Bardoel, 2016).

Temporal separation may, however, create new tensions. For example, if IBM had developed its global policies in the context of operations in Western, developed nations, the policies might contravene local law and culture in Asian or African nations (e.g., regarding LGBTQ rights).

## **Synthesis**

*Synthesis* refers to developing new terms and concepts to resolve tensions. The switch among HR researchers from the term "work-family" to "work-life" for relevant programs during the 1990s (Harrington, 2007; Lewis et al., 2007) may be seen as involving such a synthesis. That switch made the programs beneficial to a larger segment of the workforce, and arguably allowed corporations to attract and retain valuable employees, regardless of parental or caregiver status.

Another example of synthesis involves efforts to cast work-life initiatives as part-and-parcel of broader diversity initiatives. Like diversity initiatives in general, work-life initiatives recognize and respect the unique circumstances and talents of individual employees, and strive to turn conflicts around difference into sources of synergy and success (Childs, 2005; Dallimore & Mickel, 2006). Somewhat differently, it is possible to conceive of a synthesis approach to strategic versus operational tensions wherein line managers are provided with the flexibility to develop and implement work-life initiatives that are copacetic with workplace operational needs (for examples, see Rapoport, Bailyn, Fletcher, & Pruitt, 2002).

An alternative synthesis approach lies in incorporating work-life discussions into annual reviews, thereby combining evaluation along other metrics with work-life. Relatedly, evaluations of supervisors might include explicit measures of support for work-life (e.g., proportion of employees using flexible work arrangements).

Another synthesis approach might involve incorporating unique, local issues into a global work-life programs. For example, making physical and temporal space for religious observances, or addressing a local health problem (e.g., vaccinations during an outbreak). For a specific example, a US MNE proactively sought to hire disabled employees using a business case rationale. A regional HR director in China responded that the business case did not make sense in China, but ultimately implemented the policy as consistent with corporate social responsibility (Bardoel, 2016).

Synthesis is arguably more likely when the management allows ideas to percolate up from far-flung operations, which is arguably more consistent with a decentralized and contextual/institutional approach. That is, to the extent headquarters support local innovation in work-life programs, the more innovations are likely to occur. However, there is no obvious reason to expect those ideas to be coherent at the organizational or strategic levels, which may simultaneously reduce their usefulness in terms of synthesis for global operations, and even create new tensions (e.g., around religious practices in one society).

More broadly, while synthesis may often sound promising, in practice it is as likely as the spatial or temporal separation approaches to create new tensions. For example, the switch from "work-family" to "work-life" programs may have led some non-parent employees to believe that child care supports are unfair since many employees will never use them. More generally, many non-caregivers might often perceive what are pitched as work-life programs as in fact work-family programs, which are not inclusive of the needs of all employees, which could in turn generate perceptions of unfair treatment.

#### **No Recognition**

The fifth option is *no recognition* of tension. Logically, this approach is prior to the opposition strategy, since it may take time to identify tensions. However, it might continue indefinitely in some cases. Returning to the example of reduced-hours options creating tension with operational needs, if corporate culture is such that local managers are uncomfortable reporting on the tension to headquarters, as suggested by research on employee voice and silence (e.g., Morrison, 2011), the MNE might leave the tension festering over time. Alternatively, if information flows are poor within a facility, local managers might have difficulty identifying relevant tensions if the only evidence available is indirect, as in poor quality or low productivity.

A key problem for global work-life strategies, which is relevant to the "no recognition" approach, is that they may be difficult to monitor and assess (Masi & Jacobson, 2003). Unlike measures of quantity and quality for manufactured goods, it may be costly to ascertain whether work-life policies and practices are useful and effective, and the results might be ambiguous. Absent accurate measures, a no recognition approach may easily emerge.

The strategies of spatial separation, temporal separation and synthesis are not mutually exclusive, and "can be combined in practice" (Poole & Van de Ven, 1989, p. 567). For example, the provision of identical paid parental leave to new mothers and fathers might be viewed as equitable in one society while, in another, it would create a backlash against new fathers who use the policy either because usage represents a betrayal of norms around masculinity or because most fathers are assumed to use the leave for purposes other than childcare (e.g., Rhoads & Rhoads, 2012). Distinct policies could be implemented in the two societies (spatial separation), with an eye toward implementing a uniform policy over time if attitudes around fatherhood shift (temporal separation), while simultaneously either strategically blurring or sharpening the distinction between time off policies for purposes of unpaid care or for leisure (synthesis).

# **Discussion**

For managers, it is reasonable to conclude that consideration of the three axes of tensions around global work-life initiatives, and of the five potential responses considered here, will often be of value. Ignoring the potential for tensions may invite the creation of irrelevant, ineffective, low-value programs. The findings also suggest that absent a corporate culture that is open to discussion of tensions, work-life initiatives may exacerbate existing tensions and global HQ may not even recognize that it is creating problems, or at least not in a timely fashion. A further implication of this possibility is that monitoring of global work-life initiatives may be valuable.

There are at least two potential sources of bias which may relate to both managers and researchers from Western cultures. One source may emerge in cases where Western notions of progress yield an inaccurate projection of future cultural movement in non-Western societies such that temporal separation may seem reasonable when in fact spatial separation is more appropriate. Another potential source of bias lies in the possibility that some readers, perhaps including some managers, will tend to respond most favorably to the synthesis strategy. In Poole and Van de Ven's (1989) article, synthesis is presented last, and arguably includes some of the most innovative and interesting practices, which involve "new concepts or a new perspective" (p. 567). If this bias exists, it might also be attributable to Western notions of progress.

The discussion leads to four questions relevant to future research and practice:


If these questions are answered in the affirmative, this would support the utilization of a tensions approach to global work-life initiatives. That is, each question asks whether efforts to address potential tensions are necessary for effective global work-life programs.

In terms of limitations, it is possible that I have overstated the value of the tensions approach. Perhaps most obviously, efforts to search for broad principles (RQ1) could result in overly amorphous, ambiguous and perhaps even operationally meaningless initiatives. Relatedly, expansion of the terrain covered by global work-life initiatives (RQ3) could lead to an overly broad program which strives to serve everyone in terms of diversity, family issues, work-life, job stress, and so forth, which is ultimately ineffective because efforts are spread too thinly. Finally, it is debatable whether a tensions approach is necessary to believe that monitoring of global work-life program effectiveness is valuable (RQ4), as managers may simply believe that any new practice or procedure should be monitored.

Although the results of future research cannot be known in advance, there are sound reasons, provided above, for pursuing these research questions and a tensions approach to global work-life initiatives.

# **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Part VI**

# **The Next Decade of Research**

# **24**

# **Childcare Indicators for the Next Generation of Research**

**Sebastian Sirén, Laure Doctrinal, Wim Van Lancker, and Rense Nieuwenhuis**

# **Introduction**

The contributions to this handbook are all testimonies of the significant progress that has been made during the last couple of decades in terms of mapping policies and analyzing the associated political processes and impacts across countries, time, and multiple levels of policy-making. The existing literature arguably also reflects how the availability of data on relevant aspects of family policy shapes the research conducted and the kind of questions researchers are able to answer. Further advancements of this research agenda, as envisioned in several of the preceding contributions, as well as in the final and concluding chapter of this book, are accordingly also constrained by the restricted availability of relevant indicators on important aspects of

S. Sirén · L. Doctrinal

L. Doctrinal e-mail: laure.doctrinal@sofi.su.se

W. Van Lancker (B)

R. Nieuwenhuis Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden e-mail: rense.nieuwenhuis@sofi.su.se

Swedish Institute for Social Research, Stockholm University, Stockholm, Sweden e-mail: sebastian.siren@sofi.su.se

Centre for Sociological Research, University of Leuven, Leuven, Belgium e-mail: wim.vanlancker@kuleuven.be

institutional variation across countries. Pushing the availability of policyrelated indicators further should accordingly be a primary concern for those interested in the development of this research field.

In this chapter, we elaborate on the need to develop theoretically grounded family policy indicators. Since an encompassing overview of all family-related policies is beyond the scope of this chapter, we will focus specifically on early childhood education and care (ECEC, in the remainder referred to as childcare), an area for which the development of relevant policy indicators is clearly challenging, but where overcoming these challenges is a precondition to achieve further scientific progress. Although we limit the discussion in this chapter to this specific policy field, the insights highlighted below very much reflect issues that are relevant to scholars within the area of family policy research more broadly.

The first section of this chapter critically introduces the conceptual frameworks underpinning the most prevalent currents in comparative research. The next section then presents the most prominent empirical approaches utilized in existing studies, and how these depend on the availability of relevant data as well as the analytical outlooks of the respective scholars. The third section then maps the availability of comparative data on the most widely used indicators and discusses the main sources from which this data originates. The final section concludes by pointing toward some challenges for the current research agenda, along with some tentative solutions. In particular, we argue for the need to engage in the development of indicators on social rights, reflecting the institutional variation across countries, across time, as well as across various levels of policy-making, with regard to childcare in particular and family policy in general.

# **Conceptualizing the Role of Policy in Comparative Research on Childcare**

The historical changes in the organization of our societies and the (at least partial) recasting of traditional gender roles within the familial sphere, have created new needs and real challenges for existing family policies (Daly, 2011; Lewis, 2001; Montanari, 2000). While undoubtedly driven by the emancipatory struggles of the women's movement, the weakening of the male breadwinner model and the increased marital instability has in recent decades been described as generating "new social risks" to which increasing women's employment is proposed as a functional solution (Armingeon & Bonoli, 2006). Along the same lines, the recognition of gender equality and the challenges associated with changing family arrangements have been incorporated as distinctive features of the social investment discourse, which surfaced as a central paradigm in debates about the future of welfare states (Saraceno, 2017). Partly as a result of this turn, much interest has recently been oriented toward the variation of public provisions that facilitate the reconciliation of work and family life (Abendroth & Den Dulk, 2011; Hegewisch & Gornick, 2011; OECD, 2007; Plantenga & Remery, 2005), of which childcare services constitute a cornerstone.

Childcare services are accordingly considered instrumental in fostering mothers' employment (Gornick, Meyers, & Ross, 1997; Keck & Saraceno, 2013; Olivetti & Petrongolo, 2017) and greater gender equality in earnings (Evertsson et al., 2009; Grönlund & Magnusson, 2016; Nieuwenhuis, Need, & Van der Kolk, 2019). Childcare policy is moreover often promoted as a tool to reduce the risk of vulnerable families falling into poverty, by allowing mothers, and especially single mothers, to remain in employment (Maldonado & Nieuwenhuis, 2015; Misra, Moller, Strader, & Wemlinger, 2012). Finally, childcare services of high quality are also conducive for cognitive and noncognitive development of children as well, improving school readiness of, in particular, disadvantaged children. This allows for longer-term improvements in social mobility and poverty (Cunha & Heckman, 2007; Leseman & Slot, 2014; see Chapter 8 by Vandenbroeck in this volume). As a result, childcare provision has risen high on the agenda of international organizations such as the OECD (2017) and the European Commission (2013), unmistakably demonstrated in the proposal for a directive on work– life balance for parents and carers as part of the European pillar for social rights (European Commission, 2017; see also Chapter 3 by Jenson in this volume).

This development clearly reflects the increasing traction of the "adult worker" as an emerging norm in international discourses (Daly, 2011; Lewis & Giullari, 2005). Analytically, these trends are often conceptualized in terms of "defamilization," defined as the degree to which one's dependency on the family is lessened by provisions from the state or by recourse to the market (Leitner & Lessenich, 2007; Lohmann & Zagel, 2016; see also Chapter 6 by Zagel & Lohmann in this volume). These and associated concepts are then often used to construct clusters of countries adhering principally to each of these policy orientations, respectively (Ferrarini, 2006; Korpi, 2000; Leitner, 2003; Saraceno & Keck, 2010).

Critics of how care is treated within the social investment framework have pertinently highlighted the tendency to promote defamilization of care as a means toward the commodification of women's work, in order to counter poverty risks and falling fertility rates associated with changing family arrangements (e.g., Esping-Andersen, Gallie, Hemerijck, & Myles, 2002). Thereby such a discourse runs the risk of contributing to a devaluation of unpaid care work traditionally carried out by women in the position of mothers, daughters, or spouses (Daly, 2011; Jenson, 2009; Saraceno, 2017). Likewise, an instrumental focus on child development and maternal employment leaves little room for women's agency and choice, for example, to organize childcare according to their own desires (Yerkes & Javornik, 2018). Similarly, while the role of childcare policies for greater equality of opportunity in coming generations is becoming widely recognized (Burger, 2013; Esping-Andersen, 2015), current policy-oriented research generally takes the parents or the family as the main subject of study. Children are in this context often perceived as "becomings," rather than "beings" (Lister, 2003), and are thus devoid of inherent rights and agency (Daly, 2019).

The conceptual frameworks used in much of the existing literature often conflate causes, institutions, and outcomes, thereby encumbering analyses of causal relationships between these (Korpi, 2000; see also Chapter 6 by Zagel & Lohmann in this volume). The current ambition in the literature to move toward explanatory analyses of how institutions evolve and subsequently shape outcomes thus requires critical re-examination of existing concepts. While separate indicators on outcomes tend to be relatively accessible, an increased awareness vis-à-vis the role of institutional variation highlights the need to develop conceptually informed indicators also in this regard. The increasingly common emphasis on the multidimensionality of family policies also points toward a need for these indicators to reflect qualitative differences between different modes of provision (Keck & Saraceno, 2013; Korpi, Ferrarini, & Englund, 2013; Pettit & Hook, 2009; Yerkes & Javornik, 2018).

# **Empirical Approaches in Previous Research**

In line with the divergent historical patterns of family arrangements and policy developments, childcare policies show considerable cross-country variation. Roughly since the mid-1990s, a number of studies have sought to classify systems of childcare services (Anttonen & Sipilä, 1996; Gornick et al., 1997; Kamerman, 2000; OECD, 2001; Randall, 2000; Rauch, 2007; Rostgaard, 2002; Saraceno & Keck, 2010). As a result of these and other research efforts, there seems to be widespread consensus that ideal childcare policies ensure that the services are available, affordable, and of good quality (Gambaro, Stewart, & Waldfogel, 2015; see also Chapter 8 by Vandenbroeck in this volume). These three dimensions are interrelated to some extent. For a service to be affordable and/or providing high-quality care, it needs to be available in the first place (Yerkes & Javornik, 2018). Affordable childcare will fail to increase the number of children enrolled if their parents do not have access to childcare places. However, comprehensive collection of comparable indicators on all three dimensions that could allow for systematic quantitative analyses of the institutional configuration of childcare services and their consequences is still scarcely developed.

In the literature on social policy in general, and family policy in particular, three types of policy indicators can be discerned: (1) indicators based on expenditure data; (2) indicators capturing receipt and usage of policies based on survey or administrative data; and (3) indicators capturing the design of policies based on institutional or legislative data (Marchal & Van Lancker, 2019; Otto, 2018). Many comparative studies of childcare provisions have used public expenditures and enrolment rates as the main policy indicators (Bonoli & Reber, 2010; Huber & Stephens, 2000; Jensen, 2009; Pavolini & Van Lancker, 2018; Rauch, 2007). As the levels of expenditure on these services is largely a function of their take-up, it should come as no surprise that these indicators are highly correlated (Jensen, 2009, p. 12). Both measures accordingly share a number of features with implications for their usefulness in comparative research.

Like in other fields of social policy studies, expenditure data have figured prominently as an indicator of public effort and the degree of socialization of family-related activities. As a result of increased efforts by international organizations, government spending has been systematically collected for a significant number of countries, at least across the affluent welfare states of the OECD. Also, disaggregated expenditure data for different subsets of policies, separating spending on parental leave, child benefits, and childcare services, have become increasingly available (Bonoli & Reber, 2010; Castles, 2003). Still, expenditure data suffers from drawbacks not always recognized in the literature. Especially, changes in expenditure levels not necessarily reflect qualitative changes in policies, since these can also be driven by demographic changes. Such measures are also strongly influenced by short-term fluctuations in the denominator, in this case often GDP, especially during times of economic turmoil. To account for this, spending indicators should be adjusted for need as well as for GDP fluctuations (Otto, 2018). This however affects the interpretability of the indicator.

Other issues cannot be readily solved. Varying definitions and principles for data collection, for example, might weaken the comparability of spending indicators. These inconsistencies can result from different funding streams not adequately captured in the data, especially regarding expenditure at levels below the national government, as well as fiscal expenditures. This is particularly relevant for childcare, as these services are provided by regional and/or local governments in many countries (Javornik, 2014), while other countries subsidize families using childcare services through the tax system. Finally, changes in spending, even if they are adjusted for need and economic cycles, can mean different things in different policy contexts. An increase in spending on childcare could refer to an increase in the generosity of vouchers in one country or to an increase in the number of available places in the other. The outcomes of these two parametric changes are likely to be substantially different.

As an alternative, many studies have relied on survey or administrative data to calculate measures of the uptake of childcare policies, in the form of enrolment rates (OECD, 2007; Pavolini & Van Lancker, 2018; Plantenga & Remery, 2009; Thévenon, 2011). In particular, the efforts by Eurostat, the European statistical office, to collect harmonized statistics on the take-up of childcare services in the European Union Statistics on Income and Living Conditions (EU-SILC) surveys, made these indicators increasingly available to researchers. However, enrolment rates and expenditures are inherently problematic when used as an indicator of government effort, or more generally as an indicator of public policy (see Clasen & Siegel, 2007). The main reason is that such measures are functions of both state provision and public demand, and that the uptake of a policy as well as the resulting expenditure levels should essentially be seen as outcomes rather than policy variables (Kangas & Palme, 2007). Supply of childcare places is indeed influenced by public policies, while the demand for childcare is shaped mostly by the number of working mothers, which in turn is a product of prevailing norms, labor market conditions, and other related policies (Javornik, 2014). As a result, childcare enrolment and expenditure may change in a country even if there is no change in public policy in that area, for instance in times of high unemployment or changing ideals regarding parental care (Evertsson & Grunow, 2016).

Accordingly, aggregates such as expenditures or enrolment rates reveal little about the "nuts and bolts" of national childcare systems, including features related to relevant aspects such as availability, affordability, and quality. Nor do such aggregate indicators allow for examining how features of policies affect specific subgroups of the population differently. This is a significant shortcoming as the causes of social stratification related to work–family reconciliation measures is a major concern in current scholarship (Chapter 11 by Hook & Li; Bonoli, Cantillon, & Van Lancker, 2017; Korpi et al., 2013; Pavolini & Van Lancker, 2018; Pettit & Hook, 2009; Van Lancker & Ghysels, 2016; Yerkes & Javornik, 2018). Additionally, addressing complex interactions between different sets of policies that often jointly produce the outcomes that scholars of family policy impacts are concerned with, requires detailed information on policy design (Javornik, 2014). Within the domain of family policy, the interaction between leave policies and childcare provision is one such theme (Blum, Koslowski, Macht, & Moss, 2018). Another is the extent to which availability, affordability, and quality jointly shapes the uptake of childcare among different categories of families (Yerkes & Javornik, 2018). Data that would allow analyzing such interactions are however scarce, and this situation becomes particularly dissatisfying as the research agenda aims to move from descriptive toward explanatory ambitions.

As an alternative to expenditures and take-up, institutional indicators reflecting the actual entitlements of families have been solicited (Hegewisch & Gornick, 2011; Javornik, 2014; Korpi et al., 2013; Nieuwenhuis, Need, & Van Der Kolk, 2012). Such an institutional approach has previously been proposed as an operationalization of the concept of social rights, applicable across different welfare states and policy fields (Korpi, 2000; Marshall, 1950; Orloff, 1993). This has been the guiding principle of some of the most cited empirical works within comparative welfare state research more broadly (Allan & Scruggs, 2004; Esping-Andersen, 1990; Korpi & Palme, 1998), but the development of equivalent indicators regarding childcare services is still in its infancy.

Such indicators are generated by systematizing legislative information into indicators of entitlements or social rights, and can be presented in textual form. However, in order to reduce the level of complexity and thereby facilitate comparative statistical analyses, these indicators are often expressed in quantitative terms. In practice, such quantitative indicators are constructed using model family methods, by which benefit levels and other institutional features are deduced based on assumptions regarding household compositions and employment status (Bradshaw, Ditch, Holmes, & Whiteford, 1993). By holding such factors constant across observations this methodology aims to provide a measure of the temporal and cross-sectional variation in the institutional design of family policies. Researchers have come a long way in constructing useful comparative datasets along these lines regarding cash benefits, including child benefits, paid parental leave, and family-related tax credits, (Ferrarini, 2006; Gauthier, 1996; Saraceno & Keck, 2008; Van Mechelen, Marchal, Goedemé, Marx, & Cantillon, 2011). Institutional indicators have moreover been applied to study differences in policy design between distinct welfare states (Ferrarini, 2006; Ferrarini, Nelson, & Höög, 2012; Javornik, 2014; Korpi, 2000; Marchal & Van Lancker, 2019), as well as the impact of such differences on poverty and inequality (Bäckman & Ferrarini, 2010; Keck & Saraceno, 2013; Korpi et al., 2013; Misra et al., 2012; Nieuwenhuis et al., 2019; Pettit & Hook, 2009; Van Lancker & Van Mechelen, 2015). Institutional indicators have been employed to study the extent to which policies support female, and especially maternal, employment (Hegewisch & Gornick, 2011; Nieuwenhuis et al., 2012; Pettit & Hook, 2005), and to which family policies are conducive toward independence from family relationships, in terms of defamilization (Javornik, 2014; Leitner, 2003; Leitner & Lessenich, 2007; Lohmann & Zagel, 2016; see also Chapter 6 by Zagel & Lohmann in this volume). Such indicators also featured in discussions about how family policies influence the range of opportunities actually available to families in contemporary political economies, in terms of families' "capability sets" (Hobson, 2013; Yerkes, Jana, & Kurowska, 2019; Chapter 7 by Javornik & Yerkes).

Continuous measures based on institutional information, often in combination with indicators on outcomes such as labor market structures and incomes, have moreover been used by several scholars to classify countries into different ideal-typical family policy or care regimes (Anttonen & Sipilä, 1996; Bettio & Plantenga, 2004; Leitner, 2003; Meyers & Gornick, 2003; Saraceno & Keck, 2010). While ideal-typical welfare state regimes can be important theoretical constructs and useful to demonstrate systematic differences between countries, or to present country variation in a concise manner, their analytical usefulness is however limited. The limitations of the regime approach are related to the difficulties in accounting for differences within clusters, changes over time, specific features of policy designs, and their interactions with other policies (for an approach that remedies some of these shortcomings see Javornik, 2014).

# **Databases on Childcare Services: The State of Affairs**

As discernible from the preceding sections, comparative studies of childcare systems based on quantitative indicators have been facilitated by several important data collection efforts. An early example is that of the European Commission Childcare Network (Gauthier, 2000), which through a series of reports reviewed the childcare services and policies throughout the European Union (Cohen, 1989; Moss, 1990, 1996). Although primarily descriptive in their contents, these publications nevertheless provided a broad documentation of childcare provision in Europe, opening the field for further research based on these data and the development of a number of new datasets in the field. For example, Kamerman (2000) constructed a range of indicators, some based on this data, to develop a typology of childcare services in the OECD, differentiating systems by age of children, administrative auspices, and types of providers. In another study, Gornick et al. (1997) mapped the provisions of a number of family policy areas, including publicly regulated and subsidized childcare, linking such provisions to the employment of mothers. The data collected for this latter study was subsequently deposited with the Luxembourg Income Study (LIS) data center and made available to the public.

The important progress made in this field during the 1990s caused observers to proclaim that "the provision at national level of publicly funded services providing care for children with employed parents (at least up to ten years of age) […] is by now fairly well mapped" (Deven et al., 1999, p. 5). Nevertheless, a subsequent report from Eurostat underlined the weaknesses and limits of the available statistics, pointing out the difficulty "to obtain comparable statistics covering childcare services" (Eurostat, 2004, p. 1). The issue lay largely in the harmonization of statistics available at the national level. In particular, the way data were collected varied widely between countries, for example, regarding the method of data collection (survey or administrative data) or the categories used when coding the data (age groups, attendance, source of funding, type of childcare considered). The report concluded that the figures presented with regard to coverage and enrolment of children "should be treated with a great deal of caution, especially as regards their comparability between countries" (Eurostat, 2004, p. 29).

With the adoption of the Barcelona targets by the European Council in 2002, introducing explicit goals for the expansion of childcare services on offer, the need to monitor progress was brought into the spotlight. The subsequent need to map and evaluate childcare policies underscored the deficit in existing comparable data and prompted the inclusion of childcare indicators in the EU-SILC survey as a solution (Plantenga & Remery, 2009). In 2005, with most Member States covered, the EU-SILC became the reference source for statistics about childcare, used among others by the European Commission to measure the attainment of the Barcelona objectives.

The definition of formal childcare in EU-SILC covers preschool or equivalent, compulsory education, as well as professional childminders and centerbased childcare services, including out-of-school care, irrespective of the provider. A distinction is also made between full and part-time care, with the threshold set at 30 hours per week. Figures are moreover available on the use of informal forms of childcare, including friends and family members other than the parents. The Eurostat statistics based on EU-SILC also report aggregated data on the average duration of childcare per week in each country. Data are updated on a yearly basis. In 2016, variables referring to the affordability of childcare services, unmet needs for such services, and reasons for not making use of such services were added through an ad hoc module.

The OECD made use of these indicators, complementing them with information regarding non-European OECD countries, to set up a comparative database on family policies (Adema, Huerta, Panzera, Thévenon, & Pearson, 2009; OECD, 2007). These increased efforts by international entities to collect and publish comparative indicators have meant important progress toward increasing the cross-country comparability of data on childcare services, and have rendered a number of studies analyzing the impact of childcare policies, mainly with regard to inequality, labor market, and fertility outcomes (e.g., Kröger, 2011; Pavolini & Van Lancker, 2018; Thévenon, 2011; Thévenon & Solaz, 2014; Van Lancker & Ghysels, 2016).

Although the EU-SILC overcomes previous difficulties in terms of data harmonization, certain issues remain, several of which are mentioned in the report of the European Commission (2013) about the attainment of the Barcelona objectives. Data on the percentage of children under 3 cared for in formal structures in 2010–2011 are for instance "compiled from small samples and are statistically unreliable" for several countries<sup>1</sup> (European Commission, 2013, p. 7). Moreover, conceptual and methodological problems, for instance, regarding the lack of distinction between private and public provision of childcare, the small sample sizes for more than half of the countries in the survey, and errors in the coding process have been noted by others as well (Keck & Saraceno, 2011).

This is a real issue. In the questions on childcare enrolment in EU-SILC, a distinction is made between compulsory school (ISCED 1), preschool (ISCED 0), and formal childcare services (including crèches, childcare centers, and paid childminders which is a common way of organizing formal childcare in many countries). However, these different categories do not always match the types of services and preschools or the common age of enrolment in these services in particular countries. In England, for example, according to official numbers 72% of 2-year olds were enrolled in early education places in 2018, including private providers, state-funded primary schools and nurseries, and childminders (Department for Education, 2019). For 3 to 4-year olds, this was 95%. Our own calculations based on the EU-SILC data for 2018 show that 66% of 2-year olds were enrolled in early education

<sup>1</sup>The countries for which issues of reliability are raised are: Austria, Bulgaria, Cyprus, Czech Republic, Greece, Hungary, Lithuania, Malta, Poland, Romania, and Slovakia.

places, increasing to 83% for 3- to 4-year olds. Formal registration of enrolment of children does not need to be the same as coverage rates referring to the number of children being in formal childcare service during the past week, as it is asked in the EU-SILC questionnaire. But still, the difference is striking, and there is no apparent explanation for it.

A possible explanation is that questions are misunderstood by the respondents. To give one example, in the EU-SILC questionnaire, parents are asked to provide the number of hours each child spends in compulsory school, preschool, and/or formal childcare services during a regular week. The distribution of the responses sometimes indicates that parents interpreted this as number of hours per day, with mean scores of 7 (e.g., in EU-SILC 2012 for compulsory school in the UK) or 5.8 (childcare at center-based services in the UK in EU-SILC 2018). Finally, tracking evolutions in childcare coverage over time seems to be a problem for some countries as well. For instance, Eurostat (2019) data for Italy reports childcare coverage for under threes of 27% in 2008, dropping to 20% in 2012, increasing to 34.4% in 2016, only to drop again to 25.7% in 2018. Yet, there is no concomitant change to be observed in terms of expenditure on childcare in Italy (own calculations based on the OECD Social Expenditures database). The real issue here is that these problems are hardly or not even documented in the yearly quality reports released by Eurostat (Eurostat, 2020). Nevertheless, EU-SILC remains a prominent source of data on childcare services, in particular with regard to enrolment rates.

Apart from Eurostat and the OECD, a number of additional entities have compiled and published comparative indicators on the organization of childcare services. Table 24.1 gives a non-exhaustive overview of existing databases, as regards their geographical coverage as well as information regarding whether or not they are still updated and the last year for which data are available. As pertinently highlighted in the framing of this handbook, family policies in general, and perhaps particularly childcare services, are formulated and implemented at different levels. The role of supranational organizations in issuing recommendations and collecting data has been touched on above. As have the decisive role of regional and local authorities in providing services and regulating nonpublic providers. Despite this, previous research has almost unanimously focused on variation on the national level, partly for reasons of data availability but also given the theoretical underpinnings of studies and the need to reduce complexity in order to make comparisons possible. All databases reviewed below accordingly refer to the national level.


**Table 24.1** Geographical coverage and updating of databases

(continued)



aLast checked in May 2020. For references, see Sect. "Databases Included in the Tables"

The geographical coverage varies somewhat across the databases. While West European countries are included in all databases, except in the Trans-MONEE database, East European and Central Asian countries feature only in the UNECE and TransMONEE datasets.<sup>2</sup> The NOSOSCO database collects data regarding Nordic countries only. Data are generally available

<sup>2</sup>The Cesifo DICE report also includes some East European countries on the top of those included in the EU.

for fewer countries with regard to care services for the youngest children, compared to preprimary education services. Most databases are regularly updated with new waves of data, with the exception of Multilinks, the Family policy database, and the Work–family policy indicators.

The Family policy database, DICE, and the Work–family policy indicators provide data collected in several countries but referring to a single point in time. Other databases provide longitudinal data, although data availability varies across countries and indicators. In general, data about childcare in Central and Eastern European countries, as well as Latin America, African and Asian countries is scarce, especially as regards the quality and the affordability of childcare services.

Comparing the different datasets reveals some of the challenges involved when trying to find valid indicators, in particular as regards the conceptualizations of childcare formulated in relation to each database. The definition of formal childcare sometimes covers different realities depending on what kind of service is included in the definition. The OECD Family Database, for example, includes children enrolled with registered childminders, while other databases, such as the LIS Family Policy Database, excludes them. A central issue in most databases is that it is often not possible to distinguish between public and private services, or to define whether services are publicly subsidized or not, the extent of public subsidies, nor to assess whether service providers are obliged to comply with standards and regulations set up by public authorities. As an example, both the OECD Family Database and the Multilinks Project clearly set the objective to provide indicators on coverage by public or publicly subsidized formal childcare services. A closer look at the sources used in both databases shows nevertheless that the EU-SILC, which includes places in private facilities and services, is used in the OECD Family Database to define the enrolment rate of children aged 0–2. Data on coverage from the Eurostat 2004 report, also including private arrangements, is also used in both OECD and Multilinks databases.

Table 24.2 provides a non-exhaustive overview of the content of these databases, structured along four dimensions of ECEC services; governance, availability, quality, and . While the significant number of databases included in this table might seem to indicate a fairly high availability of useful indicators, it is important to emphasize that the original sources of many indicators are the same in many of the separate datasets. In particular, most childcare policy databases rely heavily on EU-SILC data. Data provided in the DICE database rely mainly on the Eurydice country reports and are not aggregated in a single database but made accessible in separate files. Similarly, data included in the Eurydice reports (2014/15, 2019) are not aggregated in a database but available through the publications.


**Table 24.2** Overview of data availability on childcare services

(continued)


*Note* Last updated May 2020

The governance and the organization of childcare services are key instruments to ensure available and accessible services. Fully integrated systems and unitary systems seem to offer greater coherence in the supply of places throughout the age groups and to be associated with improved care quality and more equitable provision of services (European Commission, 2019). Information about governance is only available in Eurydice publications (2014/15, 2019) and in the OECD Family policy database, although information is less detailed in the latter.

Availability is perhaps the most basic feature of childcare provision, as services need to be available in the first place before other features can become relevant. Although sometimes conceptualized as separate features, issues regarding accessibility and flexibility might also be seen as aspects of availability. Generally, availability is operationalized using enrolment rates, and a quick look at Table 24.1 reveals that most of the data collection efforts have indeed focused on this indicator. This, as argued before, is somewhat unfortunate since the actual use of childcare services is the result of a combination of several factors (Javornik, 2014). As an alternative institutional measure some have instead opted for an indicator of the existence of a legal entitlement to a place in a childcare setting (Saraceno & Keck, 2008).

Although the entitlement to be granted a place in public childcare is by now fairly well mapped, the data has until recently not allowed grasping crosscountry differences in terms of opening hours and flexibility in this regard, including the availability of services during nonstandard hours. The inclusion of such indicators in the Eurydice key data on Early Childhood Education and Care has however greatly improved the possibility to assess these aspects within a comparative approach (Yerkes & Javornik, 2018). Less is however known about conditions under which a place in childcare is granted, and how factors related to the family, such as family composition, income, or the parents' employment status, influence the entitlements of children. Being unemployed or on parental leave for instance might affect the number of hours a child is allowed to be in a childcare setting, as well as the associated fee. This aspect is partially tackled in the Eurydice report (2019) but only as regards children living in poverty.

Regarding the affordability of childcare, few databases have given an attempt to calculate the costs of childcare for different family types. The contribution of the OECD (further used by Multilinks and DICE) are particularly valuable in this context. Here, net costs of using childcare services are calculated for six family scenarios involving two types of households (lone parent and couple households with two children) at different earnings levels. Based on Eurydice publications,<sup>3</sup> DICE has moreover collected data on financial support that parents can be granted in order to afford the cost of childcare. Fees are also provided in Eurydice's comparative reports, but these figures might not be fully comparable across countries. Finally, the ad hoc module "Access to services" implemented in EU-SILC in 2016 includes indicators referring to the affordability of childcare services. With available indicators, it might however be difficult to assess how costs vary depending on family income, parental employment, and type of setting. More detailed information about funding schemes and cost regulations is needed in order to assess how affordability varies across groups within countries, and between different types of providers (Yerkes & Javornik, 2018).

As mentioned earlier, the quality of care matters from the perspective of child development, and is often considered as necessary in order for childcare services to actually promote social integration and equality of opportunity (European Commission, 2013; Van Lancker, 2013). Perceptions of quality may moreover influence parents' attitudes toward the usage of childcare services (Yerkes & Javornik, 2018). While a common definition of childcare quality is lacking, existing accounts often distinguish between structural and process quality. Structural quality here refers to different aspects relating for instance to the conditions of the staff (educational requirements, salaries), or to the features of the care setting (group sizes, child-to-staff ratios), while process quality is situated at the level of interaction between staff and children and refers to the care-related and pedagogical activities taking place in the childcare setting (OECD, 2018; Penn, 2012; Van Lancker, 2013; Yerkes & Javornik, 2018).

In practice, process quality figures scarcely in comparative research, much due to the lack of comparable indicators (Yerkes & Javornik, 2018). Structural quality is most commonly operationalized with reference to child-tostaff ratios and the required educational qualifications for the staff in childcare settings. Moreover, only the OECD Family database and DICE provide updated information about the requirements in terms of education for the staff, as well as regarding the legislation on group size and/or child-to-staff ratios. Again, the role of policy in ensuring a certain quality of the services is not always easy to disentangle, as substantial variation in actual provision between providers might remain despite the norms set by governments. Also, the available data do not allow to distinguish between public and private settings, again potentially obscuring the role of policy in shaping actual provision.

<sup>3</sup>Available from 2014.

In addition to direct measures of different features of quality in childcare settings, parents' subjective perception of quality might also relevant for their attitudes toward childcare and sequentially for their inclination to make use of these services.4 This is accordingly a highly relevant indicator to understand why families chose different work–family arrangements. However, it can be hard for parents to assess the actual quality of the services on offer (Vandenbroeck & Lazari, 2014), and there might be multiple factors shaping parents' perceptions of quality. Accordingly, one needs to be cautious when analyzing public policies, structural and process quality, and parent's perceptions thereof.

To summarize, data on childcare services have been made increasingly available, on a growing number of policy dimensions. However, the development has been highly uneven across time and policy fields, resulting in a patchwork of datasets all using different definitions, concepts, and methods, while several databases rely on the same original sources. An important observation is also that data collection has often been initiated from the political sphere, motivated by a need to monitor progress on prioritized issues. In the concluding section, we briefly discuss these observations and call for more academically driven efforts to collect data-based well-defined and theoretically grounded concepts.

## **Conclusion: A Look Ahead**

This article has been concerned with the availability of conceptually relevant indicators for comparative family policy research. The existing challenges and some potential solutions have been discussed using childcare policy as an example, in order to illustrate the challenges associated with comparative research in this field. Underlying the narrative above is the awareness that analytical research, engaged in explaining important social phenomena, is dependent on reliable empirical indicators, grounded in relevant theory. The consequences of data limitations will accordingly become increasingly evident as the interest of researchers become oriented toward explanation rather than description.

Improved data could, for example, facilitate inquiries about why countries differ substantially with respect to the availability, affordability, and quality of their childcare services, as well as with respect to the timing of its implementation and the mode of governance. As the distributional consequences

<sup>4</sup>Eurofound has a question on quality of services in their European Quality of Life surveys (waves in 2003, 2007, 2011, and 2016), which is based on a subjective assessment of respondents.

associated with the reorientation of welfare states according to a social investment logic is being increasingly scrutinized (Bonoli et al., 2017), researchers interested in childcare policy might in addition be in a better position to explain variation in the uptake of childcare services across socioeconomic groups, with reference to variation between national policy frameworks (Korpi et al., 2013). Another question, for which more detailed institutional indicators might be of high relevance relates to how and to what extent the outcomes of childcare policies depend on the interaction between different policy instruments, that is, exploring the role of institutional complementarities. Also, increasing marketization in childcare provision is expected to sharpen inequalities in usage and quality of the services on offer, but to empirically scrutinize such expectation indicators differentiating between public and private provision should be available. Corresponding questions and associated implications with regard to the requirements on available data can be raised in relation to other family policy instruments.

Having access to well-defined institutional indicators not only facilitates studies with explanatory aims. Disentangling institutions from their associated outcomes makes it possible for researchers to redefine the normatively based purposes of specific policies and to analyze policy impacts on outcomes deemed as relevant from a normative viewpoint. Better data can accordingly also contribute to more critical assessments about the ends and goals of contemporary family policies, and how these relate to changing perceptions about families, work, gender, old age, and childhood. Innovative research on childcare policy have in this vein recently come to ask how policy variations influence cross-country differences in parents' agency, in terms of their real opportunities to arrange childcare according to their potentially divergent desires (Yerkes & Javornik, 2018).

Although an important step forward, developing institutional indicators that reflect national policy frameworks will not be a panacea for all the challenges that comparative family policy analysts currently grapple with. The question of how to deal with variations on subnational levels still remains to be thoroughly discussed and examined, as current practices of using averages or "typical" policies will increasingly be called into question. Also, the issue of implementation and evaluation unescapably comes to the fore. As the focus is shifted toward policy scripts, questions regarding the extent to which policies reflect the actual services delivered by providers, or whether policies are merely a dead letter, will also undoubtedly arise. The need for research that links indicators of policy "supply" with related outcomes will potentially help to critically address this issue. And for some questions, such as the question what actually happens within childcare centers, quantitative approaches should be complemented with qualitative approaches as well. Studies of the relationship between structural and process quality are cases in point.

Although this review of the existing literature and databases reveals increasing efforts made over the last decades in the analysis of childcare services, a coherent research infrastructure with institutional childcare indicators, based on theoretically defined concepts, remains elusive. As the above review has indicated, there is a clear need to improve comparability of existing indicators, at the same time as the conceptual framework needs further elaboration in order to improve our understanding of causes and consequences of cross-country differences in childcare policies. We suggest a framework for data collection emphasizing the role of public institutions in structuring availability, affordability, and quality of services. We believe that pursuing the effort to capture cross-country policy variation, is likely to further the research agenda on the role of family policy for social inequality among families with children, between households with and without children, as well as between men and women. Moreover, we think that such an agenda would benefit from a perspective that integrates family policies, including social care services, as essential components of social citizenship.

**Acknowledgment** The research leading to these results has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No. 730998, InGRID-2—Integrating Research Infrastructure for European expertise on Inclusive Growth from data to policy.

## **Databases Included in the Tables**


## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **25**

# **Family Policy: Neglected Determinant of Vertical Income Inequality**

**Rense Nieuwenhuis**

Family policies have played a pivotal role in facilitating the rise of employment and earnings of women (Gornick & Meyers, 2003; Nieuwenhuis, Need, & Van der Kolk, 2012; Olivetti & Petrongolo, 2017; Thévenon, 2011), one of the most prominent developments in economic activity in OECD countries in recent decades. As has been demonstrated in an exhaustive literature and throughout this Handbook, family policies of various kinds have featured prominently in analyses of the differences in the economic activity between groups, such as women and men or mothers and women without children. These are examples of *horizontal* income inequality: income differences between groups of individuals. At the same time, a concern for high and rising levels of income inequality between households (re)emerged, focusing on factors such as globalization, technological change, wealth accumulation, and austerity as explanations for rising income differences between households (Atkinson, 2015; Iversen & Soskice, 2019; Milanovic, 2016; OECD, 2015; Piketty, 2014; Taylor-Gooby, Leruth, & Chung, 2017). This is often referred to as *vertical* income inequality: inequality between individuals or households based on their place in the income distribution. In this literature on vertical income inequality, the potential role of family policies has been neglected. This is surprising, as horizontal and vertical income inequality are inherently linked: as family policies reduce the income differences between

R. Nieuwenhuis (B)

e-mail: Rense.nieuwenhuis@sofi.su.se

© The Author(s) 2020

Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden

R. Nieuwenhuis and W. Van Lancker (eds.), *The Palgrave Handbook of Family Policy*, https://doi.org/10.1007/978-3-030-54618-2\_25

women and men—or other groups of individuals—this affects not only income differences within households, but also between. In this chapter I set the record straight and develop a research agenda that incorporates family policies as an institutional determinant of vertical income inequality between households.

First, I elaborate upon how analyses of *vertical* income inequality and *horizontal* income inequality have generally considered widely different types of determinants. In line with the scope of this Handbook, the emphasis will be on institutional determinants. I will then demonstrate that women's employment and earnings directly relate to levels of economic inequality among households; as levels of vertical and horizontal inequality are interrelated, so are their determinants. This, finally, brings into focus the perspective on how family policies might affect vertical income inequality between households, and what research questions this raises. This new perspective is, naturally, well-informed by the existing literature on the economic outcomes of family policies, of which key references are highlighted here. It also raises new questions, that are explored in conclusion.

# **The Different Determinants of Horizontal and Vertical Income Inequality**

Contemporary analyses of income inequality emphasize how in most advanced capitalist societies (including most European and OECD countries) inequality has been rising since the 1980s, although to varying degrees. Between the mid-1980s and 2013, vertical income inequality, as measured by the so-called Gini coefficient, increased on an average with over 10% in the OECD, with inequality rising fastest in Sweden with a 35% increase (OECD, 2015). Gini coefficients measure the income differences among all households (accounting for the number of individuals living in these households). As such, this is a form of *vertical* income inequality, pertaining to the overall shape of the income distribution and thus analyzing how far all individuals or households in a given population are apart. This is depicted on the vertical axis in Fig. 25.1. Naturally, such analyses can focus on a variety of income concepts, including earnings and wages.

Prominent explanations have been formulated for rising levels of vertical income inequality. Examples of key determinants have been depicted in Fig. 25.1, and include the impact of globalization (on levels of inequality within countries) (Milanovic, 2016), European integration (Beckfield, 2019), robotization and algorithms, or more precisely skills-biased technological

**Fig. 25.1** Selected determinants of vertical and horizontal economic inequality

change (Atkinson, 2015; Autor, 2014; Frey, 2019), the accumulation of capital (Piketty, 2014), pension privatization (Ebbinghaus, 2011), polarization between high- and low-skilled workers in the transition to the knowledge economy (Iversen & Soskice, 2019), precarious work and the return to mass unemployment (Bernstein, 2016; Kalleberg, 2009; Lohmann & Marx, 2018), decline in unionization (Ebbinghaus & Visser, 2000; Korpi, 1983), tax avoidance (Saez & Zucman, 2019), austerity (Taylor-Gooby et al., 2017), and longer term processes of welfare state retrenchment—in particular with respect to social assistance and other aspects of minimum income protection (Marx & Nelson, 2013; Nelson, 2013). Importantly, these developments are generally interrelated.

Income differences between women and men, a form of *horizontal* income inequality, have generally been declining in most countries over recent decades—although they have certainly not ceased to exist (Charles, 2011). Women have entered the labor market at much faster rates than men, closing the gender gap between women and men having an income of their own. Yet, while for instance in the EU 78% of men were employed in 2017 (EU & EIGE, 2020), only 66.5% of women were. Among those who are employed, the gender pay gap has been closing slowly, but in the EU, in 2017 the gross hourly wage of women was only 84% that of men (ibid.). Trends and levels of women's employment and gender inequality have been explained by a combination of demographic and institutional determinants (Nieuwenhuis et al., 2012). Demographic explanations of women's rising participation in the labor market (Lesthaeghe, 2010) include their rising educational levels (Bussemakers, Van Oosterhout, Kraaykamp, & Spierings, 2017) and declining fertility (Van der Lippe & Van Dijk, 2002). Institutional explanations, a selection of which are depicted on the horizontal axis of Fig. 25.1, pertain to determinants that include gender role norms in relation to "doing gender" (Evertsson, 2014; Knight & Brinton, 2017; West & Zimmerman, 2009), expansion of the service sector, horizontal and vertical occupational segregation (Charles & Grusky, 2004), and overly long work hours expected in high-wage occupations (Goldin, 2014).

When it comes to family policies, it has often been observed—and demonstrated throughout this Handbook—that women's economic activity tends to be more equal to men's in welfare states based on dual-earner/caregiver policy models, characterized by public childcare for young children, brief periods of well-paid parental leave for both parents, as well as individual taxation, compared to in the traditional breadwinner model—characterized by very long periods of leave for mothers, childcare only for somewhat older children and joint taxation (Korpi, 2000; Lewis, 1992). With respect to specific policies, public childcare has been a key determinant for women's employment (Olivetti & Petrongolo, 2017), as well as the motherhood wage penalty (Halldén, Levanon, & Kricheli-Katz, 2016), in particular when it is available, affordable, and of adequate quality (Gambaro, Stewart, & Waldfogel, 2015; also see Chapter 8 by Vandenbroeck in this volume). The provision of parental leave has also been instrumental to women's employment (Moss, Duvander, & Koslowski, 2019), as long as it has been adequately paid (Gornick & Meyers, 2003) and not overly long (Nieuwenhuis, Need, & Van der Kolk, 2017a).

Although depicted as separate axes, there is a small literature showing that some of the determinants of vertical income inequality between households have also affected levels of horizontal income inequality between women and men. This is shown in Fig. 25.1 by the dashed, black arrow. For instance, Iversen and Soskice (2019) argue that the shift to the knowledge economy has been beneficial to the employment opportunities of in particular high-skilled women, whereas among the lower skilled women's jobs are considered at greater risk than men's of being replaced by automation (Brussevich et al., 2018). Albeit with huge differences across welfare states, austerity has rendered the social security on which women rely less effective, and has reduced wages and the number of jobs in the public sector in which women are overrepresented (Rubery, 2015). As women are more likely to have more restricted and incomplete contribution histories—in part related to a lack of formal family policy measures over the life course—women's old-age incomes might be hit harder than men's in relation to pension privatization (Möhring, 2015, 2016).

A clear gap in the literature is how family policies have affected levels and trends of vertical income inequality. It is of course true that some lines of research examine the link between family policies on overall levels of inequality. An example is literature studying whether particularly children growing up with socioeconomically disadvantages parents benefit from enrolment in early childhood education and care (ECEC), as such promoting overall levels of equality of opportunity (Gambaro et al., 2015, also see Chapter 8 by Vandenbroeck). Yet, this line of research does not explicitly address vertical *income* inequality, even though there are very good reasons and some initial evidence to suggest that family policies, indeed, can affect levels of inequality among households. This lacuna in the literature is the focus of this chapter, and is depicted in Fig. 25.1 by the long red arrow.

Developing a research agenda on the potential impact of family policies on levels and trends of vertical income inequality among households is important for at least three reasons. First and foremost, of substantive importance, it may demonstrate that the literature on income inequality needs to include an additional set of determinants, and the results may give policymakers an extra set of instruments to curb inequality. This does require, however, a detailed understanding of the mechanisms through which, and conditions under which, family policy instruments can attenuate levels of income inequality. Secondly, of methodological nature, the potential influence of family policies may be further underlined in relation to a gender perspective of how income inequality is typically measured. Many such analyses of income inequality are based on a measurement of *household* income, which is equivalized for the household size and composition to approximate levels of income inequality among individuals (Cowell, 2011). This assumes, however, that all household income is shared equally even though this is not the case (Bennett, 2013). As it is known that women who have an income of their own have more access to and control over household resources (Sen, 1990), family policies that facilitate women's employment and own incomes may reduce overall levels of vertical inequality in ways that are not captured by traditional measurements of inequality—including control over household resources. The alternative, measuring inequality solely based on individual incomes assumes no sharing of income and resources among household members, which seems equally unrealistic. As such, a focus on family policies that not only affect income differences between women and men in general, but also income differences within households, will provide a more comprehensive analysis of vertical income inequality as well. Finally, with respect to the academic community, it should be observed that researchers studying the determinants of economic inequality, and researchers studying family policy outcomes, tend to discuss their work at different conferences (or at the very least in different tracks of conferences) and tend to publish their work in different journals. This research agenda thus has the potential to bring these communities and literatures together, and foster a multidisciplinary cross-pollination.

# **How Women's Earnings Affect Vertical Inequality**

Family policies can affect vertical income inequality by facilitating women's employment rates and average incomes, and subsequently by influencing *who* is employed and the distribution of women's incomes. Regarding the latter mechanism of family policy, little is known about how this affects vertical income inequality—as will be discussed in detail in the next sections. Regarding the former, however, there is by now a fairly clear consensus that without the rise in women's employment and earnings inequality between households would have been higher. This literature, however, only covered inequality among couples, and only shows the average effect of women's rising employment on vertical income inequality.

Early studies hypothesized that with more women entering the labor market, vertical income inequality between households was bound to rise when spouses' earnings were positively correlated, and to decline with spouses' earnings were negatively correlated (Mincer, 1962). There is indeed a tendency for higher educated women and women with higher earnings to be married to (/live together with) spouses who themselves have high levels of education and earnings, which is referred to as homogamy or endogamy (Kalmijn, 1998; Schwartz, 2013). Indeed, educational homogamy is found to boost the correlation between spouses' earnings (Breen & Salazar, 2011), and this homogamy of earnings in itself is a factor that contributes to more vertical inequality among households (Nieuwenhuis, Van der Kolk, & Need, 2017b). However, when inequality is measured in relative terms—as it typically is—a positive correlation between the earnings of spouses is a required but not a sufficient condition for women's earnings to increase inequality between households (Lam, 1997). Even when spouses' earnings are positively correlated, inequality between households will only be exacerbated by women's earnings when the inequality among women is sufficiently high compared to inequality among men's earnings. This is an important point, as it shows that to understand how women's earnings affect inequality between households not only depends on the degree to which spouses' earnings are correlated, but also on the level of inequality among women themselves. As more women entered the labor market, their (potential) earnings were argued to become more important when it comes to partner selection, thus promoting homogamy (Oppenheimer, 1988, 1994; Sweeney & Cancian, 2004). Yet, at the same time, more women having an income of their own reduced inequality among women because of the simple fact that fewer women had zero earnings (Cancian & Reed, 1998; Gregory, 2011). Thus, the rise in women's earnings observed in recent decades set into motion two dynamics with opposite outcomes (Nieuwenhuis et al., 2017b). On the one hand, women´s earnings tend to be higher in households where their partner also has high earnings, thus increasing vertical inequality. On the other, more women have earnings. This reduces income differences among women, which in turn reduces vertical inequality. The latter dynamic has been more powerful in recent decades, explaining how the rise of women's earnings has reduced vertical inequality.

With only a few exceptions (Esping-Andersen, 2007, 2009), the inequality-reducing impact of women's employment and earnings was successfully replicated in contexts as widely different as Sweden (Björklund, 1992), Norway (Mastekaasa & Birkelund, 2011), the United Kingdom (Harkness, Machin, & Meguir, 1996; Machin & Waldfogel, 1996), the United States (Betson & Van der Gaag, 1984; Cancian & Reed, 1998; Treas, 1987), Ireland (Callan, Nolan, O'Neill, & Sweetman, 1998), Mexico (Campos-Vázquez, Hincapié, & Rojas-Valdés, 2012), and Brazil (Sotomayor, 2009). In many of these countries, women's earnings reduced inequality even though they were positively correlated to those of their spouses.

The degree to which women's earnings reduce vertical income inequality between households varies substantially across countries, as was highlighted in a number of country-comparative studies. Cancian and Schoeni (1998) compared 10 developed countries to show that women's earnings reduced vertical inequality in all countries, and that this attenuating impact became stronger over time *despite* the employment rising faster among women married to high-income spouses than among women married to spouses with middle-class incomes. They further calculated that the correlation between spouses' earnings would have to increase substantially in order to nullify this attenuating impact of women's earnings on vertical inequality. Examining 14 European countries with widely different employment rates among women, Pasqua (2008) also finds that in all countries women's earnings reduce inequality between households. In countries where women's participation in the labor market was lower, inequality among women was higher while in these countries there was more potential to reduce inequality between households further by promoting more universal women's employment rates. In line with these findings, Harkness (2013) also finds that women's earnings reduced inequality more strongly in the Nordic countries with high rates of women's employment, compared to for instance southern European countries. Moreover, she further contributed the insight that closing the gender pay gap would further reduce inequality among households, although this impact would be smaller than what could be achieved by further raising women's employment rates. Would more women work full-time jobs across OECD countries, this would further reduce inequality between households (OECD, 2015). Nieuwenhuis, Van Lancker, Collado, and Cantillon (2020) also found the rise of women's employment to be associated with a reduction in relative poverty rates.

To illustrate cross-country variation in the degree to which trends in women's employment and earnings have reduced inequality between households, Fig. 25.2 summarizes results published elsewhere (Nieuwenhuis et al., 2017b; Nieuwenhuis, Need, & Van der Kolk, 2019). Among heterosexual

**Fig. 25.2** Rise in women's earnings (1981–2008) associated with lower inequality among households of couples

couples across 18 OECD countries, the share of total household earnings contributed by women was measured around 1980 and 2010. In all countries—at both points in time—women's earnings constitute less than half of average household earnings, but their share increased over time. It was then determined how much the *change* in women's earnings in total household earnings (*x*-axis) had contributed to vertical inequality between households (*y*-axis). In other words, how much higher/lower observed levels of inequality are around 2010 compared to a simulated scenario in which women's earnings had remained the same since the 1980s. These analyses were based on the strong assumption that nothing else changed in this period, but are still informative about the potential magnitude of how much changes in women's earnings affected inequality between households. Figure 25.2 shows how the increased share of women's earnings in total household earnings was associated with a reduction of vertical inequality in each of the countries, and how this impact was strongest in countries that had relatively low rates of women's employment (and consequently earnings) in the 1980s, such as the Netherlands. This is a figure that shows the impact of trends in women's share in total household earnings, which explains why the Nordic countries—already showing high rates of women's employment and earnings in the 1980s—show such a small reduction in vertical inequality.

An early analysis of the impact of family policies, based on the data summarized in Fig. 25.2 (Nieuwenhuis et al., 2019), showed that work– family reconciliation policies as paid parental leave and public childcare were indeed associated with women's earnings more strongly reducing vertical inequality. Although these policies were found associated with a somewhat elevated correlation between spouses' earnings (increasing inequality), this effect was outweighed by the attenuating impact of a reduction in the inequality among women themselves. The study also examined the impact of policies typically associated with the traditional breadwinner model, such as family allowances and tax benefits for families with children, but found these not to be associated with vertical inequality between households.

# **Family Policies as Determinant of Vertical Economic Inequality: A Research Agenda**

The literature summarized in the previous section demonstrates *that* women's earnings have attenuated inequality between households, and suggests that family policies can indeed facilitate this, but leaves unanswered important questions pertaining to the *mechanisms* through which family policies affect inequality. A research agenda on understanding how family policies might affect vertical inequality between households addresses a number of questions. Who uses family policies, what are the benefits of this use in terms of their earnings/incomes, and with whom do they form a household? These questions are interrelated, and the vast literature on family policy outcomes addresses each of these questions, as has been demonstrated throughout this Handbook. The research agenda proposed here builds on what has already been learned. Below, I revisit some of the ongoing debates in the family policy literature and relate these debates to new questions from the perspective of vertical inequality between households.

## **Who Uses?**

It has long been acknowledged that the benefits of family policies (just like many other types of policy) are not the same for everyone. For example, higher educated parents are substantially more likely to enroll their children in public childcare compared to lower educated parents (Van Lancker, 2018a; Van Lancker & Ghysels, 2012). Cash-for-care schemes (and other forms of very long, low-paid leave) may be considered a mechanism of exclusion from the labor market. Indeed, the Norwegian cash-for-care system was found more likely to be used by mothers with lower levels of education, lower income, and with a migration background (Bugum & Kvande, 2013). In contrast, child benefits are also not uniformly distributed but sometimes have policy designs that benefit higher income families (Van Lancker & Van Mechelen, 2015). Fathers are more likely to take parental leave when they are highly educated, but less likely when they are self-employed, work on temporary contracts, or in the private sector (Geisler & Kreyenfeld, 2011; Romero-Balsas, 2012).

For the research agenda developed here, the use of family policies is to be differentiated by people in different positions on the income distribution (e.g., by income decile). This will shed light on the question whether different family policies have the potential to lift up the bottom of the income distribution, strengthen or hollow out the middle, or ensure that the top of the income distribution gets ahead even further. The question of who uses family policies of course also relates to fathers, to which I will return in the section on "with whom."

#### **What Income Effect?**

It is not only important to understand where in the income distribution people are who use family policies, but also what the effect of that use is in terms of earnings and income. The additional income earned in relation to the use of family policy may also vary across the income distribution. In other words, it matters where in the income distribution people end up after using the family policy, or alternatively where they would have been had it not been for using family policies. Two debates come to the fore here.

The first ongoing debate pertains to potential trade-offs in family policy outcomes (Pettit & Hook, 2009), not only with respect to gender but also class (see Chapter 11 by Hook and Li in this volume). In countries where a large proportion of women are integrated into the labor market, women were less likely to be in high-status and (importantly here) well-paid positions (Semyonov, 1980). Welfare state interventions that include paid maternity leave and public childcare (as well as a large public sector) were found to facilitate women's employment, but also found to promote occupational segregation with women overrepresented in "female-typed" occupations and underrepresented in managerial positions (Mandel & Semyonov, 2006), which in turn was associated with wage penalties (Mandel, 2013). Taking a different position in this debate, Korpi, Ferrarini, and Englund (2013) argue that these studies only find a lower likelihood for women to reach top positions because they do not account for the large numbers of women remaining outside of the labor market in countries without work–family reconciliation family policies. Instead, they argue that these policies only affect the likelihood that women reach a top position relative to all working women (with the relative proportion of women in top positions relative to all working women being lower in countries with policies that also integrate a large number of lower skilled into employment) but do not affect the likelihood that any woman reaches a top position (also see: Grönlund & Magnusson, 2016).

Secondly, the question remains open how family policies affect wage gaps across the income distribution (including, but not necessarily limited to, gender and motherhood wage gaps). Although there is plenty of literature on (determinants of ) wage gaps (Blau, 2016; Blau & Kahn, 2017), such analyses often focus on socioeconomic characteristics such as level of education, rather than income position. For instance, Halldén et al. (2016) found that although public childcare helped reduce the motherhood pay gap, this effect did not seem to differ among higher and lower educated mothers. On the other hand, Budig and Hodges (2010) examined motherhood wage penalty across the wage distribution, finding it to be bigger for low-wage mothers (for a methodological comment, see Killewald & Bearak, 2014). Examining the impact of time spent on housework on wages, Cooke and Hook (2018) find that men do the least housework and incur the largest wage penalty for doing so when they are at the top of men's wage distribution, whereas for women incur a larger wage penalty at the bottom of their wage distribution. Morosow and Cooke (2018) found that Finnish fathers who take parental leave incur a wage penalty, but only those fathers at the bottom of the wage distribution.

Taking together who uses family policies and to what income effect raises the question to what extent people at different positions in the income distribution use various family policies, and how the income or wage effects associated with using these policies differ across the income distribution. A family policy that is beneficial to wages may reduce vertical levels of income inequality when it is predominantly used by people at the lower end of the income distribution *and/or* when its effects are strongest at the bottom. Yet, if a policy is mostly used by people at relatively high-income positions and/or the effects are strongest at the top, such policy may exacerbate vertical income inequality.

## **With Whom?**

Perhaps the cornerstone of the research agenda developed here is to examine the family policy outcomes in relation to household composition and processes of household formation. Whether family policies will affect income inequality between households not only depends on who the individuals are that use family policies and to what income effect, but also on the incomes of others with whom people form a household. If those with a higher (a priori) income are more likely to use a specific family policy and/or to benefit from this use to a greater extent, this will exacerbate inequality to a larger extent if these individuals live with (for instance) a high-earning partner. In other words, this research agenda brings to the fore a dyadic perspective on family policy outcomes. Such a dyadic perspective brings into focus at least three avenues for further developing research in family policy outcomes—in general as well as with a specific focus on vertical income inequality. The first is how the use and associated income effects of family policy relate to other household members (if any). The second is how family policies may affect household composition. The third pertains to the within-household sharing of—and control over—resources.

It is well-documented that the uses and benefits of family policies vary across individuals, and often along the socioeconomic spectrum (see sections "Who Uses?" and "What Income Effect?"). However, to understand how family policies can affect vertical inequality between households, it should also be understood how use and benefits thereof depend on socioeconomic characteristics of other household members. So far, only a small literature on this seems to be emerging.

In recent years, fathers are increasingly given—and have become more likely to make use of—parental leave rights, which is in part related to how the leave policy is implemented (Eydal et al., 2015; Karu & Tremblay, 2017; see Chapter 15 by Bartova and Keizer). Within couples, mothers took longer leave when their (male) partner held traditional attitudes, whereas how long leave he took was not affected by her views. Fathers did take more leave if their partner had a high level of education (Stertz, Grether, & Wiese, 2017). Also in Germany, fathers were most likely to take leave when their partner had a higher level of education than themselves (Geisler & Kreyenfeld, 2011). Moreover, fathers were more likely to take leave when their (female) partner worked full-time and had a high income (Reich, 2011). As such, this dyadic perspective not only brings into focus the question how partners influence each others' use of parental leave and other family policies, but also how their partners' use of family policies affect their own incomes.

The question of how family policy outcomes vary by with whom people live, also needs to include people living without a partner, with single parents being a case in point in this context. The rise of single parenthood has been associated with a rise in household inequality (Zagel & Breen, 2019), but it seems likely that this impact depends on the institutional context (of which family policies may be an important part) that shapes how well single parents are doing financially. When it comes to single mothers, it was found that they use parental leave and childcare to similar extents as mothers living with a partner, but that the use of parental level and childcare was associated with future employment more strongly among single mothers (Van Lancker, 2018b). In Sweden, single fathers were found to continue using parental leave after they had separated from the mother of their child(ren) (Duvander & Korsell, 2018). Both parental leave and child benefits were found to reduce income poverty to a larger extent among single parents compared to among couples, which in the case of parental leave was mediated by single parents' higher levels of employment (Maldonado & Nieuwenhuis, 2015).

The second way in which family policy outcomes are linked to with whom people live, is through how processes of family formation are related to family policy contexts. The most developed literature in this regard deals with the link between family policies and fertility. The relevance of this literature for vertical inequality lies, of course, in the profound consequences of having children for employment, division of labor within households, and the way equivalent income per household member is calculated. Both work–family reconciliation policies (paid leave, childcare) and financial support policies to families with children (child benefits) were found associated with higher fertility in a panel study covering 18 OECD countries from 1982 to 2007 (Luci-Greulich & Thévenon, 2013; similar conclusion in: Chapter 9 by Adema, Clarke, and Thévenon; Diprete, Mogan, Engelhardt, and Pacalova, 2003; Rovny, 2011). However, such associations were generally small and not differentiated by for instance education and income, which would be required for the study of vertical inequality.

Little is known about the degree to which family policies affect homogamy—the degree to which highly educated and/or high-income partners form couples—or hypogamy in which women have higher levels of education than their male partner (Van Bavel, Schwartz, & Esteve, 2018). It has, however, been shown how with rising female labor force participation rates (due to family policy or otherwise), as well as with women's rising levels of education, women's economic position and prospects became more important on the marriage market and resulted in higher degrees of homogamy (Oppenheimer, 1994). Mediated by female labor force participation rates and women's share in household earning, paid leave, and childcare were indeed found to be associated with a higher correlation between spouses' earnings (Nieuwenhuis et al., 2019).

Family policies can further relate to patterns of relationship formation and dissolution. For instance, family policies characterized by a high degree of familization were found to be associated with lower complexity of family formation trajectories—thus with fewer separations—whereas individualization (e.g., childcare policies) was associated with a higher degree of complexity (Van Winkle, 2019).

Third, the focus on with whom people live brings attention to the issue that it does not only matter how much income the household receives, but also who in the household receives the income. Many measures of inequality (as well as poverty) assume that all household income is shared equally. However, it has often been documented that this is not the case, and that women often have less control over household resources than men (Bennett, 2013). As a consequence, income levels (or inequality thereof ) may hide that women experience lower living standards than their partners (Cantillon, 2013). On the one hand, this may mean that mainstream inequality indicators based on household income underestimate real development in the inequality of income or living standards between individuals. On the other hand—and more relevant to the focus of the research agenda developed here—it was shown that in countries where more women have an income of their own, household poverty (as a form of inequality) was lower (Nieuwenhuis, Munzi, Neugschwender, Omar, & Palmisano, 2019).

Although it was shown that in countries with more extensive work– family reconciliation policies (e.g., paid leave and ECEC) a larger share of total household earnings is earned by women (Nieuwenhuis et al., 2019; Stier & Mandel, 2009), and although it seems plausible that the employment facilitated by these policies promotes economic independence, this does not explicitly examine the inequality of control over household earnings. How family policies relate to women's economic independence within households across the income distribution, and how that relates to household-level income inequality, remains open for future investigation.

# **Conclusion**

Women's earnings have a strong tendency to reduce relative income inequality among the households of couples, and through this mechanism work– family reconciliation policies are also related to lower inequality. This means that family policies should not only be considered in analyses of horizontal inequality between women and men (or mothers and women without children), but also as determinants of vertical income inequality between households. The literature on high and rising levels of vertical inequality so far has not been particularly sensitive to issues of gender (in)equality, and it may become so by expanding its focus to issues of household formation and earnings across the income distribution.

This chapter developed a research agenda for examining family policy outcomes with respect to vertical economic inequality, arguing that family policies wrongly have been neglected as a determinant of vertical economic inequality. Three questions are central to this research agenda: who uses family policy, to what income effect, and with whom do people live? These questions are not new, although some have received more attention than others, but when examined in conjunction with, and with attention to, variation across the income distribution, important insights will be gained in the mechanisms through which family policies can be a determinant of vertical economic inequality.

Several other issues in the family policy literature are relevant to this research agenda, but were not integrated as these issues apply more generally. To separate design or implementation of family policy from its use and the income effects of that use, it is important to have family policy indicators that measure the social rights provided by these policies—in line with the discussion on indicators developed in Chapter 24 by Sirén, Doctrinal, Van Lancker, and Nieuwenhuis. Details of implementation might matter substantially for inequality, for instance, with respect to who in the family receives the child benefits (e.g., the household as a whole, or both parents individually receiving half ), or the degree of low-income targeting (Marchall & Van Lancker, 2019). In addition, even though most of the research cited here pertained to national-level family policies, it should come as no surprise in this Handbook that other levels of family policy implementation may matter. Even though it was demonstrated in Chapters 3 by Jenson, 4 by White, and 5 by Razavi that supranational and international organizations do not implement family policies themselves, they do have an interest in reducing economic inequality. Subnational variation in family policy availability may in fact give rise to economic differences between regions. To the extent that rural areas provide more policies to support dual-earner families, this may exacerbate rural–urban income differences. To the extent that the "final availability" (see Chapter 21 by Chung) of family policies in organizations is greater for high-skilled employees and/or those in managerial positions (see Chapter 22 by Begall and Van der Lippe), this also may exacerbate inequality between workers and consequently between households.

Developing an understanding of the mechanisms underlying family policy as a determinant of vertical economic inequality will not only further the literature on family policy outcomes, but also contribute to tackle one of the key challenges of our time: high and rising levels of inequality. As such, family policy might be incorporated in broader analyses of economic inequality. The focus on family policies may also shed light on two more specific challenges of inequality. The first relates to stalled trends in improving gender equality in the labor market (England, 2010), as for instance seen at comparatively high levels of equality in the Nordic countries but also in the United States. In these countries, the question is how future inequality will evolve as it is no longer attenuated by the rise in women's employment and earnings. Here, an understanding of the mechanisms through which family policies affect economic inequality is pertinent, as it may help propose policies that not only raise the number of employed women, but do so in a way that equalizes opportunities and earnings across households. The second relates to the challenges of family diversity in the dual-earner model. On the one hand, family policies such as paid leave and public childcare have facilitated the rise of dual-earner families. This has benefitted gender equality, the income stability of these families, and reduced income inequality among couples, but it may also have raised the living standards in a country (or for instance urban areas). While, for instance, single parents undoubtedly and greatly benefit from such work–family reconciliation policies, the presence of a large number of dualearner families may also make it more difficult for singles and single parents alike to achieve what are considered acceptable living standards (Alm, Nelson, & Nieuwenhuis, 2020).

Family policies have empirically been linked to women's employment and earnings, and consequently to lower vertical income inequality. Yet, the literature also makes abundantly clear that family policies come with trade-offs along the lines of gender and class, as well as Matthew effects. These mechanisms need to be better understood to integrate family policy in analyses of—and recommendation against—high and rising inequality. The challenge ahead is to understand what (combination of ) family policies may be inclusive to a wide range of families across the full width of the income distribution.

**Acknowledgements** This research was financially supported by FORTE: the Swedish Research Council for Health, Working Life, and Welfare (grant number 2018-00988).

## **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **26**

# **Conclusion: The Next Decade of Family Policy Research**

**Wim Van Lancker and Rense Nieuwenhuis**

Family policies are influenced, formulated, or implemented at levels that range from the supra-national, over the national and the sub-national to the level of organizations. At these different levels, family policies serve similar functions that include child income supports, childcare services, parental leaves, leaves to provide care to frail and elderly family members, and support similar goals that include improving children's development and equal opportunities, promoting gender equality, regulating fertility, and stimulating productivity. This has been extensively documented in the chapters of this handbook.

The purpose of this final chapter is to look ahead. We introduce what we believe are five major societal challenges for the future outlook and outcomes of family policies, and reflect on what the handbook teaches us on how to effectively address these challenges, as well as what there is yet to learn. With the latter, we hope to contribute to setting the stage for the next decade of family policy research.

W. Van Lancker (B)

R. Nieuwenhuis

Swedish Institute for Social Research (SOFI), Stockholm University, Stockholm, Sweden

e-mail: rense.nieuwenhuis@sofi.su.se

Centre for Sociological Research, University of Leuven, Leuven, Belgium e-mail: wim.vanlancker@kuleuven.be

# **Levels of Policy Implementation: Globalization and Decentralization**

This handbook has been developed based on the premise that bringing together (insights into) different levels of family policy making provides a relevant and comprehensive understanding of family policies. Processes of globalization and regional integration (such as in the EU) on the one hand, and decentralization on the other, make that increasingly families are subject to policies at multiple levels of (potential) decision making. Family policies are increasingly available and studied across the globe; supra-national and international organizations are increasingly involved in analyzing, recommending, or implementing family policies, while regions and even cities are being delegated—or taking—responsibilities in family policy implementation. All at the same time, organizations play a crucial role in what access workers have to family policy arrangements, either limiting access to or supplementing publicly provided family policy provisions. The challenge for a research agenda that can successfully account for these developments is to develop a clear analytical and conceptual focus that goes beyond the nationstate, as is argued by Zagel and Lohmann (Chapter 6). We highlight a number of themes that emerged from the chapters in this Handbook that work toward such an agenda and conceptual focus.

First, there are clear differences between supra-national and international organizations when it comes to their normative views on family policy. In part, the different positions these organizations take with respect to family policy relates to their origins, mandate, and initial goals and geographic area of competence. Razavi demonstrates in Chapter 5 how there is not "one United Nations," with a labor-centric focus in the ILO, a focus on children in UNICEF and UN Women centering key feminist concerns. UNIFEM (that later merged with DAW to become UN Women) focused on women's economic rights, but operated on the assumption that various other familyand social policies were more relevant for middle- and high-income countries. Jenson (Chapter 3) shows how the OECD and the EU framed family policies in terms of their concern about too *low* fertility and social investment policies to foster growth, whereas the World Bank was concerned about too high fertility globally because of poverty and "limits to growth." We point out here, that although the social investment perspective is now center stage in country-comparative work on social and family policy, it lacks clear analytical focus and misses out on key concepts in the family policy literature, such as unpaid work and gender inequality in paid and unpaid work (Cantillon & Van Lancker, 2013).

A second theme that emerged is that of the power of national governments. International Organizations are crucial to facilitate globalization and affect the interdependencies between countries, by being venues of mutual learning and cooperation, aligning collective interests, to spread norms and practices among member states, but also as a stage of confrontation of norms and influence. Policy ideas diffuse across nations through mechanisms that include competition, coercion, socialization/learning, and emulation (White in Chapter 4). Yet, the diffusion of ideas toward national policies is imperfect, norms and ideas are filtered through domestic mediating factors, including many factors affecting gender equality such as the number of women in parliament, veto players et cetera. In a globalized world, family policy is hardly formulated—let alone implemented—at a global or even supra-national level (Jenson in Chapter 3). Supra-national organizations have little hard power, although, for instance, the World Bank can use soft power by providing conditional loans. Even in the European Union (EU), where member states have delegated some of their legislative power, top-down influence remains opaque when it comes to family policy—although for example the 2019 Directive on Work-Life Balance suggests some change in this regard. Globalization notwithstanding, the national (or federal) level of family policy making remains key to understand family policies. Showing evidence that the multilevel policy levels act as communicating vessels, Parolin and Daiger von Gleichen report in Chapter 18 that in the United States federal-level policies compensated for state-level changes support for families. Although organizations play an important role in setting the "final availability" of family policies to their workers (more on which below), Chung argues in Chapter 21 that the national context in terms of attitudes and policies set the limits and expectation in which organizations and their managers can operate. A similar adaptation to national-level conditions has also been identified in Bardoel's Chapter 23 on multinational enterprises. As enterprises in rich countries became more active in developing countries through acquisitions and mergers, local managers had to adapt their central human resource policy to the local situation.

Thirdly, regional differences and decentralization processes were shown to matter greatly for family policies as well—despite the strong position of the nation-state discussed above. In Chapter 17, Engeman, shows how in the United States, state-level and even city-level developments innovate paid family leave policies with more entitlements than the federal government provides. As a form of bottom-up policy learning, these sub-national policy developments may provide leverage and hope to those who advocate national-level policy adoption. Nonetheless, as long as states have leeway in the way federal grants are spent, as in the case of *Temporary Assistance for Needy Families* (TANF), for instance, income protection for families can be jeopardized because some states pursue other objectives (such as marriage promotion). Great diversity in federalist context is not necessarily for the better for families, argue Parolin and Daiger von Gleichen in Chapter 18. Indeed, local implementations of (family) policy can differ widely from the national/federal policies, and this variation has been shown to have consequences for both the use and the outcomes of different policies (see Chapter 17 by Engeman, Chapter 18 by Parolin and Daiger von Gleichen, and Chapter 20 by Emery). Moreover, as argued by Schober in Chapter 19, even if policies are implemented in the same way, regional differences in norms or economic conditions may result in sub-national variation in the use and outcomes of these policies. This challenges some cross-national analyses, as well as provides interesting insights in the development of family policy over time with a focus on sub-national actors. As discussed by Engeman in Chapter 17, even though it may seem that the window for paid leave in the United States has closed, the policy developments at the state-level may provide an impetus to "breaking the liberal-market mold."

Fourth, a global research agenda should avoid or at least be aware of Western assumptions (Bardoel in Chapter 23) and go beyond the heteronormative family idea, taking due account of families that do not adhere directly to prevailing norms (Evertsson, Jaspers, & Moberg in Chapter 16).

Finally, understanding family policy in this multilevel context poses incredible demands to data. Data infrastructures require major investments and long-term commitments, for which funding seems increasingly scarce (e.g., LIS, 2020; Nelson et al., 2020; Scruggs, 2013). Not only is greater investment in the quality, availability, and comparability of current databases imperative, making further progress requires the development of institutional data infrastructures along the lines developed by Sirén, Doctrinal, Van Lancker, and Nieuwenhuis in Chapter 24. Moreover, this work should not stop at the national level as is typically the case. Similar indicators at the subnational and organizational levels would greatly further the research field as well. In that context, Chung calls in Chapter 21 for more research to genuinely gauge how company-level policies are shaped in the context of national and sub-national policies, and to what extent they are a reaction to existing policies, or are used as a tool to improve productivity rather than to foster work-life balance of workers. This is in particular a plea for gathering more cross-national comparative data on companies of different sizes in different sectors (see also the data presented by Begall and Van der Lippe in Chapter 22), so that a more integrative picture can be sketched of how

international, national, regional, and company-level policies interact with one another, shaping actual work/family decisions and opportunities for families from different socio-economic backgrounds.

## **Austerity and Marketization**

Although large parts of the (industrialized) world in Europe, the OECD and beyond have seen austerity and welfare state retrenchment (Beckfield, 2019; Taylor-Gooby, Leruth, & Chung, 2017), there seems to be no clear evidence that family policies were affected (e.g., Adema, Clarke, & Thévenon in Chapter 9). The European Union, for instance, seeks to expand paid parental leave to both parents, and paid family leave has become more common in some US states or cities (Engeman in Chapter 17). A challenge is to understand exactly why core family policies have expanded while other areas of policy were retrenched in many countries. Specific countries can of course show exceptions to this overall pattern, such as related to the introduction of the Universal Credit in the United Kingdom (Millar & Bennett, 2017). Parolin and Daiger von Gleichen show in Chapter 18 that state-level redistributive programs such as Temporary Assistance for Needy Families (TANF) have become less generous in a number of states, but this has (partially) been compensated by federal programs such as the Supplemental Nutrition Assistance Program (SNAP).

Two insights into this challenge arise from the chapters in this Handbook. First, it is worth noting in this context that supra-national organizations as the OECD and the EU changed their perspective from "family policy as a burden" in the 1980s to "family policies as a precondition for growth" (Chapter 3 by Jenson). Moreover, not all family policies were treated the same. Under the social investment paradigm in European countries, as well as the perspective on *active* social policy more generally, family policies that enable employment and foster human capital are front and center. Policies providing income protection such as family benefits are less popular, and were more subject to cuts under austerity and welfare retrenchment (see also below).

Second, the mode of family policy provision did change, and future research could pay more attention to how marketization is gaining more prominence. Razavi details in Chapter 5 how, for instance, the ILO has tried to maintain a high-road perspective on the public provision of care and regulation of care-sector jobs, while the UN sought more private sector solutions. Razavi identifies this as a risk, for "history tells us that market-based solutions are unlikely to provide the kind of universal social and family policies that can reign in gender, class and other intersecting inequalities." In Chapter 8, Vandenbroeck raises similar concerns about marketization and demand-side financing of childcare provision, as it is associated with lower quality of service provision. Marketization tends to come with budget cuts on staff. Emery shows in Chapter 20 that in the Netherlands (a country with demand-side financing of ECEC) 70% of childcare facilities are now privately run, coming from 30% in 2005. Dykstra and Djundeva analyze in Chapter 14 how marketization in care services for later life families resulted in dualization. Witnessing a shift away from residential care toward home care, this potentially puts pressure on family relations in ageing societies and particularly among those who do not have the means to compensate for inadequate provision of care by purchasing additional care on the private market.

### **Economic Inequality**

Economic inequality has once again taken a center stage position in public and academic debates in recent years. And it is, and will continue to be, highly relevant for family policy as well. Although the employment rates of women have grown closer to those of men over the last decades, the trends toward gender equality have slowed or even stalled in a number of countries (for the US, see England, Levine, & Mishel, 2020), the gender pay gap has not closed (Goldin, 2014), occupational segregation is persistent (Charles & Grusky, 2004), and the different work histories of women and men contribute to gender gaps in old-age poverty (Möhring, 2015, 2016) that although closing, still persist (Doctrinal & Nieuwenhuis, 2019). Levels of wealth inequality (Piketty, 2014) and income inequality (Milanovi´c, 2016) are high and rising within countries, as well as, for instance, in the European Union as a whole (Beckfield, 2019), and trends in poverty in Europe and beyond are "disappointing" (Jenkins, 2020; Vandenbroucke & Vleminckx, 2011). These are not mere concerns of accounting, but they have real-life consequences: The life chances of children growing up in different family forms are considered to be "diverging" (at least in the US, see McLanahan, 2004), more unequal societies impair equality of opportunity (Corak, 2013) and economic growth (OECD, 2015), and growing up in poverty has severe consequences for later life chances—even in rich societies (Van Lancker & Vinck, 2019).

Although these forms of inequality are intrinsically linked, it is useful to distinguish between regional, horizontal and vertical inequality. The distinction between horizontal and vertical inequality introduced in Chapter 3 by Jenson and Chapter 25 by Nieuwenhuis is relevant here to organize some of the key lessons learned. *Horizontal* inequality refers to differences between groups (such as men and women, parents and people without children, or single-parent families and two-parent families), *vertical* inequality refers of the overall differences between households (such as income inequality or poverty rates). Reminiscing the lessons learned in the previous section on the level of policy implementation, when considering how family policies may address horizontal and/or vertical inequalities it should first be recognized that there are also vast geographic inequalities in the availability of family policies.

#### **Regional Inequality**

Regarding the geographic inequalities, the overview of family policy development in Chapter 10 by Filgueira and Rossell underlined the vast inequality across the globe in the fiscal capacity to implement family policies (and hence in the effectiveness of policies). In this context it is relevant to note that the majority of the world's poor (defined in absolute income terms) no longer live in low-income countries but in fact live in countries that by now have evolved into middle-income countries (and within these countries overrepresented in specific, rural regions) (Sumner, 2016). These countries have—at least in theory—some budget to spend on family policies, which opens the question what family policies can mean for the global poor. Indeed, most countries provide rights and entitlements to at least some form of family policy (Heymann & Earle, 2010). Yet, implementation of, for instance, paid maternity leave and conditional cash transfers can be lacking with respect to coverage, eligibility criteria, entitlements, and non-take-up (see Chapter 10 by Filgueira and Rossel).

Chapter 18 by Parolin and Daiger von Gleichen demonstrated inequality in family policy availability, but then with regards to levels and take-up of family policies across states in the US. While living in the same country under the same federal state, this means that similar families have less/more access to better/poorer policies in terms of income, time and services because of the place where they (happen to) live. Schober in Chapter 19 shows similar evidence with regards to ECEC in Germany. So not only inequality across countries or between families within countries matters, but also between regions and states. Even beyond federalist states sub-national variation is relevant. Emery in Chapter 20 shows for the Netherlands that proximity of childcare facilities near one's place of living determines women's labor market opportunities. If childcare places are concentrated in well-off neighborhoods, as is discussed by Vandenbroeck in Chapter 8, this creates new barriers to employment. National policy frameworks with affordable care can still play out very differently across regions, so that once again where you live partly determines your opportunities.

#### **Horizontal Inequality**

Family policies have played a central role in reducing economic inequality between groups that include women and men, and single-parent and twoparent households. Yet, although there is a fair amount of consensus on the overall effects of paid leave and childcare on improving gender equality, little remains of this consensus when it comes to the impact of these family policies on class inequality. As Hook and Li synthesize the literature in Chapter 11, childcare increases the employment of women at any level of education, but the question remains who benefits most. Some studies find that the lower educated benefit more from the provision of childcare, others find the opposite. Clearly, such findings are closely connected to—and have implications for—vertical inequality, as discussed below. For paid leave there is a similar lack of consensus in the literature. How the intersection between gender and class plays out in relation to family policy, may very well depend on additional factors. Hook and Li suggest the type of (coordinated or liberal) market economy, cultural norms, or the overall level of income inequality. Along similar lines, in Chapter 7 Javornik and Yerkes call for analyses of the interplay between different types of family policies, and with other institutional and otherwise contextual conditions. Indeed, numerous outcomes envisaged with family policies can and are also achieved with a broader set of welfare state policies such as minimum income protection, unemployment benefits, housing benefits, labor market regulations, and overall redistribution (Alm, Nelson, & Nieuwenhuis, 2020; Bradshaw, Keung, & Chzhen, 2018; Horemans & Marx, 2018; Verbist, Diris, & Vandenbroucke, 2020).

With respect to class inequality in family policy outcomes, it has often been reported that higher-educated and higher-income parents are more likely to use formal childcare. Moreover, there is also evidence—as summarized in Schober's Chapter 19—that shows that higher-educated and lower educated parents have different *types* of preferences regarding childcare. When selecting a childcare center, higher-educated parents tended to care more about the quality of the care on offer and the pedagogical curriculum, compared to lower educated parents. Native-German parents were less likely to choose childcare centers in which many children from migrant parents were enrolled. The implication of such findings is that childcare policies may not only reduce, but also perpetuate existing socio-economic inequality between groups of parents.

The perpetuation of inequality can also be seen at the company level. Family-friendly policies and flexible working arrangements are not available to all workers to the same extent. Usually high-skilled workers and higherstatus jobs have more access to these company policies. Access is shaped by an individual's real bargaining power (see Begall and Van der Lippe in Chapter 22). Moreover, there is also country-level inequality involved, since there is evidence provided by Chung in Chapter 21 that there is a positive relationship between family policies at the national or sub-national level and the company-level: companies tend to provide more generous policies in countries with more generous family policies. Linking regional and horizontal forms of inequality, it becomes clear that families living in countries with few family policies in place, or family policies disincentivizing women's employment and preserving traditional gender norms in work and care, are facing a double jeopardy.

#### **Vertical Inequality**

A continuing challenge for family policies in OECD countries is to reach those families most in financial need, in order to play a more important role in the reduction of vertical economic inequality—including poverty. These goals and outcomes are clear at various levels of family policy making, as, for instance, demonstrated in the conceptual work on child income protection by Daly in Chapter 2, at the supra-national level as charted by Jenson in Chapter 3, the national level in Chapter 9 by Adema, Clarke and Thévenon and Chapter 13 by Maldonado and Nieuwenhuis, and the sub-national level of the state as demonstrated by Parolin and Daiger von Gleichen in Chapter 18.

Family policies indeed have the potential to reduce (child) poverty, but their effectiveness will be reduced if these family policies are only available in some sub-national areas in a country, are only used by highly-educated or high-income parents, or when companies or managers provide access to family-related policies only if they expect this to benefit the productivity of the workers—along the lines described above. This is where the link between horizontal and vertical forms of inequality in terms of family policy outcomes becomes apparent. This link was further developed by Nieuwenhuis in Chapter 25, arguing that there is some evidence that paid leave and childcare policies facilitate women's employment to such extent that this helps reduce vertical economic inequality among the households of couples. Yet, to fully understand how family policies may affect vertical income inequality, it was argued, requires not only to consider the income effects of using family policies but also to consider who uses family policies and with whom they live. Although the argument focused on national-level policies only, in the context of this handbook this immediately brings into focus to other levels of policy making. From this perspective, sub-national variation in the availability of family policies may not only affect economic differences within regions (or states) but also between. Who uses family policies is also determined by the "final availability" determined by organizations. This can, for instance, relate to organizations restricting access for (some) workers to flexible working arrangements, to workers avoiding to use their legal entitlements in anticipation of repercussions, or companies providing workers with more generous family policy arrangements than publicly provided or mandated. The extent to which organizations provide or limit access to worker's access to such policies, and whether they do so following a socio-economic gradient, may be an important additional mechanism shaping vertical economic inequality.

# **Changing Family Relations**

The diversity of family forms vastly exceeds that of the number of idealtypical, model family types that family policy makers implicitly or explicitly have in mind. Family configurations consist of a wide range of interdependencies among family members, that need not be based on kinship, need not live in the same home, and that are subject to change over time (Widmer, 2010). The ideal-typical "nuclear family," consisting of a married husband and wife with dependent children, by no means describes the reality of a majority of families (UN Women, 2019). The capability approach, as developed in Chapter 7 by Javornik and Yerkes and Chapter 19 by Schober, provides a framework that is inherently sensitive to a diversity of life-courses and to family diversity. In its recognition of individual's agency as socially embedded, it helps explain how family policies can have widely different consequences for different individuals or families. The challenge for family policy makers and scholars alike is how to implement family policies that support this wide range of families. This resolves around a number of issues, including the definition of family types, whether or not different family types require specific policies, and solidarity among family types.

The importance of definitions was raised prominently by Evertsson, Jaspers and Moberg, who in Chapter 16 introduce the concept of parentalization to address the issues of who can become a parent and how, and who can share in the care of the child. Applied to female same-sex couples, this chapter highlights the importance of the definition of concepts as "parents." Even in countries that are typically considered to be rather liberal, laws, policies and ideas about who can become a parent (in a variety of interpretations of "parent" that include the social, the legal, and the biological) lag behind the reality of couples who want to become a parent and want to be able to take upon themselves all the responsibilities that come with parenthood. Family policies enacted in law can deliberately exclude certain types of families or favor one particular type of family. In reality, many parents who do not adhere to the gender norm of the heteronormative family are less or not entitled to parental leave schemes or face steep barriers to become parents in the first place. A number of developments, including extending IVF and adoptive rights to same-sex couples and expanding the number of legal parents a child can have, work toward more inclusive notions of parenthood. Here, it is to an important extent the changing of the definition of a "parent" that includes more people in the reproductive, social, and family rights that were already enjoyed by others. In the area of care (leave) for frail and elderly family members, similar debates arise regarding who is considered "family" (Ivanova & Dykstra, 2015).

The issue whether there is a need for group-specific policies was addressed in the two chapters on single parents. Chapter 12 by Skinner and Hakovirta discuss child support policies that were designed to specifically address the needs of (children growing up with) separated parents. Child support arrangements often represent long-term commitments, as payments are due typically until children reach adult age. During this time, a lot can change in the lives of both separated parents, in terms of, for instance, income and employment, re-partnering, and having more children (with other partners). Over time, family relations have been changing away from the traditional breadwinner model. An important challenge for (the administration of ) child support systems is to adapt to these new situations—sometimes this challenge seems nigh insurmountable (Meyer, Skinner, & Davidson, 2011). They show that only in a minority of countries child support systems adapted to reflect changes in maternal labor market participation, e.g., by taking account of mother's earnings in calculating child support amounts, or acknowledge the role of fathers in childcare. In addition, as the chapter concludes, the goals of promoting gender equality and of child income support can be competing. This analysis is complemented in Chapter 13 by Maldonado and Nieuwenhuis with the argument that single parents often benefit from policies that are aimed at all families with children—not just for single parents. So, instead of child support, this chapter examined whether single parents benefit just as well as two-parent families from child income supports, childcare and paid parental leave. The results for family benefits were unequivocal: when it comes to poverty reduction, single parents benefit more from family benefits than two-parent families. Other work argued that the poverty reduction associated with family benefits can exceed those of child support (Nieuwenhuis & Maldonado, 2018). With respect to parental leave and childcare, the results are less clear-cut: single parents receive similar or slightly higher payments during parental leave and pay similar or slightly lower fees for childcare. However, expressed relative to their household income the replacement rate of parental leave is lower, and the childcare fees higher, for single-parent families compared to two-parent families. While single parents are in pressing need of reconciliation policies if they want to work, those policies are less affordable for them.

The third issue, solidarity among family types, can best be illustrated with chapters on care for elderly and frail family members. With increasing longevity and population ageing, it is increasingly important to also consider care relations between (adult) children and their elderly parents, and policies that can support this form of care as well. In Chapter 14, Dykstra and Djundeva chart such policies for long-term care (LTC), and it becomes clear that such policies can comprise of a combination of providing care directly to the elderly (elderly homes, home care) and providing support for the family care giver (e.g., in the form of leave, or awarding pension credits to care givers). Nevertheless, despite the coordinating efforts of the European Commission in the *Pillar of Social Rights*, many welfare states (in particular in Southern and Eastern Europe) fail to ensure adequate care for their elderly and frail citizens. Witnessing a shift away from residential care toward home care, this puts pressure on family relations in ageing societies. In-kind policies such as residential care come with different trade-offs compared to cash-for-care schemes, for instance, with respect to refamilization and gender equality (in both care and work). Dykstra and Djundeva call in Chapter 14 for policy evaluations that "*cut across* policy domains". This issue of cross-domain evaluations relates to the starting point of Birnbaum, Ferrarini, Nelson, and Palme (2017), who showed that welfare states providing similar levels of financial support to children, working-age people, and the elderly tended to have higher levels of overall support compared to countries that focused their financial support on only one age group. Such findings are important to advocates for family policy as well, as shown by Engeman in Chapter 17. Advocates for paid family leave in the United States made sure not to frame their policy proposals as aimed at care for a specific group, or to support women, but instead anticipated greater support if their proposals were formulated in a gender-neutral and universal frame to provide care for all family members—irrespective of their age.

# **Gender Revolution: Adapting to Women's Empowered Roles?**

Developments across OECD countries in terms of family policies helped reduce gender inequality, and improved fertility rates (or rather, perhaps, slowed the decline in fertility). Although the initial rise in female labor force participation was linked to fertility decline and relationship dissolutions, the stagnation or even reversal of these trends was linked to societies—and in particular men—adapting to the norm that women are highly-educated and have empowered roles (Esping-Andersen, 2016; Van Bavel, Schwartz, & Esteve, 2018). Nonetheless, the care for children, as well as for elderly parents (and other family members) is predominantly shouldered by women, which also has important implications for their working lives and later their own retirement income (Dykstra and Djundeva in Chapter 14). It is a challenge to design and provide family policies that adequately promote gender equality in terms of labor, care, and leisure.

Adult worker models that are often used to theorize comparative family policy research, usually lack a focus on gender equality in the labor market as well as in unpaid work—as detailed by Zagel and Lohmann in Chapter 6. To be able to think about changing gender and class relations simultaneously, it is important to further develop concepts and theories to examine asymmetrical and heterogeneous policy effects (also see Chapter 11 by Hook and Li).

Family policies, and welfare states more generally, have been struggling to adapt to changing gender relations and women's empowered roles, which was detailed in numerous chapters of this handbook. We outline three challenges. In many countries paid parental leave for fathers continues to lag far behind parental leave provisions to mothers (Adema, Clarke, and Thévenon in Chapter 9), but welfare states are in fact adapting their parental leave policies to also allow—and encourage—fathers to take parental leave. Fathers' taking of leave can in itself be considered as a form of gender equality, and it may foster other forms of gender equality that reduce income differences within households, and may represent more gender-equal role models for their children. Still, in Chapter 15 Bartova and Keizer show vast differences between countries, in terms of the duration of leave that fathers can take, the level of wage replacement, as well as in which way leave is provided to fathers (e.g., non-transferable leave, or individual entitlements or family entitlements). Individual, non-transferable leave for fathers was found to be the most effective policy design to encourage fathers to take parental leave.

A second challenge for welfare states to adapt to women's changing roles is also clear with respect to child support systems. Skinner and Hakovirta show in Chapter 12 that child support systems effective in reducing child poverty are not always the ones in which gender equality is taken on board—as was discussed in more detail in the previous section. In systems which are based on a male breadwinner model, the income of the mother is usually not taken into account for calculating support amounts. This runs against ambitions to promote gender equality, while these systems are more effective in reducing poverty.

Also related to single parents, and not covered in this handbook, is how the rise of shared residence also represents changing gender relations. Shared residence is the practice that children continue to live about equal amounts of time with both their parents after they separated (Fransson, Låftman, Östberg, & Bergström, 2018). To the extent that separated fathers and mothers are both actively involved in the care for their children challenges the notion that most single parents are mothers (Nieuwenhuis, 2020). Shared residence seems to be on the rise in a number of European countries, and there is evidence to suggest that this benefits the well-being of these children (Baude, Pearson, & Drapeau, 2016; Nielsen, 2014), although it should be acknowledged that parents doing shared residence tend to be a selective, rather well-resourced group and that the shared residence living arrangement is relatively unstable (Poortman & Van Gaalen, 2017). Shared residence is to be understood as relationships between (individuals in) multiple households, and therefore typically not captured well in large-scale surveys—including those used to create the indicators that policy makers rely on. Yet, it is important to better understand the driving forces and outcomes of shared residence, and which policies might promote it and stimulate positive outcomes.

A third (set of ) challenge(s) pertain to the observation that welfare states adapting their family policies to changing gender roles may not be enough. The role employers and organizations can play in this respect, is not yet fully understood. On the one hand, Goldin (2014, p. 1091) argued that employers have a central role to play in the "last chapter" of gender convergence in pay: a sizeable part of the gender pay gap in high-paying occupations in the

United States is related to working conditions that favor very long work hours and are inflexible with respect to when and where the hours are worked. On the other hand, based on data from multiple countries, Chung discusses in Chapter 21 how family-friendly arrangements and flexible working at the company level might actually contribute to stronger patterns of inequality in work and care. In a flexible environment, men are more likely to increase their overtime hours while women are more likely to increase time spent on household chores and care work. These outcomes of workplace flexibility are shaped by gendered norms and expectations attached to gender (as well as to level of education, as shown by Begall & Van der Lippe in Chapter 22). The flip side of this, of course, is that in countries with more egalitarian gender norms, the effect of flexibility will be different as well. This goes to show that gender inequalities are likely to be reproduced at the organization level if dominant country norms are not challenged. At the same time, flexible working and family-friendly working arrangements might reduce gender inequality in the labor market. When women are able to retain control over their working time, this might reduce the need for transitioning into part-time jobs, which come with wage and career penalties, and the gender wage gap tends to be smaller in companies with more family-friendly policies, in particular flexible working arrangements. Multinational enterprises often operate in contexts with very different gender norms and expectations about work and family, resulting in tensions between (typically) the Western notion of work-life balance and national norms, practices, and challenges. Bardoel (in Chapter 23) presents ample examples of the very different types of support human resource managers provide to their workers in different parts of the world. Being able to adjust to these very different needs in the context of an enterprise that operates in multiple countries requires the recognition and understanding of tensions. Such tensions can be distinguished along dimensions that include strategic vs. operational concerns, centralization vs. decentralization, or institutional versus contextual awareness. The need to addressing such tensions effectively and based on evidence, raises new questions of human resource management in multinational enterprises.

# **The Next Decade of Research: Family Policy in Extraordinary Times**

Much of the research on family policies examines how and why family policies change over time, and how changing family policies (or differences in policies across countries, regions, or organizations) are related to a variety of outcomes for a variety of families. Much less, however, is known about the development and effectiveness of family policies when societies undergo rapid change. What does the evidence generated in *ordinary times* tell us about the role of family policies in *extraordinary times*? At the time of writing this conclusion, early May 2020, the rapid spread of the COVID-19 pandemic has had a massive impact on societies across the globe. It is too early to write anything definitive, but too late not to write anything. Countries almost universally shut down large parts of their economies to reduce the spread of the pandemic and to prevent healthcare systems to become overwhelmed. Marked increases of all-cause mortality were visible almost everywhere, in particular among the elderly (EuroMOMO, 2020). Quite possibly, at the time of reading, it has become clear that this was only the beginning. But the current crisis reminds us of the recurring nature of crises: the 2008 financial and economic crisis ("Great Recession"), the 2015 wave of migration in European countries, partly caused by violent conflicts in Syria, Iraq, Afghanistan and other countries, and concerns about the consequences of climate change are all examples of crises that can alter the course of societies. The very Swedish model was motivated by concerns about emigration and low fertility, in a book famously titled "Crisis in the population question" (Myrdal & Myrdal, 1934). Although we never know when crises hit, we can prepare for the eventual next one. Yet, very little seems to be known about the fundamental question of whether and how family policies function in times of societal upheaval, and for whom. The five challenges we put forward in the previous section can give us direction on how exogenous shocks, crises, and in particular their economic consequences, raise pertinent questions for the next decade of family policy research.

The evidence collected in this handbook suggests that supra-national organizations have had a limited direct influence on the making of family policy, although OECD, ILO and UN Women all provide analyses and recommendations, and the EU has had tangible impact on member states' family policies. While budgetary restrictions in the Eurozone have been associated with austerity (Beckfield, 2019), in times of societal upheaval and economic crisis, organizations like the EU can provide financial support to member states to keep their economies afloat. The Great Recession of 2008, for instance, was followed by European coordinated stimulus policies to absorb the shock and to avoid the Eurozone from collapsing in the short-term. In the longer-term, however, a straightjacket of fiscal austerity was imposed on several countries by the European Commission, the European Central Bank and the International Monetary Fund. This led to austerity measures being implemented in many member states, affecting some family policies as well. Most affected were cash transfers such as family and child allowances. Following the crisis, in countries such as Greece, the Netherlands, the UK, Hungary but also Finland, cuts in benefits were implemented under fiscal consolidation measures including a freeze of benefit levels, tighter eligibility conditions, the abolition of tax breaks, or actual reductions of child benefits (Richardson, 2010; Thévenon, Adema, & Ali, 2014). The effect of the 2008 crisis in terms of the labor market as well as the austerity measures put in place in the aftermath affected children particularly hard (Cantillon, Chzhen, Handa, & Nolan, 2017; Chzhen, 2017). However, as family policies are still the prerogative of EU member states, little is known about how a more coordinated approach toward family policies would affect their capacity to cope with shocks of various kinds. While family transfers were subjected to cuts, the effect on leaves and particularly childcare services was different. Usually, pre-crisis reforms were carried out as planned, and childcare is a policy area to which governments increasingly devoted public resources (Van Lancker & Ghysels, 2014). However, here too, the challenges for research we identified are relevant. An increase in spending is meaningless as such; what matters is how public budgets are spent. Concern has been raised that the 2008 economic crisis has put additional pressure on public budgets, accelerating a process of marketization of public provision of care services, including long-term care and childcare services. This raises the question what the consequences might be for the generational conflict (Birnbaum et al., 2017). For instance, in the face of budget constraints, how will public care for children be prioritized relative to care for people later in life? Marketization of public service provision was linked to a dualization between those who can afford services and who cannot, ranging from the provision of childcare to elderly care. Yet, little is known about whether the crisis exacerbated existing processes of dualization to a larger extent in marketized systems of provision compared to public provision of (care) services.

With respect to decentralization, for instance, municipalities providing public services to families, a concern is whether regional variation in the provision of services as well as in the degree to which municipalities are affected by crises, translates into—or exacerbates—regional inequality with respect to access to services and how well people are protected by these services.

A primary concern regarding (the economic consequences of ) crises is rising inequality. Social policies have been described as *automatic stabilizers* in times of economic downturn. This has most notably been described in the context of unemployment insurance, that automatically stabilizes incomes of workers and their families (at least to some degree and typically for a limited period of time) even when unemployment rises rapidly. The benefit is not only to the families receiving the benefits, but also to the economy at large because unemployment benefits stabilize purchasing power and demand for goods, thus helping to prevent further collapse of companies. Family leave policies, including the ability to take time off from work for own illness or to care for a family member, have been described in a similar manner (Boushey, 2016). For instance, at the time of writing, many countries including Slovenia, Belgium, Finland and Poland had implemented or expanded leave policies for parents who cared for young children during the COVID-19 lockdown measures. It remains to be seen whether these expansions will be temporary, or lead to structural adjustments of leave policies in these countries. Child benefits are not an automatic stabilizer per se (as they are always provided, not only in times of economic turmoil), but are wellknown to be highly effective to help reduce poverty in large parts of the world and among a wide range of families with children (see, for instance, Chapter 9 by Adema, Clarke, & Thévenon, Chapter 10 by Filgueira & Rossel, and Chapter 13 by Maldonado and Nieuwenhuis). We have yet to learn how automatic stabilizing mechanisms provided by family policies in combination with other income protection policies will hold up in the face of (health- and economic) crisis. It will be an important question to examine how their presence may help families as well as societies endure the pandemic, equalizing risks, and hasten recovery.

Greater disruptions might be expected when it comes to (public or private) service provision. This includes—but is certainly not limited to—childcare, schooling and care for elderly. In relation to distancing measures, many childcare facilities and schools have been closed. Otherwise, services provided by the company (see Chapter 21 by Chung for examples and how common such company-level services are across European countries) are tied to employment, and workers lose access when they are unemployed or when the company goes bankrupt. More generally, it was found that in times of economic crisis, managers put more emphasis on whether the company might benefit from providing company-level services, rather than on the needs of the worker (Been, Den Dulk, & Van der Lippe, 2016).

Childcare and other forms of education serve an important function in equalizing development and equal opportunities in children. From research on long holidays we can learn about potential consequences of long closures of childcare facilities and schools (Campbell, Watson, & Watters, 2015), and the evidence strongly suggests that long closures will exacerbate inequality in the skills, knowledge and development of children with different socioeconomic background. Moreover, schools also provide supervision on other aspects of well-being and (social) safety, and support in the form of school meals. As such, long closures of childcare and schools—or parents losing access to company provided care services—are a risk factor for heightened inequality (Van Lancker & Parolin, 2020). Family policy researchers should do their part examining to what extent such arising inequalities have been most effectively prevented—or remedied—by different modes of (for instance) childcare provision. Horizontal inequalities in terms of ageing and caring are also highly relevant to study in relation to crisis. The spread of COVID-19 infections is particularly dangerous for the elderly. How families live and care together, and to what extent welfare states provide adequate care for frail elderly may prove to be relevant factors to understand why and to what extent some countries are more severely affected by the pandemic than others. As such, how families and welfare states are able to absorb shocks is determined by how welfare states organize care relations.

From the perspective of family diversity, it should be recognized that some family forms might be in a better position to deal with (the consequences of ) a crisis than others. Therefore, the importance of family policy may also vary. From the perspective of family policies, the potential consequences of a crisis are myriad. How will the closure of childcare facilities in response to a crisis, for instance, affect the challenge for single parents to combine work and family responsibilities? How are child support payments affected if parents lose their jobs and are no longer able to make the payment how will this affect family relations? How do separated parents (re-)negotiate shared parenting arrangements during societal upheaval? How effective can care regimes for later-in-life families that rely on family members providing care operate when the financial or health situation of these family members changes rapidly? What will be the consequences of closing down (or no longer financially supporting) IVF centers—even if only temporary—for the ability of a wide range of families to become a parent? It should finally be recognized that family policies might not be enough for all families: lowered levels and stricter eligibility criteria in unemployment insurance benefits have rendered families without a second earner in the household into a new risk group for poverty (Alm et al., 2020). As such, the capacity of unemployment benefits to act as an automatic stabilizer as discussed above may be inadequate for different types of families in times of crisis.

Finally, it will remain to be seen—and extensively studied—what direction the gender revolution will take in times of crisis, and what role family policy can continue to play here. Women spend more time than men on care work, and when childcare services (or schools) are not available it is a real concern that this gender inequality increases. It has been shown that fathers who took parental leave are more involved with the care for their children later on (Duvander & Jans, 2009): it will be interesting to see whether this (change in) behavior persists in times of crisis. Evaluations of the (austerity following) the 2008 financial crisis showed that particularly the public sector was hit by a reduced number of jobs and wage cuts. Since more women than men work in these sectors, they were more likely to be affected by these measures (Rubery, 2015). As it is an ongoing debate whether family policies reduce or increase class-based inequality, including occupational segregation, this raises the issue whether current constellations of family policies facilitate women's activity in sectors that are more vulnerable during and after crisis and other forms of major societal change.

As policies to support families in terms of work, care, leisure, and incomes evolved and developed across the world, from companies, over cities and regions, to nation-states and supra-national and international organizations, a clear research agenda for the next decade emerges. Many societal challenges lie ahead of us, of which we identified five in this concluding chapter, and how family policies develop in the future will affect how these challenges unfold and affect families. In addition, we identified that there are clear gaps in our knowledge on how to adequately support ordinary families in extraordinary times. We sincerely hope our handbook will prove to be an anchor point, synthesizing what we know while contributing to the research agenda on what we need to know.

### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Index**

#### **A**

access/accessibility 11, 17–19, 31, 36, 37, 49, 50, 52, 57, 60, 62, 73, 80, 95–97, 101, 104–107, 142, 144, 147, 149–152, 154, 155, 157, 160, 170–172, 174, 177–179, 181, 182, 198, 204, 205, 210, 211, 228–230, 233, 240, 241, 243, 257, 282, 283, 323, 354, 357, 359, 379, 398–402, 409, 415, 418, 422, 432, 434, 439, 441, 446, 449, 451, 453, 466–468, 472, 476, 487, 495–497, 499, 502, 503, 511, 512, 518, 521, 528, 529, 536, 540–543, 553, 555, 557–559, 561–564, 575–581, 583, 585, 588, 590, 594, 596–599, 609, 610, 630, 631, 640, 643, 646, 661, 684, 689, 691, 692, 699–701 accommodation model of childcare choices 17, 486, 494, 496, 502 adoption 7, 52, 70, 75, 77, 78, 80, 90, 92, 95, 309, 402, 403,

407–410, 415, 416, 418–420, 445–447, 450, 495, 635, 685 joint 399, 403, 407, 409 second-parent 403, 407, 411, 416–418, 421 adult worker model 126, 127, 299, 516, 695 affordability 101, 106, 149, 151, 152, 154, 155, 158, 160, 179, 182, 205–207, 256, 317, 441, 495, 511, 512, 606, 632, 633, 636, 640, 643–645, 647 Africa 61, 92, 104, 221, 233, 234, 236, 604, 610, 611, 614, 640 ageing in place 358, 359 agency 6, 38, 88, 90, 96, 105, 142, 144, 145, 148, 275–278, 280, 283, 296, 297, 495, 496, 556, 607, 630, 646, 692 algeria 233 angola 235 Argentina 237, 238, 240 Asia 47, 106, 231, 242, 605

austerity 21, 94, 98, 102, 103, 111, 657, 659, 660, 687, 698, 699, 702 Australia 8, 78, 80, 200, 202, 205, 208, 210, 228–230, 268–271, 274–276, 281, 290, 291, 293, 295–298, 313, 317, 432, 513 Austria 30, 32, 33, 202, 210, 255, 334, 336, 432, 636 autonomy 49, 70, 76, 107, 119, 122, 143, 150, 220, 333, 335, 578, 581, 585, 597 working time 537, 540, 541 availability 5, 18–20, 36, 101, 123, 141, 147, 149, 150, 152, 154, 160, 172, 206, 226, 241, 256, 336, 353, 447, 493–496, 498–501, 504, 512–515, 518, 519, 522, 528, 536, 576

#### **B**

*Babies and Bosses* 56, 57, 63 Bahrein 235, 236 ballot measure 440, 446, 451 Baltic 202 Belgium 30, 32, 33, 35, 170, 177, 178, 197, 199–201, 203, 208, 225, 268–271, 275, 277, 280, 284, 290–292, 352, 356, 432, 500, 501, 542, 700 belize 239 benin 235 between-country differences 581 Bolivia 238, 239 botswana 235 Brazil 239, 240, 663 Bulgaria 319, 582, 636 Burkina Faso 234 burundi 235 business case 102, 103, 559, 615

**C** cabo verde 235 California 180, 181, 209, 436, 437, 439–443, 446, 447, 453, 470, 472 Cameroon 235 Canada 8, 75, 80, 180, 200–202, 209, 226–228, 230, 311, 432, 449 capabilities 107, 142–154, 157–161, 241, 243, 494–496, 500, 501 capability approach 17, 142–149, 160, 486, 494–496, 502, 692 capital 52, 61, 62, 79, 80, 100, 123, 255, 543, 659, 687 care credits 333, 352, 353, 356 giver 70, 101, 105, 120, 130, 133, 135, 250, 305, 332, 338, 355, 369, 401, 450, 486, 610, 614, 660 ideals 422, 485, 492, 494, 502 leaves 35, 319, 320, 333, 338, 352, 370, 377 receiver 120, 133, 135 caregiving 16, 104, 132, 258, 305, 333, 353–357, 359, 370, 372, 387, 433, 434, 436, 441, 444–449, 452, 453, 560, 610 risks 355 Caribbean 237 case-oriented approach 124 cash benefits 37, 93, 203–206, 228, 229, 231, 233, 242, 332, 334, 338, 355, 357, 358, 417, 461, 468, 633 cash for care 15, 333, 335, 338, 353, 356–358, 449, 488, 666, 694 categorical concept 120, 124, 136 central african republic 235 central europe 222 chad 235 child benefits/family benefits 9, 20, 30, 32, 33, 89, 92, 97, 101,

102, 106, 108, 204, 229, 231, 288, 303, 304, 306, 307, 310–317, 322, 323, 399, 401, 631, 633, 666, 669, 670, 672, 687, 694, 699, 700 childcare 149 accessibility 149, 151, 154, 157, 495, 529 affordability 160, 495 availability 150, 172, 493, 495, 499, 513, 518, 522 fathers 15, 157, 272, 310, 370, 371, 375, 380, 387, 389 flexibility 149, 153, 159, 160, 495, 517, 518 governance 640, 643, 645 quality 11, 31, 52, 97, 103, 147, 159, 177, 308, 500, 502 childhood 4, 14, 25, 29, 36, 37, 53, 56, 61, 99–103, 106–109, 175, 183, 205, 206, 223, 224, 241, 242, 291, 308, 463, 479, 493, 519, 611, 661 a good childhood 36 childminder 18, 486, 511–513, 516, 517, 519, 527, 528, 635, 636, 640 child support/child maintenance 14, 49, 211, 229, 267–270, 273–276, 280, 290–299, 303, 322, 693, 694, 696, 701 Chile 96, 101, 104, 198, 202, 203, 208, 237–241, 319 Civil Rights Act of 1964 443 class 5, 21, 111, 119, 125, 128, 129, 144–149, 151, 157, 158, 160, 170, 172, 253, 255, 256, 324, 353, 561, 673, 690 climate change 102, 452, 698 coercion 71, 72, 76, 108, 177, 685 cohabitation 89, 197, 420 Colombia 238–240 commodification 132, 160, 177, 178, 181, 629

company level 7, 535, 536, 542, 543, 552–554, 557, 563, 691, 697 comparative methods 123, 269, 273, 274 comparative welfare state research 119–121, 123, 125, 128, 633 competition 4, 63, 71, 81, 170, 553, 610, 685 Conditional Cash Transfers (CCTs) 61, 238, 689 Congo 235 Contradictory family policy model 129 conversion factors 144–146, 148, 149, 157, 160, 495, 496, 499–502 Costa Rica 237, 239 côte d'ivoire 235 county 17, 432, 485 covid-19/corona 698, 700, 701 crisis 96, 98, 103, 110, 170, 201, 233, 237, 698–702 Croatia 179, 182 culture 72, 148, 195, 258, 537, 554, 604–606, 609, 614, 616, 617 work-care culture 503 work culture 554 custody 49, 200, 272, 284, 295, 296, 298, 303, 399, 409 joint 272, 288 Cyprus 179, 338, 540, 636 Czech Republic 202, 204, 207, 336, 636

#### **D**

data 13, 15, 17–21, 33, 47, 50, 51, 56–58, 61, 73, 76, 108, 123, 133, 141, 149, 150, 152–155, 161, 172, 196, 198–200, 203, 220, 228, 230, 237, 239, 250–257, 260, 268–270, 273, 274, 280, 291,

293, 295, 298, 304, 310, 311, 315–317, 319–321, 324, 325, 336, 337, 354–356, 359, 369, 370, 377–380, 383, 384, 386, 410, 412, 420, 453, 466–468, 471, 486, 498, 499, 503, 504, 511–515, 517, 519–524, 528, 529, 536, 538–540, 554, 555, 558, 563, 576–578, 580, 582, 588, 598, 608, 627, 628, 631–637, 639, 640, 643–647, 665, 686, 697 Database for Institutional Comparisons of Economies (DICE) 156, 639, 640, 643, 644 decentralization 16, 21, 28, 465, 490, 492, 501, 608, 609, 612, 684, 685, 697, 699 decommodification 10, 123, 124, 129, 130, 400 dedomestication 130, 132, 133, 135 defamilialism/defamilism 160 defamilization/defamilialization 9, 10, 124, 125, 129–136, 400, 401, 419, 629, 634 degenderization 130, 132–134 demotherization 130, 132, 134, 135 Denmark 35, 169, 178, 182, 197, 201–205, 208, 225, 268–272, 275, 276, 280, 282, 291, 298, 313, 337, 352, 355, 356, 374, 398, 403, 407–413, 415–417, 419–421, 432, 539, 541 development 4, 5, 7, 11, 12, 16, 20, 26, 31, 35–38, 45, 47, 53, 56–61, 64, 73, 78, 80, 91, 99–101, 103, 104, 106, 109, 110, 121, 123, 145, 150–153, 158, 170, 172–176, 183, 194, 204, 205, 211, 220, 221, 230, 235, 236, 242, 295, 307, 319, 335, 337, 357, 370–372, 374, 387, 388, 398, 432–434, 436, 437, 441, 444, 445, 451, 452,

472, 487, 488, 503, 511, 529, 582, 610, 614, 628–630, 633, 634, 644, 645, 657, 659, 670, 683–686, 689, 693, 698, 700 different-sex couple 401–403, 407–409, 414, 420 diffusion 48, 51, 70–72, 81, 453, 555, 685 discrimination 49, 51, 62, 63, 75, 77, 79, 93, 108, 252, 253, 255, 370, 412, 420, 435, 443 anti-discrimination laws 437 distance 18, 153, 179, 334, 517, 520, 522, 523, 526, 527 diversity 5, 18, 29, 76, 89, 108–110, 176, 178, 183, 274, 369, 437, 459, 465, 468, 472–474, 476, 478, 479, 518, 524, 539, 600, 609, 611, 612, 615, 617, 672, 686, 692, 701 divorce 49, 195–197, 199, 220, 235, 237, 278, 296, 298, 309, 334, 371, 399, 420, 524 Djibouti 235 doing gender 660 domestic policymaking 4, 7, 8, 27 Dominican Republic 237, 239, 240 donor 96, 106, 107, 234, 407, 408 anonymous 407, 408 known 407, 408 non-anonymous 408, 418, 422 unknown 407 dual-carer 15, 251, 257, 308 dual-earner 79, 81, 128, 129, 141, 153, 158, 196, 203, 205, 251, 257, 259, 308, 317, 318, 320–325, 553, 672, 673 caregiver model 489, 660 support model 128 dualization of care 358

Early Childhood Education and Care (ECEC) 11, 12, 25, 28, 34–36, 53, 57, 61, 63, 80, 100, 101, 103, 169–183, 194, 195, 201, 204–207, 211, 219–221, 225, 226, 228, 230, 240–243, 304, 305, 307, 308, 310, 317, 318, 323–325, 485–491, 493, 494, 497, 499–501, 628, 640, 643, 644, 661, 671, 688, 689 earner-carer model 257, 270 East Asia 196, 220, 231, 232 East Germany 488, 491, 498–500 Ecuador 101, 238, 239 education 17–20, 26, 34, 37, 47, 48, 53, 56–58, 60, 61, 63, 64, 80, 89, 97, 104, 106–109, 142, 145, 147, 150, 153, 157–160, 169, 172, 175, 176, 180, 182, 196, 199, 201, 204, 205, 228, 257, 305, 356, 463, 485, 486, 493, 498, 501, 503, 524, 537, 556, 582, 587, 590, 594, 596, 597, 635, 644, 666, 670 educational attainment 15, 172, 196–198, 201, 249, 255, 257–259, 261, 475, 477, 500 educational gradient 305, 576, 578, 579, 581, 587, 590, 596–599 Egypt 236 El Salvador 239 employee benefits 575, 597 employment 4–7, 9, 10, 13, 14, 17, 18, 20, 21, 25–27, 31, 32, 34, 37, 47–58, 60–64, 78–81, 97, 99, 101, 103–105, 127–129, 142, 146, 150, 151, 153, 158, 171, 179, 194–196, 201, 202, 205, 211, 228, 229, 233, 236, 250–261, 269, 270, 272, 295–298, 304, 305, 307–309, 311, 316, 323, 324, 334, 338,

356, 374, 400, 437, 443, 447, 461, 463, 466, 468, 474, 475, 486, 488, 492, 495–500, 502, 503, 511–513, 519, 608 emulation 71, 72, 685 endogamy 662 equatorial guinea 235 Eritrea 235 Establishment Survey for Working Time and Work–Life Balance (ESWT) 538, 539 Estonia 35, 178, 197, 199, 200, 268–271, 275, 277, 280, 284, 290, 294–296, 298, 313, 323, 336, 337 Ethiopia 235 Europe 15, 25, 29, 31–34, 37, 47, 50, 52, 77, 81, 127, 170, 172, 223–226, 230, 238, 242, 274, 305, 311, 332, 333, 338, 354, 356–358, 370, 387, 388, 514, 535, 536, 538, 541–543, 558, 634, 687, 688, 694 European Directive on work-life balance 332 European integration 658 European Union (EU) 4, 7, 15, 26, 27, 33, 35, 36, 45, 46, 48–55, 62, 63, 69–71, 73, 77–79, 81, 161, 175, 221, 225, 228, 237, 256, 310, 334, 338, 379, 418, 423, 468, 634, 639, 647, 685, 687, 688 European Union Statistics on Income and Living Conditions (EU-SILC) 632, 635–637, 640, 644 European Working Conditions Survey 540 Eurydice 35, 149, 153, 175, 640, 643, 644

**F** familialism/familism 10, 124, 131, 134, 135, 142, 154, 250, 259, 488, 489 family multi-generational 331 multi-parent 399, 422 relations 14, 21, 29, 50, 53, 56, 130, 219, 325, 520, 576, 634, 688, 692, 693, 694, 701 separated 14, 267, 272, 273, 298 structure 5, 29–31, 38, 51, 55, 221, 323, 334 The FAMILY Act 433, 453 well-being 3, 28, 121 Family and Medical Leave Act (FMLA) 431–436, 439, 440, 444, 445, 449, 451, 464 family diversity 5, 672, 692, 701 family leave 5, 16, 17, 94, 397–399, 401, 414–421, 433–437, 440, 444, 445, 447–451, 453, 459, 472, 477, 577, 579, 584, 611, 685, 687, 695, 700 insurance 439–442, 446, 447, 449–452 family life courses 309, 661 family policy concepts 10 database 271, 640, 643 explicit 28, 50 implicit 7, 28 motive 121, 122 typology 129, 251, 635 father-non-resident 275, 283, 293 fathers 12, 15, 30, 32, 35, 36, 38, 62, 135, 157, 205, 207, 208, 224, 230, 234, 240, 257, 272, 291, 295, 297, 305, 310, 321, 369–387, 389, 420, 447, 616, 669, 696 federal/federalism 4, 16, 17, 75, 78, 80, 206, 209, 227, 431, 433–436, 439, 442, 448,

451–453, 459, 461, 463–465, 469–472, 477–479, 485, 486, 489, 493, 685, 687 federal state 4, 17, 18, 485–487, 501, 502, 689 female employment 80, 92, 180, 195, 201, 254, 269, 272, 460, 466, 468, 474, 475, 477, 480, 492, 494, 502, 511–515, 517–519, 524, 527, 528 feminist tradition 122, 128 fertility 33, 51, 55, 56, 59, 60, 62–64, 194, 196, 198, 199, 209, 211, 220–222, 228, 231–233, 242, 306, 334, 373, 402, 407, 409, 412, 423, 520, 610, 636, 660, 669, 670, 683, 684, 698 fertility rates 50, 55, 198, 220–224, 226, 231, 233, 237, 334, 489, 629, 695 final availability 19, 536, 672, 685, 692 Finland 35, 180, 199, 202, 224, 226, 268–272, 278, 280, 287, 291, 293, 295, 296, 298, 308, 319, 321, 334, 336, 337, 352, 354, 356, 398, 401, 403, 407–412, 415, 417–422, 432, 513, 539, 541, 543, 582, 699, 700 flexibility stigma 537, 561, 562 flexible use law 437 flexible working 4, 7, 12, 18, 205, 210, 211, 535, 536, 542, 557–562 arrangements 143, 159, 210, 211, 537, 540, 542, 543, 553–562, 564, 691, 692, 697 times 575 flexitime 537–541, 543, 553, 554 frail family members 356, 694 France 30–32, 35, 79, 126, 170, 173, 177, 178, 197, 199–201,

205, 207, 224, 225, 268, 269, 271, 275, 278, 280, 285, 292, 294, 321, 334, 336, 337, 356, 375, 378, 381, 382, 387, 432, 611 functionings 144, 146, 495, 496, 499–503 achieved 144, 146, 496, 497, 502 valued 146

#### **G**

Gabon 235 Gambia 235 gay 371, 398, 399, 410, 412, 415, 417, 421, 422, 610 gender gender equality 4, 6, 8, 14, 15, 32, 34, 36, 50, 51, 60, 69–71, 75, 77–79, 81, 88, 90, 93, 94, 96–98, 101–104, 106, 107, 110, 127, 128, 134, 142, 146, 194, 196, 249, 256, 260, 268, 269, 272–275, 290, 295–299, 309, 359, 373, 388, 453, 563, 628, 629, 672, 683, 685, 688, 690, 694–696 gender norms 148, 258, 389, 535, 554, 561, 610, 691, 693, 697 gender relations 51, 64, 126, 134, 606, 695, 696 gender revolution 695, 701 gender roles 89, 107, 203, 353, 388, 401, 560, 561, 628, 660, 696 gender wage gap 259, 260, 374, 562, 697 generational interdependence 333 geo-spatial 529 Germany 17, 18, 30, 33, 35, 80, 177, 200, 202, 206, 208, 224, 254, 255, 257, 259, 268, 269, 271, 275, 290, 291, 293, 321,

334, 336–338, 356, 372, 380, 381, 385, 386, 388, 389, 432, 485–487, 489, 490, 492, 493, 497, 500–503, 514, 582, 669, 689 Ghana 234 globalization 4, 21, 76, 91, 105, 657, 658, 684, 685 global poor 689 governance 3, 5, 7, 12, 48, 73, 74, 87, 90, 91, 94, 95, 105, 465, 640, 643, 645 gradual concept 120, 124, 125, 129, 135, 136 Greece 33, 199, 205, 225, 226, 356, 432, 449, 540, 636, 699 guardian 270, 409, 410, 422 legal guardian 409, 448, 449 Guatemala 237, 239, 240 Guinea 235 guinea-bissau 235

#### **H**

H1N1 outbreak 446 Haiti 239 Hawaii 439, 440, 442, 445, 446 healthcare 18, 451, 460, 461, 463–466, 477, 479, 698 home care 35, 189, 209, 257, 336, 337, 356–359, 373, 688, 694 home working 537 homogamy 147, 196, 662, 663, 670 Honduras 237, 239, 240 horizontal inequality 658, 671, 689, 690 Human Resource (HR) corporate 20 global 20 local 20 manager 20, 577, 582–584, 588, 589, 597–599, 614

Hungary 27, 35, 178, 182, 200, 202, 204, 207, 255, 334, 337, 352, 582, 636, 699

**I**

Iceland 178, 199, 200, 208, 225, 268, 271, 278, 280, 288, 290–292, 336, 374, 380, 381, 383, 384 income 15, 17, 20, 21, 28–32, 46, 53, 55–57, 60–64, 92, 96, 98–102, 105, 106, 108, 128, 132, 147, 151, 152, 157–159, 178–182, 196, 197, 205, 211, 221, 223, 226, 228–230, 238, 240, 243, 256, 259, 275, 293, 295–297, 306, 311, 314, 317, 318, 322, 325, 333, 386, 415, 423, 464, 472, 479, 493, 517, 643, 658, 659, 661, 662, 667, 668 security 45, 46, 49, 50, 53, 78, 79, 95–97 support 25, 30–34, 37, 63, 219, 304–306, 308, 310, 316, 320, 323, 332, 683, 694 indicators 11, 20, 58, 62, 123, 124, 127, 128, 133, 141, 153, 172, 251, 255, 257, 313, 315–317, 319, 324, 356, 359, 378, 460, 465–469, 490, 597, 598, 627, 628, 631, 632, 634, 640, 646, 670, 671 individual entitlements 16, 374, 375, 378, 380, 383–386, 388, 389, 433, 435, 436, 439, 441, 444, 447, 448, 450, 452, 453, 696 individualization 670 industrial relations 543, 553 inequality 5, 11–13, 18, 19, 21, 75, 106, 125, 128, 144, 146, 159, 195, 211, 230, 243, 249, 253, 259, 261, 303, 304, 311, 352,

353, 356, 357, 359, 488, 563, 576, 577, 600, 634, 636, 647, 657, 658, 661–665, 668, 671, 672, 699 inequalities in caregiving 353, 356, 357, 359 informal care 133, 228, 256, 307, 333, 335, 356 in-kind services 332 International Human Resource Management (IHRM) 603, 611 International Labour Organisation (ILO) 7, 8, 71, 78, 87–98, 100, 104–107, 110, 220, 224, 231, 233, 234, 239, 242, 373, 684, 687, 698 International Organizations (IOs) 7–9, 45–47, 58, 62, 64, 69–74, 78, 81, 88, 89, 99, 100, 238, 334, 629, 631, 672, 684, 685, 702 In Vitro Fertilization (IVF) 407, 408, 411, 412, 415–417, 420, 421, 693, 701 Iran 236 Iraq 698 Ireland 126, 172, 179, 200, 201, 207, 224, 304, 317, 318, 334, 352, 432, 663 Israel 200, 202 Italy 30, 32, 33, 172, 177, 198, 199, 225, 226, 304, 336, 337, 356, 357, 381, 385, 386, 388, 389, 432, 449, 542, 543, 637

**J**

Jamaica 59, 169 Japan 196, 198, 200, 202, 203, 208, 231–233, 242, 514 job characteristics 579, 583 Jordan 236

**K** Kentucky 439, 444 Kenya 94 knowledge economy 659, 660 Korea 196, 198, 203, 206–208, 231–233, 242 Kuwait 236

**L**

labor force participation 10, 45, 46, 50, 52, 54–57, 62, 101, 107, 109, 194, 224, 226, 228, 232, 235, 250, 334, 356, 370, 453, 610, 670, 695 LAC 239 late life 352 Latin America 61, 201, 221, 235, 237–240, 243, 640 Latvia 27, 35, 178, 199, 202, 208, 338 learning/policy learning 3, 6, 9, 34, 48, 54, 71–73, 102, 110, 152, 169, 173, 614, 685 leave carers leave 449, 452, 453 childcare leave 35, 319, 320, 370, 374, 375, 377 child-related leave policies 372, 374, 376, 377, 386, 389 family 5, 16, 17, 94, 110, 397–399, 401, 414, 416, 417, 419–421, 433–437, 439–442, 444–451, 453, 459, 472, 477, 577, 579, 584, 611, 685, 687, 695, 700 flexibility 224, 242, 352, 375, 376, 386, 516 job-protected 36, 397, 401, 431–433, 436, 438, 440, 442–445, 447, 452 maternity 16, 36, 70, 73, 78, 100, 102, 104, 106, 108, 110, 157, 219, 224, 227–230,

233, 234, 237, 239, 240, 242, 319, 320, 370, 372, 373, 377, 384, 417–419, 434, 443–445, 447–450, 611, 667, 689 paid leave 16, 36, 106, 107, 151, 201, 205, 207, 208, 230, 255, 257, 309, 319–321, 382, 434, 439, 441, 447, 449–453, 464, 666, 670–672, 686, 690, 692 paid sick leave 435, 437, 438, 440, 441, 446, 451 parental 7, 10, 12, 13, 16, 35, 36, 51–53, 57, 58, 62, 69, 70, 73, 78, 79, 81, 93–95, 97, 101, 102, 104, 105, 108, 142, 145, 147, 148, 151, 154, 156–158, 194, 195, 204, 207–209, 219, 224, 226–228, 230, 233, 237, 239, 240, 242, 250–254, 256–260, 304, 306, 308–310, 319–324, 372–375, 377, 379, 382, 385, 386, 388, 389, 401, 403, 414–422, 431, 435, 440, 444–449, 464, 472, 488, 489, 498, 516–518, 537, 538, 575, 580, 584, 616, 631, 633, 643, 660, 665, 666, 668, 669, 683, 687, 693–696, 702 paternity 7, 31, 36, 53, 69, 70, 100, 208, 209, 230, 234, 240, 310, 370, 373–375, 377, 380, 385, 387, 389, 398, 414, 417–419, 447, 448, 584, 611 sick 435, 437, 438, 440, 446, 451 transferability 375, 376, 389 type of entitlement 374, 377, 381 lebanon 236 lesbian 398, 408, 411, 412, 414–416, 418, 420, 422, 610 Lesotho 235 LGBTQI 398 Liberia 235 Libya 236 linkage 17, 50, 160, 486

Lithuania 27, 178, 199, 204, 208, 336, 636 lone parents 56, 57, 80, 289 Long-Term Care (LTC) 15, 106, 108, 331–338, 353–359, 694, 699 Luxembourg 35, 208, 225, 311, 313, 318, 323, 337, 432 Luxembourg Income Study (LIS) database 252, 253, 293, 310–314, 635

#### **M**

macro-micro perspective 494 Madagascar 235 malawi 235 Mali 235 Malta v, 35, 178, 225, 423, 542, 636 manager 20, 94, 180, 210, 256, 553, 556–558, 563, 564, 576, 580, 582–584, 588, 596, 599, 691, 700 market-based family support 687 marketization 21, 176, 177, 181, 354, 357, 358, 646, 687, 688, 699 of care 180, 354, 357, 358 marketized services 142, 154, 157, 158, 160, 699 marriage 9, 16, 29, 30, 37, 50, 55, 89, 195–199, 203, 211, 223, 233, 257, 332, 398, 403, 407, 408, 415, 419, 420, 448, 463, 610, 611, 670, 686 Massachusetts 439, 440, 443–447, 449–451 Maternity Leave Act of 1972 444 maternal/maternalist 35, 99, 226, 275, 489, 512, 634 employment 17, 78, 80, 201, 202, 257, 259, 261, 269, 270, 295, 297, 298, 374, 486, 488,

498–500, 502, 503, 513, 514, 519, 630 role 8, 103 mauritania 235 Mauritius 234 means 149 Medically Assisted Insemination (MAI) 402, 407, 408, 411, 412, 415–417, 420, 421 methodological practice 119, 121, 123, 136 Mexico 61, 198, 200, 202, 203, 205, 208, 237, 239–241, 663 Middle East 235 middle-income countries 221, 689 migration 17, 54, 309, 357, 410, 416, 498, 501, 666, 698 model dual-earner dual-caregiver 79, 81, 128, 129, 141, 153, 158, 196, 203, 205, 251, 257, 259, 308, 317, 318, 320–325, 553, 660, 672 earner-carer 257, 270 family support 124, 125, 128, 135 male breadwinner 10, 30, 79, 81, 124–129, 298, 628, 696 market-oriented support 128 residual model of long-term care 333 universalistic model of long-term care 333 mother 8, 13, 30–32, 35, 36, 46, 50, 61–64, 89, 93, 99, 100, 103, 122, 134, 152, 180, 202, 233, 258, 270, 297, 369, 384, 512, 669 birth 407, 408, 415, 417–419, 422, 472 social 408, 414, 415, 417, 420, 422 Morocco 235 mother-labour-force participation 46

mozambique 235 multidimensionality 128, 129, 134, 136, 630 multilevel v, 3, 5, 6, 19, 251, 523, 576, 577, 584, 598, 599, 685, 686 Multilinks 149, 640, 643 Multi-National Enterprises (MNEs) 20, 603–608, 610–613 municipality 16, 17, 486, 489, 493, 501, 502, 524, 527

#### **N**

Namibia 234 national contexts 543, 560, 563, 685 Netherlands 16, 18, 30, 32, 170, 176–181, 196, 197, 200–204, 207, 210, 211, 225, 308, 333, 336–338, 352, 354–356, 375, 381, 382, 387, 389, 398, 401, 403, 407, 409, 410, 412, 414, 418, 419, 421, 422, 432, 497, 512–517, 520, 521, 526–529, 539, 541–543, 582, 665, 688, 690, 699 Netherlands Kinship Panel Study (NKPS) 511, 512, 516, 519–523, 527 New Jersey 436, 439, 440, 442, 446, 470, 472 New York 5, 104, 436, 439–442, 446, 447, 449, 459, 470, 472 New Zealand 8, 78, 80, 183, 197, 203–205, 210, 228–230, 268–270, 274, 275, 290, 292, 293, 295–298, 432 Niger 235 Nigeria 234, 611 non-familialization 46, 63 Nordic Social Statistical Committee (NOSOSCO) 639 North Africa 233

North America 226 Norway 35, 36, 172, 197, 200, 201, 203, 208, 210, 224, 225, 254, 268, 269, 275, 280, 290, 291, 295–299, 321, 336, 337, 352, 374, 380, 381, 383, 388, 398, 403, 408, 410–412, 415–417, 419, 421, 432, 540, 663

#### **O**

Oceania 228, 230 Oman 236 Oregon 440, 445, 446, 450 Organisation for Economic Cooperation and Development (OECD) 4, 7, 12, 27, 35, 45–48, 56, 61, 63, 145, 149–153, 157, 158, 183, 194–203, 205, 207, 210, 269, 373, 644 countries 9, 12, 54, 69, 79, 158, 195–211, 226, 228, 237, 242, 304, 308, 332, 372, 515, 636, 658, 664, 665, 670, 691, 695 family database 33, 196, 200, 202, 204, 206, 209, 220, 317, 319, 640, 644 organization 4–7, 18–20, 29–31, 37, 46, 48, 63, 69, 77, 95, 106, 132, 181, 335, 552, 581, 583, 587, 590, 611, 687 survey 576, 577 work-family policies 148, 253, 575–578, 581, 584, 598, 599

#### **P**

Pacific 222, 238 Palestine 236 Panama 239, 240 Paraguay 240 parent labour-force participation 252

legal 399, 407–410, 415, 416, 418, 420, 422, 693 parenting 16, 25, 32, 57, 93, 101, 103, 148, 157, 173, 174, 198, 199, 273, 290, 295, 369, 371, 401, 403, 413, 418, 420, 422, 701 single/lone 13, 14, 56, 57, 80, 158, 179, 180, 200, 211, 224, 267, 292–294, 298, 303–314, 318, 321, 324, 488, 576, 643, 669, 673 parentalization 15, 398, 399, 401–403, 414, 419–423, 693 Parental Leave Database (PLB) 319–321 performance outcomes 557, 559, 562 Peru 239, 240 Pillar of Social Rights 7, 53, 63, 357, 694 Poland 33, 199, 204, 206–208, 268, 269, 271, 275, 291, 292, 294, 304, 336, 380, 384, 387, 561, 636, 700 policy availability 672, 689 implementation 21, 30, 70, 604, 672, 684, 689 principles 268, 269, 274, 280, 281, 290, 291, 293–298, 486 provision 16, 122, 308, 323–325, 491, 492, 501, 576, 577, 590, 684, 687 Portugal 33, 196, 197, 208, 210, 224, 226, 337, 356, 374, 380, 381, 385–388, 432, 449, 540, 582 poverty 4, 7, 13, 14, 17, 30, 46, 47, 52–57, 59–62, 75, 76, 127, 170, 171, 173, 177, 194, 197, 211, 225, 226, 228, 230, 233, 243, 267, 290–294, 298, 299, 304–306, 309, 311–314, 352,

453, 460, 464, 472, 473, 478, 664, 671, 688, 701 precarious work 178, 659 pregnancy disability leave 442–444 Pregnancy Discrimination Act of 1978 443 privatization 170, 178, 179, 280, 659, 661 productivity 99, 107, 243, 557–559, 578, 583, 609, 616, 683, 686, 691 public services 5, 89, 106, 107, 205, 207, 307, 354, 357, 699

#### **Q**

Qatar 236 quality 11, 57, 60, 98, 101, 105, 143, 148, 149, 152, 153, 159, 174, 176, 177, 181, 224, 257, 355, 493, 501, 629, 644

#### **R**

recession 33, 446 reconciliation between work and family life 38, 51, 79 re-familialization 15, 358 regional 5, 6, 17, 18, 88, 220, 485, 486, 499, 519, 615, 686 inequality 689, 699 variation 16, 17, 154, 158, 160, 209, 485, 486, 489–493, 501, 504, 699 registered partnership 403, 407, 410–415, 418 regulating family 26, 32, 219 residential care 335–337, 358, 688, 694 resourcing families 38 Rhode Island 436, 439, 440, 442, 445–447, 452, 472 robotization 658 Rwanda 235

**S**

same-sex couple 16, 371, 399, 401–403, 407–412, 414, 417, 419–421, 423, 693 marriage 16, 398, 403, 419, 610, 611 parenthood 398, 402, 403, 415, 419 union 398 sao tome and principe 235 saudi arabia 236 Senegal 234 services 6, 9, 12, 13, 15, 17, 18, 21, 26–28, 30, 37, 58, 96, 98, 107, 132, 153, 157, 182, 210, 240, 257, 337, 357, 461, 463, 471, 488, 495, 539, 630, 635, 637, 647 sex segregation 252, 259, 260 Seychelles 235 shared care 14, 268, 272, 295, 296 Shared Parental Pay 157 shared residence 303, 310, 696 shock 21, 464, 698, 699, 701 single parents 13, 14, 158, 211, 292, 293, 303–309, 311–315, 317–319, 323, 324, 418, 669, 694 sierra leone 235 skill 210, 259, 577, 590, 594 gap 576, 585, 599 highly skilled employees 576, 577, 598 Slovak Republic 200, 202, 207 Slovenia 35, 196, 199, 201, 203, 210, 304, 337, 700 small necessities leave 437, 438 social class 13, 144, 145, 158, 221, 249, 252, 255, 257, 259, 260, 387, 399, 400 social inequality 19, 576, 577, 600, 647

social investment/social investment package 7, 34, 45, 52, 53, 57, 60, 61, 63, 64, 80, 81, 101, 181, 492, 629, 646, 684, 687 social mechanisms 224 Social Policy Indicator Database (SPIN) 149, 310, 315, 319 social protection 49, 54, 56, 57, 62, 80, 88, 89, 95–97, 104–110, 194, 203, 220, 224, 231, 303 spending 203, 222 social rights 31, 73, 126, 128, 135, 147, 310, 311, 315, 319, 400, 402, 629, 633, 672 Somalia 235 South Africa 234, 235, 312–314 south asia 222 Spain 30, 35, 179, 197, 198, 200, 201, 203, 208, 224, 226, 268–271, 273, 275, 280, 295, 296, 298, 336–338, 356, 379, 381, 387, 432, 540, 582 stabilizing demography 45, 46, 64 *Starting Strong* 57, 61, 63 state 3, 6, 8, 10, 13, 16, 18, 27, 28, 49, 51, 70, 72, 75, 77, 81, 90, 94, 110, 120, 122, 126, 128, 132, 136, 149, 150, 179, 220, 226, 238, 250, 259, 268, 293, 311, 322, 333, 399, 432, 433, 436, 437, 439, 440, 442, 444, 445, 449, 451–453, 460, 463, 466, 471, 473, 476, 478, 493, 552, 633, 667, 686, 698 sub-national 683, 685, 686, 690–692 sub-Saharan Africa 92, 104, 221, 234 subsidies 9, 30, 31, 150, 158, 179, 206, 233, 235, 257, 308, 459, 500, 513, 516–518, 640 Sudan 235 supervisor 588, 597, 609, 615

supra-national 683–685, 687, 691, 698, 702 surrogacy 407, 409, 410, 412, 416, 421 altruistic 409, 421 commercial 409, 421 Swaziland 235 Sweden 35, 79, 126, 142, 154, 157–160, 169, 170, 182, 197, 199, 201, 203, 204, 208–210, 224, 259, 268–272, 280, 290, 291, 304, 305, 310, 318, 321, 336, 337, 352, 354, 355, 373, 380, 381, 383, 384, 398, 403, 407, 408, 410–413, 415, 416, 420, 537, 561, 582, 663 Switzerland 32, 172, 200, 202, 203, 336, 337, 432, 540

#### **T**

Tanzania 234 targeting 31, 33, 37, 99, 229, 232, 307, 314, 335, 358, 433, 672 within universalism 33 tax avoidance 659 technological change 657, 659 teleworking 537, 541 temporary disability 439, 441, 443, 452 Temporary Disability Insurance (TDI) 439, 442, 450, 464 Tennessee 440, 445, 447 tension centered theory 603, 605–608 tensions centralized 608, 612 decentralized 608, 612 operational 608, 613, 615 strategic 608, 613, 615 Togo 235 trade-off 13, 14, 21, 98, 309, 358, 359, 599, 667, 673, 694 TransMONEE 639

transnational advocates 74 trinidad & tobago 239 triple bind 304, 305, 306 Tunisia 235 Turkey 198–203, 205, 208 typology 10, 125–127, 129, 136, 250, 251, 275, 635

#### **U**

Uganda 234 unions 48, 78, 94, 398, 543, 553, 556, 557 United Kingdom (UK) 8, 33, 35, 78, 80, 142, 147, 148, 154, 157–160, 170, 172, 179, 200–202, 205, 207, 210, 211, 234, 259, 268, 269, 271, 272, 275, 280, 292, 293, 297, 308, 313, 314, 317, 322, 334, 336, 338, 352, 354, 381, 382, 517, 543, 582, 637, 663, 687 United Nations Children's Fund (UNICEF) 8, 71, 87, 89, 90, 98–103, 684 United Nations Economic Commission for Europe (UNECE) 639 United States (US) 5, 8, 16–18, 56, 60, 80, 171, 172, 176, 177, 199, 201, 204, 206, 209, 226–230, 242, 268, 270–274, 280, 290–293, 304, 308, 311, 312, 314, 318, 431–433, 436, 447–450, 452, 453, 459, 461, 462, 464, 465, 470, 479, 500, 501, 514, 517, 535, 536, 610, 615, 663, 672, 685–689, 695, 697 universalism 11, 33, 182, 324 unpaid leave 70, 382, 432, 434, 447, 448, 464

UN Women 8, 87, 89–91, 96, 103–106, 108, 109, 303, 684, 692, 698 Uruguay 238–241

**V**

variable-oriented approach 124 Vermont 440, 445, 446, 449 vertical inequality 661–663, 665, 666, 669–671, 689–691

**W**

wages 13, 30, 32, 53, 101, 179, 210, 249, 250, 252, 256, 257, 260, 462, 463, 466, 468, 470, 472, 477–479, 658, 660, 668 Washington 47, 60, 436, 439, 440, 446, 447, 450, 451, 472 welfare state conservative 123 liberal 8, 80, 123 paradox 257, 259, 260 regime 78, 123–126, 128, 130, 220, 305, 399–401, 582 social-democratic 123 well-being 3, 8, 13, 18, 27, 28, 31, 36, 37, 54, 57, 89, 121, 130, 173, 304, 355, 370, 461, 465, 468, 472, 479, 495, 500, 579, 701

subjective 355, 499, 500 West Germany 172, 485, 487–489, 494, 498–500, 503 within-country differences 17 work-family 81, 142, 143, 147, 259, 305 balance 102 conflict 578 policies 148, 253, 575–581, 584, 585, 588, 590, 596, 598, 599 working hours 157, 202, 210, 211, 249, 254, 271, 334, 526–529, 537, 540, 542, 554, 560, 561, 584, 606 working time 51, 63, 537 arrangements 597, 598 work-life 5, 145, 194, 332, 338, 389, 575, 606, 609, 611, 614, 617 management 20, 603–609, 612 World Bank (WB) 7, 45–48, 58–60, 62, 64, 73, 100, 102, 226, 234, 684, 685

**Y** Yemen 236

**Z** Zambia 235 zimbabwe 235