Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Economics and Finance. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 04 October 2022

Maternity Leave and Paternity Leave: Evidence on the Economic Impact of Legislative Changes in High-Income Countriesfree

Maternity Leave and Paternity Leave: Evidence on the Economic Impact of Legislative Changes in High-Income Countriesfree

  • Serena Canaan, Serena CanaanDepartment of Economics, Simon Fraser University
  • Anne Sophie Lassen, Anne Sophie LassenDepartment of Economics, Copenhagen Business School
  • Philip RosenbaumPhilip RosenbaumDepartment of Economics, Copenhagen Business School
  •  and Herdis SteingrimsdottirHerdis SteingrimsdottirDepartment of Economics, Copenhagen Business School

Summary

Labor market policies for expecting and new mothers emerged at the turn of the 19th century. The main motivation for these policies was to ensure the health of mothers and their newborn children. With increased female labor market participation, the focus has gradually shifted to the effects that parental leave policies have on women’s labor market outcomes and gender equality. Proponents of extending parental leave rights for mothers in terms of duration, benefits, and job protection have argued that this will support mothers’ labor market attachment and allow them to take time off from work after childbirth and then safely return to their pre-birth jobs. Others have noted that extended maternity leave can work as a double-edged sword for mothers: If young women are likely to spend months, or even years, on leave, employers are likely to take that into consideration when hiring and promoting their employees. These policies may therefore end up adversely affecting women’s labor market outcomes. This has led to an increased focus on activating fathers to take parental leave, and in 2019, the European Parliament approved a directive requiring member states to ensure at least 2 months of earmarked paternity leave.

The literature on parental leave has proliferated during the past two decades. The increased number of studies on the topic has brought forth some consistent findings. First, the introduction of short maternity leave is beneficial for both maternal and child health and for mothers’ labor market outcomes. Second, there appear to be negligible benefits from a leave extending beyond 6 months in terms of health outcomes and children’s long-term outcomes. Furthermore, longer leaves have little, or even adverse, influence on mothers’ labor market outcomes. However, evidence suggests that there may be underlying heterogeneous effects from extended leave among different socioeconomic groups. The literature on the effect of earmarked paternity leave indicates that these policies are effective in increasing fathers’ leave-taking and involvement in child care. However, the evidence on the influence of paternity leave on gender equality in the labor market remains scarce and is somewhat mixed. Finally, recent studies that focus on the effect of parental leave policies for firms find that in general, firms are able to compensate for lost labor when their employees go on leave. However, if firms face constraints when replacing employees, it could negatively influence their performance.

Subjects

  • Health, Education, and Welfare Economics
  • Labor and Demographic Economics
  • Public Economics and Policy

One of the most notable changes in the labor market during the past century has been the significant rise in female labor force participation. As a result, parental leave systems have become increasingly important and are now a critical component of labor market policies in most high-income countries. Today, all Organisation for Economic Co-operation and Development (OECD) countries apart from the United States have federally funded parental leave programs. However, these programs vary substantially in terms of key features, such as duration and benefits. For example, the length of paid parental leave in Spain and the Netherlands is 16 weeks, whereas the total paid leave in countries such as Finland, Hungary, Estonia, and the Slovak Republic is more than 160 weeks. Moreover, policies have changed significantly and rapidly during the past few decades. In 1980, the average duration of paid leave in the OECD was 14 weeks, compared to more than 53 weeks on average in 2018. Another major change in parental leave systems is the recent focus on incentivizing fathers to take leave. Norway was the first country to introduce a fathers’ quota (paternity leave earmarked for fathers) in 1993 and was soon followed by other countries such as Sweden, Iceland, and Spain. In 2015, three-fourths of OECD countries provided at least a few days of paid leave that can only be used by a father, and in 2019, the European Parliament approved a directive requiring member states to ensure at least 2 months of earmarked paternity leave.

Evaluating the effects of different parental leave policies is a complex task. First, the aim of parental leave policies is multifaceted. Initially, the main motivation for parental leave provisions was to ensure the health and survival of infants and to allow mothers to recover after childbirth. More recently, increased attention has focused on the influence of family policies on labor market outcomes and gender equality. In addition, governments need to consider other factors such as firms’ productivity and government expenditures. Second, the effects of parental leave policies can depend significantly on the setting. Specifically, the impact of extending maternity leave can depend on social norms, availability of day care, and length of the initial leave, and it can vary by demographic group as well. Furthermore, parental leave policies can vary along several dimensions, such as the length of leave, benefits, eligibility, and division of child care responsibilities between parents.

This review focuses on studies that allow for causal inference. Many of the articles included in this review apply regression discontinuity (RD), a difference-in-differences (DD) approach, or a combination of the two to study the effects of policy reforms. When done well, and with appropriate data, using within-country policy changes has important advantages. When policy reforms happen unexpectedly, there is limited scope for manipulation into treatment and arguably little reason for concern about omitted variable bias. The studies therefore provide a causal estimate of the effects of the policy change on parents who just became eligible compared to those just rendered ineligible. However, these studies also have limitations that are important to keep in mind. First, they do not capture broader effects. For example, a parental leave policy can affect employers’ expectations about the behavior of all women in the labor market, which will affect outcomes for both the treatment and the control group. Second, these studies rely on the immediate effectiveness of the policies. If, for example, following a reform introducing earmarked paternity leave, fathers’ use of paternity leave increases only gradually, a study applying RD or DD may underestimate the true impact of the policy. Finally, it is important to keep in mind that these studies estimate the average treatment effect, referring to the weighted average effect on people who changed their behavior after the policy reform and those who did not. Any effect on the outcome of interest depends on the uptake rates and which part of the population exhibits the behavioral change.

The studies reviewed in this article focus on different settings and time periods. Dissimilarities in labor market policies, social norms, and other factors can play a significant role in outcomes and interact in various ways with different parental leave systems. However, the review highlights some key findings in the literature that are remarkably consistent across contexts. First, the introduction of short leaves is found to be beneficial, both in terms of health outcomes for mothers and their children and in terms of mothers’ labor market outcomes. Second, there are negligible benefits of leave beyond 6 months in terms of health outcomes. Longer leave often has an adverse effect on mothers’ wages and employment while appearing to have little effect on children’s long-term outcomes. However, there is evidence that there may be underlying heterogeneous effects for different socioeconomic groups.

The results on the effects of paternity leave are more mixed. Although earmarked paternity leave proves effective in terms of increasing the uptake rate of fathers, the magnitude of its success in doing so varies tremendously across settings, and the evidence of its effect on labor market outcomes for both men and women is mixed. However, the majority of studies find that introduction of earmarked paternity leave increases paternal involvement in child care, but to a lesser extent in other household tasks. If fathers are more involved in child care, this might affect child outcomes, but very few studies investigate this possibility. However, those that do explore this idea find positive effects and highlight potential complementarities from paternal and maternal care. Moreover, there are important spillovers to other aspects of family life—most notably, to couple stability and fertility. However, the evidence remains scarce and somewhat mixed, underscoring the need for further research to understand the influence of paternity leave on family outcomes.

Finally, several recent studies focus on the consequences of parental leave on firms. The general finding from this literature is that firms are able to compensate for the lost labor input from having employees go on leave by hiring more employees or increasing work hours of other workers. These measures help firms avoid incurring losses in their overall performance. However, firms may face certain barriers to replacing the lost labor input from leave-taking, resulting in negative effects on their performance.

The aim of this review is to provide a comprehensive overview of recent research on the impact of parental leave policies on key outcomes such as children’s health and development, mothers’ health, parents’ labor market outcomes, societal norms, gender roles, gender equality, and their influence on firms’ outcomes. This analysis will complement previous surveys found in the literature. First, Olivetti and Petrongolo (2017) share a detailed overview of the historical background of family policies in high-income-countries with a focus on parental leave, child care, and early childhood education. They survey the literature on the effects of these policies on fertility and women’s labor market outcomes. Second, Rossin-Slater (2018) provides an excellent review of the literature on the influence of maternity leave and family policies on children’s health and mothers’ labor market outcomes. Third, Berlinski and Vera-Hernández (2019) summarize the literature on various family policies, including maternity leave, on child development.

This review broadens the focus and brings in recent developments in the literature. In particular, it summarizes the literature on (a) the impact of family leave policies on firm outcomes; (b) the impact of policies that target fathers’ leave-taking; and (c) the role of norms, gender roles, and intra-household bargaining. The review starts with an overview of the history and motivation behind maternity leave policies. Next, it discusses the effects of maternity leave on maternal health, child health and development, fertility, couple stability, and mothers’ labor market outcomes. Then it presents the more recent history of paternity leave and summarizes how paternity leave schemes affect both fathers’ and mothers’ earnings and family outcomes. It then summarizes the studies that evaluate the effect of leave policies on firms and employers. It concludes by outlining suggestions for future research.

Maternity Leave

Overview of History and Purpose

Labor market policies for expecting and new mothers emerged at the turn of the 19th century, following the Industrial Revolution and urbanization, when women increasingly began to work outside of their homes. The main purpose of these policies was to protect the health of mothers and their newborn children. Through the Factory Act of 1877, Switzerland was the first country to prohibit employment of pregnant women 2 weeks prior to and 6 weeks after childbirth. Similar laws were passed in Germany in 1878, Hungary in 1884, Austria in 1885, the Netherlands in 1889, Norway in 1892, Sweden in 1900, Denmark in 1901, and Greece in 1912 (Wikander et al., 1995).

In the beginning, these laws focused on employment prohibition around childbirth, but these policies evolved to enable voluntary leave from work, consisting of job protection and, in some cases, income support. Most of these policies were formulated with the underlying assumption that a male breadwinner was earning a sufficient family wage, which arguably underscored women’s roles as wives and mothers (Wikander et al., 1995). By the mid-20th century, the focus of family leave policies shifted toward women’s rights and gender equality. In light of this change, parental leave systems became a means for women to reconcile their jobs and family life. Sweden introduced 3 months of maternity leave in 1955, followed by Norway in 1956, Finland in 1964, and Denmark in 1967, with payment equivalent to unemployment insurance or sickness benefits (Datta Gupta et al., 2008).

The first proposal to introduce mandated parental leave from the European Commission of the European Union (EU) in 1983 was an effort to promote equal opportunity by ensuring that leave could be used by either parent. However, the proposal was rejected, and the EU directive on parental leave was not adopted until 1996 (Fusulier, 2011). When adopted, most EU member states had already implemented some form of shareable parental leave. In 1974, Sweden was the first country to introduce shared parental leave. Slovenia and France followed in the same year, and Norway did so in 1977 (Kamerman & Moss, 2011). However, mothers almost exclusively used parental leave rights. By 1990, all OECD countries except Australia, New Zealand, Switzerland, and the United States offered at least 12 weeks of paid leave. In 2013, 98 countries provided at least 14 weeks of employment protection, and 74 countries provided at least two-thirds of the women’s pre-birth earnings for at least 14 weeks (Addati et al., 2014). Figure 1 shows the duration of provision of paid leave and job protection across the OECD countries in 2018. Givati and Troiano (2012) propose that parts of the variation throughout the world stem from societal tolerance of gender-based discrimination. Measuring attitudes with a language-based measure, they show that societies with less tolerance provide longer leave.

Figure 1. Leave provision in the OECD in 2018.

Source: OECD. Figure produced by the authors.

The United States remains an outlier. It is the only high-income country in the world with no national provided paid parental leave system. During the 1970s, 23 U.S. states passed laws that prohibited health insurance companies from treating pregnancy differently from comparable illnesses, and federal law (Pregnancy Discrimination Act of 1978) outlawed employer discrimination of pregnant women more broadly. However, until 1993, there was no national employment protection during the weeks before and after childbirth. The passage of the Family and Medical Leave Act (FMLA) in 1993 ensured any parent the right to unpaid family leave for up to 12 weeks per year. During the past few years, several states have introduced paid family leave (PFL) programs. The first of these programs emerged in California in 2004, where employees became eligible for up to 6 weeks of leave with partial wage replacement. Similar programs were subsequently implemented in New Jersey in 2009, Rhode Island in 2014, and New York in 2018—where new parents could receive paid leave for up to 6, 4, and 8 weeks, respectively.

Generally, due to the timing of the introduction and development of parental leave schemes throughout the world, most modern studies evaluating the effects of introducing a short maternity leave are conducted with U.S. data, whereas studies on the extension of leave programs stem mainly from Europe.1 Table 1 provides an overview of the reforms evaluated in the studies reviewed in this section, with details on leave duration, compensation, and eligibility.

Table 1. Maternity and Parental Leave Reforms, by Country and Reform

Reform

Details

References

Norway (1977)

Paid leave extended from 12 to 18 weeks. Eligibility dependent on prior employment. 100% replacement, increased from health insurance. Job protection extended from 3 to 12 months.

Bütikofer et al. (2021), Carneiro et al. (2015)

Norway (1987–1991)

Paid leave gradually extended from 18 to 35 weeks.

Kotsadam and Finseraas (2013)

Norway (1992)

Extended paid leave from 42 to 52 weeks. 80% replacement, 100% if leave duration is shortened. Eligibility depends on prior employment.

Corekcioglu et al. (2021), Dahl et al. (2016)

United States (1978)

Expansion of the Temporary Disability Insurance covering birthgiving mothers receiving 6–12 weeks of leave with 50–66% replacement.

Stearns (2015)

United States (1993)

Introduction of the Family Medical Leave Act that provides 12 weeks of job protection. Eligibility depends on employment time at the workplace and workplace size.

Rossin (2011)

California (2004)

Introduction of 16 weeks of paid family leave at 55% compensation rate.

Appelbaum and Milkman (2011), Milkman and Appelbaum (2013), Rossin-Slater et al. (2013), Lerner and Appelbaum (2014), Baum and Ruhm (2016), Bailey et al. (2019), Bullinger (2019), Pihl and Basso (2019), Bana et al. (2020)

New Jersey (2009)

Six weeks of additional paid family leave in addition to TDI, at two-thirds wage compensation.

Appelbaum and Milkman (2011), Lerner and Appelbaum (2014)

Rhode Island (2014)

Four weeks of additional paid family leave in addition to TDI, at 60% wage compensation.

Bartel et al. (2016)

New York (2018)

Eight weeks of paid family leave in 2018, 10 in 2019, and 12 in 2021, at wage compensation of 50% in 2018 and 67% in 2021.

Bartel et al. (2021)

Sweden (1979)

Employment requirement relaxed for parity >1.

Ginja, Jans, et al. (2020)

Sweden (1989)

Paid job protection extended to 15 months.

Liu and Skans (2010)

Germany (1979)

Paid job protection extended from 2 to 6 months. Benefits replacing income in the first 3 months and being provided at a flat rate for the remaining 3 months of one-third the average national income.

Dustmann and Schönberg (2012), Guertzgen and Hank (2018)

Germany (1986–1992)

Paid leave gradually extended from 6 to 24 months. Job protection gradually extended to 36 months. No employment criteria for benefits.

Dustmann and Schönberg (2012), Gangl and Ziefle (2015), Ejrnæs and Kunze (2013), Schönberg and Ludsteck (2014)

Germany (2007)

Paid leave reduced from 2 years to 1 year at 67% replacement of income. Flat rate for those without employment history. Introduction of 2 months paternity leave.

Raute (2019), Kluve and Schmitz (2018), Cygan-Rehm (2016), Kluve and Tamm (2013), Bergemann and Riphahn (2015), Cygan-Rehm et al. (2018), Huebener et al. (2021)

United Kingdom (1979)

29 weeks of job protection and wages. 6 weeks at 90% wage compensation and 29 weeks at a flat rate.

Gregg et al. (2007)

United Kingdom (1994)

Expanded eligibility. 90% replacement for 6 weeks; flat rate for 12 weeks. Eligibility dependent on prior employment.

Stearns (2018)

United Kingdom (2000)

Paid leave extended to 6 months. Increased duration of the flat rate. Job protection increased to 1 year.

Stearns (2018)

Denmark (1984)

Extension from 14 to 20 weeks at 90% replacement.

Andersen (2018), Rasmussen (2010)

Denmark (1994)

Paid leave extended from 24 to 76 weeks. 90% replacement for 24 weeks, then 60% eligibility depends on prior employment.

Datta Gupta et al. (2008), Friedrich and Hackmann (2021), Andersen (2018)

Denmark (2002)

Increased compensation. 90% replacement for 46 weeks. Job protection for 60 weeks.

Beuchert et al. (2016), Andersen (2018), Gallen (2019)

Austria (1990)

Paid job protection extended from 1 to 2 years. 100% replacement for 8 weeks, then a flat rate. Eligibility dependent on prior employment. Employment requirement relaxed for parity >1.

Danzer and Lavy (2018), Lalive and Zweimüller (2009), Danzer et al. (2020)

Austria (1996)

Paid job-protected leave reduced to 18 months.

Lalive and Zweimüller (2009)

Canada (2000)

Paid leave extended from 6 to 12 months at 55% replacement. Eligibility dependent on prior employment.

Baker and Milligan (2008a, 2008b), 2010)

France (1994)

Increased paid leave from 10 weeks to 3 years. Extended from second to third births. Flat rate benefit. Eligibility depends on prior employment.

Piketty (2005), Lequien (2012), Canaan (2019)

Czech Republic (1995)

Extension of flat rate benefit from 3 to 4 years. No job protection the fourth year. No eligibility requirement.

Mulleröva (2017), Bičáková and Kalíšková (2019)

Czech Republic (2008)

Possibility to shorten leave to 2 or 3 years. Total benefits kept constant.

Bičáková and Kalíšková (2019)

Note: This table is not meant to be an exhaustive list of reforms by country but, rather, contains the reforms utilized by the studies included in this review.

TDI, temporary disability insurance.

Mothers’ Health Outcomes

Although maternity leave policies initially stemmed from a desire to ensure the health of mothers and their newborn children, the evidence on the causal influence of leave on maternal health remains surprisingly scarce. Existing studies primarily focus on the effects of extending the current leave on mothers’ mental health. Overall, studies indicate that the introduction of leave proves beneficial for mothers’ health outcomes, whereas extending parental leave beyond 6 months appears to have a negligible effect. However, the average treatment effect may conceal heterogeneity within the effects of extended leave, and some evidence points to health benefits of a longer leave for women in lower socioeconomic status (SES) groups.

Bullinger (2019) examines the effects of the introduction of 6 weeks of paid leave in California in 2004, which effectively increased the average time that women stayed at home after birth from 3 weeks to 6 weeks. She notes that self-reported mental health and ability to cope with day-to-day demands improved among mothers after the reform. Chatterji and Markowitz (2005, 2012) also focus on leave duration in the United States. They use variation in state leave policies as an instrument for maternity leave length (the average leave length in their sample is 9 weeks) and find that longer leave correlates with decreased depressive symptoms and improved self-reported health. Guertzgen and Hank (2018) study the expansion of paid leave in Germany from 2 to 6 months in 1979 and find that longer leave correlates with a higher incidence of long-term sickness and absence from work, while noting that this likely stems from the impact on selection into the labor market. Avendano et al. (2015) examine changes in the duration of paid maternity leave in several European countries between the 1960s and the 1990s. In 1960, the duration of full-wage weeks ranged from 2 to 16 weeks, and by the end of the study period, it ranged from 8 to 16 weeks. They find that women who had access to more generous leave policies when they had their first child were less likely to experience depressive symptoms at age 50 years.

A paper by Bütikofer et al. (2021) significantly contributes to this literature by estimating the influence of both introducing and extending paid leave on a range of maternal health outcomes. In July 1977, Norway introduced a 4-month paid leave along with 12 months of unpaid leave. Before the policy change, working mothers only had access to 12 weeks of leave. With the reform, benefits increased from sickness benefits to full wage replacement. Observing women at approximately age 40 years, the study finds that the reform improved health outcomes such as body mass index, blood pressure, pain, and mental health. The study also explores the impact of several later leave expansions, each of which increased the duration of paid leave by 2 weeks, and finds no further improvements in the health outcomes.

Studies that focus on expanding leave length beyond 6 months suggest negligible benefits in terms of mothers’ health. Baker and Milligan (2008b) study the effects of extending paid maternity leave in Canada from 6 to 12 months. They examine mothers’ outcomes 7–24 months after giving birth and find no effect on self-reported health, depression, or other postpartum problems. Dagher et al. (2014) use employer policies as an instrument for maternity leave duration and find a U-shaped relationship between leave duration and postpartum depressive symptoms, with minimal symptoms occurring for mothers who had approximately 6 months of leave. However, a study that examines a reform in Denmark (Beuchert et al., 2016) suggests that health benefits may result from a longer leave for some women. In 2002, the reform in Denmark increased the length of parental leave with full benefit compensation and effectively increased the average leave duration from 244 days to 276 days. The study finds that the increased length of maternity leave reduced hospital admissions and the probability that mothers would receive antidepressants in the first 3 years after giving birth and that these effects were driven by mothers with less than 10 years of schooling. Finally, Liu and Skans (2010) find no evidence that increasing parental leave in Sweden from 12 to 15 months has any effect on mothers being hospitalized due to mental disorders within 3, 6, or 16 years after giving birth.

Children’s Health and Development

In the early days of parental leave policies, the central objective was to ensure infants’ health and survival. More recent years have brought an increased focus on children’s developmental outcomes. One argument that has been made for a longer parental leave is that increased parental time may have benefits in terms of cognitive development, which may influence later life outcomes such as schooling and income. The evidence suggests that introducing a short parental leave does indeed have significant benefits in terms of infant health outcomes and can bring health benefits later on by decreasing the risk of children being overweight or diagnosed with attention-deficit/hyperactivity disorder (ADHD). The evidence on the impact of longer leave is more mixed. Overall, it appears to have no effect in terms of development and schooling outcomes. However, there is evidence that long leave may prove beneficial in terms of schooling outcomes for children born in higher SES families, whereas it adversely affects verbal development for children in low SES families.

Several studies that use cross-country comparisons suggest that longer maternity leave is associated with lower infant and child fatalities (Heymann et al., 2011; Ruhm, 2000; Tanaka, 2005). Causal evidence indicates that the effect of introducing a short leave may differ from that of extending a longer leave. Studies from the United States show that ensuring that mothers can stay with their children during the first several weeks after birth has a significant and positive effect on infant health. Rossin (2011) evaluates the effects of FMLA in 1993 in the United States; FMLA mandated a minimum of 12 weeks of unpaid maternity leave for eligible women. She finds that the reform led to small increases in birth weight; decreased the likelihood of premature birth; and significantly decreased infant mortality among college-educated and married mothers, who were most able to take advantage of the unpaid leave. Stearns (2015) studies the effects of the temporary disability insurance (TDI) programs in the United States. In 1978, these programs began providing wage replacement benefits to pregnant women for 6–12 weeks. This reform particularly benefitted women who could not have afforded to take leave before TDIs were implemented. The study finds that the TDI benefits reduced incidences of low birth weight and early term birth and that it had the greatest impact among unmarried and Black mothers. The California PFL program, implemented in 2004, allowed parents up to 6 weeks of paid leave with a newborn. This increased leave-taking by mothers by 3–6 weeks (from a baseline of approximately 3 weeks). Pihl and Basso (2019) find that this reform decreased hospital admissions, and Bullinger (2019) finds it improved overall child health.

There is less evidence on the long-term health benefits of introducing a relatively short maternity leave. Lichtman-Sadot and Bell (2017) investigate the effect of PFL on various health outcomes when children are between ages 5 and 6 years. They find that the reform reduced the risk of children being overweight or being diagnosed with ADHD, hearing problems, or communication problems. Parents who had children after the introduction of PFL were also more likely to assess their child’s overall health more positively and less likely to report a history of frequent ear infections. The researchers find that these effects are driven by children from less advantaged backgrounds, consistent with the finding that PFL has the greatest effect on leave-taking among mothers who could not afford to take unpaid leave.

Whereas the studies from the United States examine the impact of providing a short leave, research from Canada and Europe focuses on the effects of extending maternity leave beyond the first several weeks after birth. Baker and Milligan (2008b) examine the effects of extending paid maternity leave in Canada from 6 to 12 months and find no influence on children’s overall health in the first 24 months after birth. In a follow-up study, Baker and Milligan (2010) find no significant effects on child development, specifically in measures of temperament and motor and social development. In a study of the 2002 reform in Denmark, Beuchert et al. (2016) estimate the effect on children’s inpatient hospital admissions and emergency room visits within 1 year and within 3 years from birth and find no significant effect. Danzer et al. (2020) estimate the effect of extending parental leave in Austria from 1 year on children’s health outcomes. Importantly, they explore regional variation in the availability of formal child care. When studying the heterogeneous effects of extending the duration of parental leave on children’s outcomes, it is critical to consider what is being replaced with the increased time that children have with their parents,2 yet this is rarely addressed. However, the findings of Danzer et al. (2020) highlight the importance of considering this factor, as they show that the extended leave had a positive effect on children’s health outcomes only in regions where formal child care was not readily available.

One mechanism through which longer maternity leave could affect infant health outcomes is by allowing mothers to breastfeed for longer periods. Huang and Yang (2015) examine the impact of PFL on breastfeeding practices in California and find that the reform led to a 10–20%-point increase in breastfeeding rates 3, 6, and 9 months after the birth of a child. Studying the same reform, Pac et al. (2019) also find a positive effect on breastfeeding, with a significantly larger effect for disadvantaged mothers. Baker and Milligan (2008b) find that the reform in Canada, which extended paid leave from 6 to 12 months, increased the duration of breastfeeding by more than 1 month. However, using this exogenous shock to assess the benefits of breastfeeding, their evidence suggests that, at least after 6 months, the benefits of increased duration of breastfeeding are trivial.

Several studies have explored the influence of parental leave on children’s long-term outcomes. Carneiro et al. (2015) study the 1977 law change in Norway, which introduced 4 months of paid maternity leave (and 12 months of unpaid leave). They find that the reform led to a 2%-point reduction in high school dropout rates and a 5% increase in wages at age 30 years. This effect is driven by families with fewer resources, where the mother would have taken very little unpaid leave before the policy change. In contrast, Rasmussen (2010) investigates a policy change in Denmark in 1984 that increased parental leave from 14 to 20 weeks and finds no significant effect on children’s long-term educational outcomes. Furthermore, Dahl et al. (2016) evaluate the effect of increasing paid leave in Norway from 18 to 35 weeks and find no effect on children’s schooling. Liu and Skans (2010) examine the effect of extending parental leave in Sweden from 12 to 15 months. They find no effect on hospitalizations within 3, 6, and 16 years after birth. Furthermore, they find no overall effect on children’s school performance. However, they find positive effects on test scores among children of highly educated mothers. Dustmann and Schönberg (2012) assess the effects of three policy changes in Germany: The first expanded paid leave from 2 to 6 months in 1979, the second extended the leave further to 10 months in 1986, and a reform in 1992 extended paid leave to 18 months. They find no evidence that any of these reforms improved children’s schooling outcomes; in fact, they observe that the 1992 expansion may even have resulted in lower children’s educational attainment. Baker and Milligan (2015) estimate the effect of increasing maternity leave in Canada to 12 months and find no positive effect on cognitive and behavioral development of children when they reach ages 4 and 5 years. Furthermore, they uncover a small negative effect on Peabody Picture Vocabulary Test scores among boys. Similarly, Canaan (2019) investigates the effect of a French reform that extended leave duration from 16 weeks to 3 years and finds that it harmed children’s verbal development at the age of 6 years.

Danzer and Lavy (2018) study the influence of extending paid leave in Austria from 12 to 24 months. They find no significant overall effect on children’s standardized test scores, but similar to Liu and Skans (2010), they find significant positive effects for children of highly educated mothers, especially for boys. Furthermore, they find a negative effect on the schooling outcomes of children whose mothers have lower levels of education, particularly for boys. Danzer et al. (2020) estimate the impact of the same reform on children’s labor market outcomes and find no effects and no indication of a systematic pattern with respect to SES of the mother or the child’s gender. Finally, Ginja, Jans, et al. (2020) investigate a policy change in Sweden that allows mothers higher benefits for a subsequent child without re-establishing eligibility through market work if two births occur within a prespecified interval. They find that this policy improves the schooling outcomes of the older child, likely due to increased maternal time.

Several countries have mandatory prenatal leave to protect the health of pregnant workers and their unborn children. Ahammer et al. (2020) study the effects of a policy reform in Austria that extended the mandatory prenatal leave from 6 to 8 weeks on children’s short- and long-term outcomes. They find that the reform had no influence on children’s health outcomes in the short or long term and had no impact on their future labor market outcomes. Furthermore, they find no effect on maternal health and subsequent fertility.

Fertility and Marriage

Parental leave policies affect the cost of having a child and might therefore affect family outcomes such as fertility and marriage decisions. Several studies have investigated the influence on fertility of a reform in Germany in 2007 that significantly increased the benefits for higher earning women while decreasing the benefits for lower earning women. This reform also reduced the total period of leave duration and implemented paternity leave. Disentangling the effects of these changes proves difficult. Raute (2019) documents a fertility increase of 23% among women with tertiary education, whereas Kluve and Schmitz (2018) observe that the policy reform reduced subsequent fertility among younger mothers. Cygan-Rehm (2016) finds that the reform had a negative impact on fertility among low-income mothers.

Lalive and Zweimüller (2009) and Danzer et al. (2020) consider the effect of extending maternity leave in Austria from 12 to 24 months. Whereas Lalive and Zweimüller (2009) find an increased probability that couples will have a second child within 10 years of the birth of their first child, Danzer et al. (2020) show no significant effects on completed fertility when they extend the horizon to 17 years after birth. They do not find that the reform had an overall influence on divorce probabilities, but they document an increased probability that mothers who were unmarried at the time of birth would get married.3 Liu and Skans (2010) observe no influence of extending the maternity leave in Sweden to 15 months on parental fertility or divorce rates; similarly, Dahl et al. (2016) find no effect of extending the maternity leave in Norway from 18 to 35 weeks on completed fertility, marriage, or divorce.

Mothers’ Labor Market Outcomes

Having children has an immense influence on women’s labor market outcomes both in the short and long term (Angelov et al., 2016; Kleven et al., 2019; Lundborg et al., 2017). This reality has spurred an increased interest in reforms in high-income countries. Policymakers face difficult trade-offs because parental leave programs often aim to accommodate multiple concerns, such as child and parental welfare, parental labor market outcomes, gender inequality, firm productivity, and governmental expenditures. Wage compensation is often combined with job protection schemes to ensure that parents can afford to take the leave and return to the labor market afterward. Proponents argue that generous parental leave policies promote gender equality and increase women’s earnings by allowing mothers to retain valuable firm- or occupation-specific human capital and match-specific human capital after childbirth. However, it is precisely these longer spells of job absenteeism that opponents worry about. They argue that more time away from work lowers women’s future labor market outcomes through human capital depreciation and possibly discrimination. This section reviews the literature on how a variety of parental leave policies affect mothers’ and fathers’ labor market outcomes.

A range of cross-country comparative studies, using variation in the availability and length of leave provision across countries, find that paid leave is associated with somewhat higher female employment rates (Jaumotte, 2003; Pettit & Hook, 2005). In a prominent early study, Ruhm (1998) investigates the effect of parental leave on female employment and wages from 1969 to 1993 in nine European countries that experienced significant changes in their respective parental leave policies. He finds that entitlement to short periods of paid leave, totaling approximately 3 months, led to an approximately 3% or 4% increase in female employment rates but little or no effect on wages. On the other hand, entitlement to longer parental leaves of more than 9 months had no additional impact on employment but a significant negative impact of approximately 3% on female wages. Work covering additional countries and later years broadly confirms Ruhm’s findings (e.g., Blau & Kahn, 2013; Ruhm & Teague, 1995; Thévenon & Solaz, 2013). Cross-country studies also suggest that parental leave length affects women across SES levels differently. Long leave schemes increase labor market participation but decrease earnings for highly educated women relative to other women (Cipollone et al., 2014; Olivetti & Petrongolo, 2017).

Although these rigorous studies provide strong comparisons of parental leave policies across countries, some concerns remain about the causal interpretation of cross-country studies. These studies are prone to overstating the true influence of parental leave because the extensions of this leave often happened over a period of time during which other family-friendly policies were implemented as well. Parental leave can vary in length, extent of job protection, income support, eligibility rules, and availability to either parent. The rules and costs governing preschool education and child care also vary considerably across countries. Some countries have enabled direct family transfers and tax allowances to low-income working parents, differing in rules and magnitudes.

A large branch of the literature has addressed this challenge by focusing on one country and considering policy changes to elicit the causal effect of parental leave policies. Reviewing the results of various studies analyzing parental leave reforms in different countries confirms the overall findings from the cross-country studies. Overall, a concave relationship exists between the length of parental leave and mothers’ labor market outcomes. Introducing and extending parental leave rights and wage compensation for up to 6 months improve mothers’ labor market outcomes. Prolonging these rights for 1 year seems to have little effect, and extending them to 1 year or longer seems to have an adverse effect on women’s wages and employment. Formal rights to maternity leave make it easier for mothers to maintain an attachment to their pre-birth job and employer, meaning that mothers do not have to start over when they return to the labor market after their childbirth and childrearing period. Where these rights already exist, however, extensions of the maternity leave period from a certain point on can have the opposite effect.

Introduction of Short Programs

Generally, introducing a short paid parental leave scheme has been shown to improve mothers’ labor market outcomes. Several studies have examined the labor market consequences of the United States’ first explicit paid parental leave policy implemented in California in 2004. Rossin-Slater et al. (2013) show that the implementation of this program doubled the use of parental leave by Californian women from 3 to 6 weeks on average and that this change primarily resulted from the greater uptake by less privileged mothers. They estimate that this change increased the weekly work hours of employed mothers of 1- to 3-year-old children by 10–17%. Most studies on this reform confirm these findings and show an increase in the labor supply on both the intensive and extensive margins in the short term, while highlighting that mothers’ likelihood of returning to their pre-leave firm increased (Bana et al., 2020; Baum & Ruhm, 2016). Conversely, Bailey et al. (2019) find no evidence that this reform improved women’s labor market outcomes and further claim that women making use of the improved leave provision experience a lower employment rate and wages 6–10 years after birth. Baker and Milligan (2008a) find similar effects when studying the introduction of 18 weeks of parental leave in Canada, which led to a sharp decrease in job separations. Gregg et al. (2007) show that the British introduction of job protection and 6 weeks of wage compensation in 1979 significantly increased mothers’ labor market attachment in the years after childbirth, where many moved from unemployment into part-time positions. On the contrary, the German extension of paid leave and job protection from 2 to 6 months in 1979 led to a decrease in mothers’ employment by 1% or 2% at 52 and 76 months after childbirth, although the effect does not persist in the long term (Dustmann & Schönberg, 2012; Guertzgen & Hank, 2018).

Extending Parental Leave Up to 12 Months

Many countries have expanded their parental leave schemes over time, and the results suggest that little or no effect on mothers’ labor market outcomes occurs until the duration of leave approaches 1 year in duration. Focusing on pre-1993 policy reforms in Norway, Dahl et al. (2016) find that expansions in government-funded maternity leave from 18 to 35 weeks had little effect on a wide variety of outcomes, including parental earnings and labor market participation in the short or long term. Datta Gupta et al. (2008) show that maternity leave approaching 1 year in duration affects Danish mothers’ wages negatively. Nielsen et al. (2004) suggest that the adverse effect is mainly driven by women employed in the private sector, whereas they find no negative effect of a longer leave for mothers in the public sector. They also find that the potentially higher wage compensation during pregnancy and better postpartum career opportunities in the public sector attract pregnant women to shift to the public sector.

In Norway, Corekcioglu et al. (2021) find that the extension of maternity leave from 30 to 52 weeks in 1993 did not help women reach top positions within their organization and indicate that it may even make them less likely to do so. Small but adverse effects on labor market attachment linked to a 52-week parental leave are also found in Germany (Schönberg & Ludsteck, 2014). Stearns (2018) is able to separately identify the effects of extending wage compensation and job protection to 52 weeks in Great Britain and finds that access to longer paid maternity leave increases the probability of returning to work in the short term but not in the long term. In contrast, making job protection available to new mothers significantly increases maternal employment rates and job tenure 5 years after childbirth. Longer leave therefore seems to increase mothers’ labor market attachment but decrease mothers’ chances of career success in terms of promotions to managerial positions.

Extending Parental Leave Beyond 12 Months

In general, studies examining the effects of a parental leave that extends beyond 1 year find adverse effects on mothers’ labor market outcomes. A range of studies examine the French 1994 reform that extended the period of paid leave for families with two children to 3 years and find that the reform induced women to exit the labor market and incur a wage penalty if returning to work, both in the short term (Canaan, 2019; Piketty, 2005) and in the long term (Lequien, 2012). Using German data, Ejrnæs and Kunze (2013) find that the increase in leave duration of up to 36 months led to detrimental effects on employment and wages for mothers. Using survey data, Gangl and Ziefle (2015) show that the expansion of leave duration changed German mothers’ work–family preferences. The affected women reported lower levels of work commitment, and fewer held a full-time position. In the same setting, Schönberg and Ludsteck (2014) find that the adverse effects on the labor market are mainly short-term effects. In particular, they find that increased leave duration reduced employment rates and earnings for up to 6 years after childbirth, but with smaller effects in the longer term. The same has been found in Austria, where an extension of paid leave from 1 to 2 years reduced mothers’ earnings in the short term but had no longer term effects (Lalive & Zweimüller, 2009).

Mullerova (2017) examines a parental benefit reform that took effect in the Czech Republic in 1995, extending the universal parental leave benefits from 3 to 4 years while keeping the job protection period at 3 years. She finds that mothers’ probability of employment decreased by 15–25% at the end of their parental leave and persisted at the same level more than 2 years later. Bičáková and Kalíšková (2019) evaluate the same reform as well as a later reform in 2008. The second reform allowed women to choose an alternative setup that shortened the paid leave from 4 to 3 or 2 years while keeping the overall amount of financial benefits received virtually unchanged. The job protection remained set at 3 years. Their findings demonstrate that the second reform had the opposite effect of the first, although with a much smaller impact.

A shortening of the parental leave also had a positive effect on German mothers. In 2007, Germany modernized its parental leave system, replacing the previous lengthy but low-benefit leave—which specifically targeted low-income families—with a 12-month universal (in principle) leave offering much more generous coverage. The new benefits were dependent on pre-birth earnings, which meant that women with high labor market participation received a higher wage replacement rate. The empirical results indicate that the reform proved effective, leading to a 12% increase in mothers’ employment probability after the end of the benefit period (Kluve & Tamm, 2013) and a positive influence on employment 3–5 years after childbirth for women with relatively high levels of education (Bergemann & Riphahn, 2015). However, these results do hide substantial heterogeneity, as women who were employed prior to giving birth increased their leave duration with the reform (Kluve & Tamm, 2013). Welteke and Wrohlich (2019) argue that the increase in benefits particularly encouraged high-income mothers to stay at home for the first 12 months following childbirth. By considering the increase in leave duration among working women and identifying female co-workers who had a child after the reform, they find substantial spillovers among the co-workers who took a longer leave themselves. The initial effect of the reform on new mothers’ use of this leave and absenteeism from the labor market is therefore greater than what is identified when only considering women in the reform window.

Summary

The surveyed research on the effects of maternity leave suggests significant benefits result from introducing a short leave, whereas the evidence of benefits for extending a longer leave is more mixed.

The literature on health outcomes provides compelling evidence for the beneficial effect on both maternal and child health of the introduction and expansion of a short maternity leave. The beneficial impact of leave extending beyond 6 months is more ambiguous, and some evidence suggests that policies implementing longer leave may increase inequality. In terms of health outcomes, low-income mothers benefit more from the provision and extension of paid leave. However, in terms of children’s long-term outcomes, such as test scores, the benefits of leave extensions appear to be concentrated among those with highly educated mothers. Extending leave duration may therefore strengthen the relationship between maternal SES and child outcomes.

Introducing maternity and family leave entitlements generally appears to improve mothers’ job continuity. The evidence shows that extending these provisions for up to 6 months improves mothers’ labor market outcomes, but longer leave might have an adverse long-term effect on wages, employment, and career opportunities, especially when the leave extends for 1 year or more. The evidence also suggests that there are heterogeneous effects of different parental leave schemes. Offering universal paid leave increases use of leave by low-earning women, whereas longer paid leave and job protection periods may harm highly educated mothers’ careers the most. In particular, women working in the private sector may experience diminished chances of reaching top positions when the paid leave duration increases. Expanding eligibility can also increase fertility, which might in turn lower mothers’ long-term earnings due to the labor cost of additional children.4

Paternity Leave

Because mothers remain the primary users of shareable leave, policymakers have to a greater extent started to target fathers. The primary goal of recent paternity leave policies has been to involve fathers more in child care and other tasks in the household to alleviate some of the responsibility carried by mothers. Indeed, correlative studies show that leave-taking fathers are more involved in subsequent child care (e.g., Boll et al., 2014; Nepomnyaschy & Waldfogel, 2007). If this relationship is causal, nontransferable paternity leave and equal sharing of parental leave should decrease household specialization. This could stem both from a direct effect on parents’ labor supply and outcomes and from a more indirect effect through changing norms and behaviors that can alter the division of labor within the household. Extensive causal evidence reveals the impact on earnings of both mothers and fathers, but perhaps due to data availability, the effect on time spent on child care and housework has been less studied. Moreover, studies have shown an effect on fertility and couple stability. Paternal engagement has a positive association with child development (del Carmen Huerta et al., 2013; Sarkadi et al., 2008) and improves later father–child relationships (Petts et al., 2020). The causality and selection aspects of this finding have only been disentangled in a few articles.

Overview of History and Purpose

The recent focus on involving fathers in parental leave-taking stems mainly from gender equality concerns. According to the EU Commission, shared responsibility between parents should be an essential part of strategies to increase equality between men and women in the labor market and to ensure fathers’ opportunity for time with their newborn child (Council of the European Union, 2019). A nontransferable (earmarked) paternity leave has been introduced sporadically since the 1990s but now often serves as a central element when modernizing the parental leave system in most OECD countries. A short paternity leave around the time of birth was introduced in Finland in 1978 and in Sweden in 1980.5 Norway became the first country to introduce earmarked paternity leave in 1993, followed by Sweden in 1995. In 2000, Iceland passed a law that earmarked one-third of a 9-month-long parental leave to fathers. In 2021, the leave was extended to 12 months, earmarking 6 months to each parent, albeit with the possibility of transferring 6 weeks from one parent to the other. Figure 1 provides an overview of father-specific leaves in OECD countries in 2018.

Although earmarked paternity leave has been praised for being an effective tool, uptake rates differ significantly by country. The Icelandic policy has proven most successful, bringing a more than 80% increase in uptake rates (Olafsson & Steingrimsdottir, 2020). Within Europe, both the German and the Danish introduction of parental leave have been the least effective. The German introduction of 2 months of paternity leave led to an increase of approximately 12%-points (Kluve & Tamm, 2013). The introduction of earmarked paternity leave took place in 2007 as part of the modernization of the German parental leave system, which also shortened the leave duration and made benefits dependent on pre-birth earnings. The Danish introduction of 2 weeks of paternity leave in 1998 implied an increase of approximately 2 days of the average leave taken by fathers, and the abolishing of the earmarked leave in 2002 barely altered the average leave duration (Andersen, 2018).

Analyzing the effect of the Californian PFL program—the first in the United States—Bartel et al. (2018) report an increase in paternity leave uptake of 0.9%-points. Using within-U.S. variation of employment protection covering fathers, Han et al. (2009) show that American men are insensitive to legislation enabling leave. As an arguably closer comparison to the United States, Patnaik (2019) reports an increase in uptake rates of more than 50%-points after a reform in Quebec, which introduced 5 weeks of paternity leave. Moreover, important differences appear to exist between the introduction and the expansion of paternity leave. Evaluating two subsequent reforms in Sweden, Duvander and Johansson (2012) find the introduction of the first month of paternity leave in 1995 to have twice as great an effect as the expansion to 2 months in 2002. They also evaluated “a gender equality bonus” in 2006, which provided mothers with a tax credit if they shared leave equally, and find close to no effect on fathers’ leave duration. Table 2 provides an overview of the reforms, fathers’ use of leave prior to the reforms, and the reforms’ effect across countries.

Table 2. Fathers’ Take-Up of Earmarked Paternity Leave (“Daddy Quotas”) Upon Reform Implementation, by Country and Reform

Reform

Replacement Rate

Prior Take-Up

Reform Effect

References

Norway (1993)

Introduction (4 weeks)

80–100% of former earningsa

3%

32%-point

Dahl et al. (2014)

25%-point

Cools et al. (2015), Johnsen et al. (2020), Lappegård and Kornstad (2020)

Norway (2002)

Extension (2 weeks)

65%

Lappegård and Kornstad (2020)

Norway (2009)

Extension (4 weeks)

75%

3 weeks

Hart et al. (2019), Lappegård and Kornstad (2020)

Sweden (1995)

Introduction (1 month)

80% of former earningsa

30–38 days

50%-point

Ekberg et al. (2013)

15 days

Avdic and Karimi (2018)

10 days

Duvander and Johansson (2012)

Sweden (2002)

Extension (1 month)

37 days

5 days

Duvander and Johansson (2012)

Sweden (2008)

“Gender equality” bonus

Tax credit per day used by father

48 days

No effect

Duvander and Johansson (2012)

Denmark (1998)

Introduction (2 weeks)

90% of former earningsa

12–15 days

1–3 days

Andersen (2018), Druedahl et al. (2019)

Denmark (2002)

Removal (2 weeks)

14–22 days

Small reduction

Andersen (2018), Beuchert et al. (2016)

Iceland (2001)

Introduction (3 months)

80% of former earningsa

~0

82.4–86.6%-point

Olafsson and Steingrimsdottir (2020)

Canada (2006)

Introduction (5 weeks)

70% of former earningsa

22%

53.6%-point

Patnaik (2019), Wray (2020), Margolis et al. (2021)

Spain (2007)

Introduction (2 weeks)

100% of former earningsa

~ 0

55%-point

Farré and González (2019)

6–8 days

González and Zoabi (2021)

Germany (2007)

Introduction (2 months)

67% of net earningsa

4%

12%-point

Kluve and Tamm (2013), Schober (2014), Unterhofer and Wrohlich (2017), Tamm (2019), Cygan-Rehm (2016), Raute (2019)

United States (1993)

Introduction (12 weeks)

Unpaid leave with job protection

7.2%

3.9%-point

Han et al. (2009)

California (2003)

Introduction (6 weeks)

55% wage replacementa

2%

0.9%-point

Bartel et al. (2018)

a Benefits are capped at a ceiling.

Note: This table is not meant to be an exhaustive list of reforms by country but, rather, contains the reforms utilized by the studies included in this review.

Many studies have explored which factors and characteristics make fathers use parental leave. Using Swedish data, Ma et al. (2020) find that men who are young, foreign born, or earn a low income are less likely to take leave, explaining that this results from unstable labor market conditions. Descriptive evidence also highlights the importance of workplace characteristics (e.g., Bygren & Duvander, 2006; Geisler & Kreyenfeld, 2019; Kaufman & Petts, 2020; Naz, 2010) as well as relative income within couples, education levels, and number of previous children. Finally, the leave system plays an important role in uptake. Hook (2006) illustrates that paternity leave serves as an effective policy tool for increasing paternal involvement. Ray et al. (2010) emphasize how generosity and gender-egalitarian design of policy interrelate. Jørgensen and Søgaard (2021) document that uptake of paternity leave may be sluggish if benefits paid to fathers are low, highlighting the importance of wage replacement rates in influencing uptake of leave. Using data from 21 European countries, Castro-García and Pazos-Moran (2016) show that fathers take leave when it is nontransferable and payments are generous, whereas only a small minority take other types of leave. Mussino et al. (2019) compare the use of paternity leave among migrants in two culturally and economically similar countries—namely Sweden, with a long paternity leave, and Finland, with a short paternity leave—and find that migrants’ leave behavior is much more similar to that of the population in their country of residence than their country of birth, showing that policies enabling paternity leave are crucial for fathers’ uptake of leave.

Gender Equality in Time Allocation and Labor Market Outcomes

Paternity leave policies might affect gender equality via two channels: (a) by improving women’s labor market earnings relative to men’s and (b) by increasing the time fathers spend on child care and other tasks in the home. In most settings, gender equality improves with the introduction of paternity leave. Importantly, this is rarely driven by a meaningful reduction in fathers’ earnings but, rather, by a positive effect on mothers’ earnings and labor supply combined with more paternal involvement at home.

Examining the introduction of earmarked paternity leave in Norway in 1993, Cools et al. (2015) find no effect on Norwegian fathers’ work hours and yearly earnings, and Kotsadam and Finseraas (2011) discover that paternity leave leads to a more equal division of specific tasks in the household. Rege and Solli (2013) find that Norwegian fathers’ earnings are reduced with their uptake of leave, and by employing time-use data, they argue that this is driven by increased long-term paternal involvement, where fathers shift time and effort from the market to home production. Combined, these studies report that household specialization decreases with the introduction of paternity leave. Moreover, girls born immediately after the reform are less likely to do household work in adolescence (Kotsadam & Finseraas, 2013), showing that the equal sharing of household tasks persists into the next generation.

Dahl et al. (2014) document another type of social spillover. They find peer effects in workplace and family networks, as both brothers and co-workers of fathers initially affected by paternity leave reform take a longer leave themselves when they have a child. This effect depends on the strength of ties, with larger point estimates for brothers than co-workers. Moreover, the effect is transmitted in networks, creating a snowball effect that amplifies the initial impact of the reform and peer influence. Peer behavior likely provides fathers with relevant information about paternity leave, eventually leading to new norms of increased paternal involvement.

A related study by Johnsen et al. (2020) investigates the variations of relative leave induced by the reform and also finds effects on co-workers caused by the leave-taking behavior of fathers. These researchers observe that fathers’ own leave-taking does not affect their labor market trajectory when controlling for their relative eligibility status within the firm. However, fathers have higher earnings if a larger share of their co-workers are eligible for paternity leave. This suggests that paternity leave may negatively affect fathers’ earnings by causing them to lose out on high-wage positions to competing co-workers who do not take leave. Importantly, this effect is driven by the difference in eligibility and, in turn, leave-taking behavior. Dahl et al. (2014) show that the effect of the policy change might be greater than what is found when only comparing the couples with children born around the reform implementation period. Similarly, Johnsen et al. (2020) demonstrate that fathers other than those in the treatment group are affected by the reform. Norway further extended its paternity leave duration from 6 to 10 weeks in 2009, but Hart et al. (2019) find no effect on fathers’ or mothers’ subsequent earnings.

Sweden followed Norway’s lead by introducing 4 weeks of paternity leave in 1995. Johansson (2010) investigates the influence of the reform and finds that it had a negative although statistically insignificant effect on fathers’ earnings. Using the same reform, Avdic and Karimi (2018) also report a small reduction in fathers’ earnings, along with a small reduction in mothers’ earnings, which is mainly driven by mothers’ increase the use of unpaid leave. By using a measure of absenteeism from work in order to care for sick children, Ekberg et al. (2013) find that the reform did not have a long-term influence on paternal involvement in child care and uncover no effect on earnings of either mothers or fathers. Druedahl et al. (2019) use the Danish introduction of 2 weeks of earmarked paternity leave in 1998 and find that women’s earnings increased significantly while men’s decreased (albeit insignificantly). They explain that this effect is primarily driven by families wherein women are employed in the private sector.

Most studies on this topic have focused on the absolute duration of leave, whereas the relative difference in length of leave between the parents can act as an important driver of the gender wage gap because it may determine the division of labor within the household. Andersen (2018) examines five separate Danish parental leave reforms and observes that an increase in paternity leave relative to maternity leave leads to higher earnings for mothers. Pylkkänen and Smith (2004) compare Sweden and Denmark, which are culturally and ideologically similar but differ remarkably in parental leave policies over time: Sweden has provided much longer maternity and paternity leave compared to Denmark. They conclude that longer fathers’ leave shortens the mothers’ period away from work.

Although evidence from outside Scandinavia remains more limited, existing studies from Spain, Canada, and Germany all find evidence that lower gender specialization results from earmarked paternity leave. Couples affected by the reforms are more likely to move toward a dual-earner, dual-caregiver model. Farré and González (2019) use time-use data to investigate the effect of the Spanish introduction of 2-week paternity leave on fathers’ participation in child care and demonstrate that eligible fathers increase their time spent on child care compared to ineligible fathers. They find that fathers’ earnings are unaffected and a positive effect on mothers’ earnings occurs, driven by a reduction in unpaid leave.

Analyzing Quebec’s introduction of 5 weeks of paternity leave, Patnaik (2019) uses within-country variation in Canada. She documents that the time mothers spend on paid work and the time fathers spend on household responsibilities, including child care, increased, with no effect on fathers’ time spent on paid work. Using the same reform, Wray (2020) shows that fathers increased the time spent on solo parenting without their partner present.

A policy change in 2007 that introduced paternity leave in Germany simultaneously introduced changes in compensation rates and a shortening of total leave from 24 months to 14 months, making it difficult to separate the effects of the different changes. Tamm (2019) relies on within-father differences between first and subsequent children and reports that fathers’ leave-taking increases the time allocated to child care after their leave. Mothers’ working hours increased and fathers’ hours were reduced after a paternity leave, but these labor market effects are short-lived. Using a more standard reform evaluation framework to address the same reform, Kluve and Tamm (2013) find a small and insignificant effect on fathers’ time allocated to housework. Using survey data from West Germany, Schober (2014) finds that fathers with children born just after the 2007 reform spend more time on child care compared to those with children born before the reform, with no effect on housework. Similar to the evidence from Norway, spillovers to individuals in close proximity to the affected fathers seem to occur. Unterhofer and Wrohlich (2017) find that grandparents, particularly grandmothers, alter their view in support of working mothers when their son is given the opportunity to take paternity leave.

Fertility and Marriage

Because paternity leave can affect the household division of labor and shift the cost of child care from mothers to fathers, it might also affect other family outcomes, such as fertility and divorces. The transition to paternity leave can have mixed effects on fertility. On the one hand, it can increase fertility because having children becomes less costly for mothers’ careers. On the other hand, if mothers’ labor market attachment increases—and thus, their opportunity cost of subsequent children does too—fertility might decrease. Changes in costs for fathers should have symmetrical effects. The empirical evidence also shows that multiple effects are at play, and findings of the effect of paternity leave on fertility are mixed. Several studies have investigated the effect on the risk of couple dissolution, and all but one study find that paternity leave has a stabilizing effect.

Doepke and Kindermann (2019) reveal that the distribution of the parental burden is a key determinant of fertility. Farré and González (2019) find that 2 weeks of paid paternity leave in Spain reduces fertility, driven by a postponement of subsequent childbirths. They suggest that higher opportunity costs for mothers reduce fertility desires, but they also mention that increased paternal involvement might lower fathers’ fertility desire as the costs related to child care become more salient. Using Norwegian data covering a period of 25 years and exploring regional variation in uptake rates, Lappegård and Kornstad (2020) find that higher uptake rates correlate with higher fertility. This effect is particularly strong for second births. Evaluating the introduction and extension of paternity leave in Norway, Cools et al. (2015) and Hart et al. (2019), respectively, do not find any effect on fertility. Using the Belgian introduction of a short paternity leave around the time of childbirth, Fontenay and Tojerow (2020) find that birth spacing increased as a result of this leave-taking. As mentioned previously, the German reform in 2007, which increased replacement rates of benefits but lowered total leave duration while earmarking 2 months to fathers, also affected fertility. Raute (2019) shows that the reform increased fertility particularly among highly educated mothers. Cygan-Rehm (2016) finds that spacing between births increased, driven by low-income mothers.

Olafsson and Steingrimsdottir (2020) investigate the effect of the 2001 Icelandic paternity leave reform and find that this reform reduces separations for up to 15 years following childbirth, with greater effects within the first 5 years. They find larger effects in households in which the mother has a higher or similar educational attainment as the father. In households in which the father is more educated than the mother, the long-term effect on marital stability is negative. Margolis et al. (2021) investigate the introduction of paid paternity leave in Quebec. This reform also expanded eligibility and increased compensation rates for both mothers and fathers. They report lower separation rates in the first 5 years after childbirth and no difference in the following 3 years. Proxying gender norms with household characteristics, they find that both paternity leave uptake and separation rates are greater among couples that are likely to hold more egalitarian views.

Farré and González (2019) also investigate divorce rates and report that up to 3 years after childbirth, paternity leave appears to have a stabilizing effect on marriages, but the effect is insignificant in the following 3 years. There is no effect on divorces in Norway from the implementation of paternity leave (Cools et al., 2015) or its extension from 6 to 10 weeks (Hart et al., 2019), although Kotsadam and Finseraas (2011) show lower levels of self-reported conflict after its introduction. Cygan-Rehm et al. (2018) evaluate the German introduction of paternal leave and find that the reform reduced the risk of single motherhood. The effects are driven by households in which mothers are working. Because paternity leave was introduced at the same time as other changes to the leave system, the researchers cannot disentangle its effect from that of related policies, but they conclude that their findings indicate mothers’ improved financial situation and increased paternal involvement in child care.

Contrasting these findings of either no or positive effects on marital stability is one study using Swedish data. Avdic and Karimi (2018) investigate the introduction of parental leave and observe an increase in divorces within the first 5 years of the child’s life among low-income mothers, showing that couples who would likely have split up later drive this result. They also investigate the extension of earmarked paternity leave from 1 to 2 months in 2002 and find no effect on divorces. When comparing the results across countries, the effects on female labor market outcomes and income might be important. Farré and González (2019) and Patnaik (2019) find a positive effect on labor supply of women in Spain and Quebec.

The German reform increased benefits for a subset of households (Cygan-Rehm et al., 2018; Kluve & Tamm, 2013). However, Avdic and Karimi (2018) find an increase in unpaid maternity leave. Although the reforms used in these studies all offered paid paternity leave, the different responses by the households and other details of the reforms led to opposite effects on household income in Sweden compared to the other countries.

Gender norms provide another potential mechanism for reconciling the findings across countries. When children are born, most couples reorganize their lives toward more traditional family patterns, and this might cause conflict in couples that hold egalitarian views. Paternity leave might then have a stabilizing influence on these couples, but it may have a destabilizing influence on couples who prefer a high degree of specialization. The results from Sweden are potentially driven by households that would have chosen a more “conservative” allocation of time, whereas the heterogeneous results reported by Cygan-Rehm et al. (2018), Margolis et al. (2021), and Olafsson and Steingrimsdottir (2020) show that stabilizing effects appear greater in couples that are more likely to hold egalitarian views. Proposing a framework wherein some couples specialize while others do not, González and Zoabi (2021) revisit the Spanish reform of 2007. They identify the part of the population wherein the reform had the greatest effect on decreased specialization and document decreased fertility and an increased risk of divorce.

Children’s Health and Development

Involving fathers more in early child care might affect child outcomes along several dimensions, such as health and educational outcomes. Only a few studies have investigated this possibility, but they suggest a complementary relationship between maternal and paternal care. Cools et al. (2015) study the introduction of 1 month of paid paternity leave in Norway in 1993 and find that children’s school performance at age 16 years improves as a result of this program. They observe the most concentrated effect in families in which the father is better educated than the mother and highlight the importance of the idea that the effect of increasing paternal care will depend on the relative quality of the care it is replacing. The size of their estimates is larger than that reported by Liu and Skans (2010), who study an expansion of maternity leave in Sweden from 12 to 15 months. The reform studied by Cools et al. (2015) introduced 4 weeks of paid paternity leave on top of an existing 12-month leave scheme almost solely used by mothers. They argue that the nontrivial effect is likely driven by the long-term effect of the reform on household specialization and paternal involvement found in other studies (e.g., Kotsadam & Finseraas, 2011; Rege & Solli, 2013).

Using the Swedish introduction of paternity leave, which simultaneously reduced the total shareable leave, Ekberg et al. (2013) find that both the male share of child sick days and the total number of sick days are unaffected by the reform. Also using Swedish data and the introduction of double days in 2012, Persson and Rossin-Slater (2019) investigate the effect of the reform on child and maternal health. The reform allowed parents to be on leave at the same time and also allowed them to take leave intermittently, implying that fathers could choose, on a day-to-day basis, to stay home with the mother and child. They find a positive effect on maternal health measured by decreased contact with health providers and a decrease in usage of prescription drugs but no effect on child health. The double days are therefore used when the mother is unavailable to care for the child due to being sick. Similar to the research of Cools et al. (2015), this finding speaks to the potential synergistic effect of maternal and paternal care.

Summary

Earmarked paternity leave can increase fathers’ use of parental leave and family involvement, in turn ameliorating mothers’ household burdens while increasing their labor market participation and work hours. The literature reviewed herein shows that the introduction of paternity leave entitlements increases fathers’ usage of leave. Most studies have found that an increase in paternal involvement in child care results from paternal leave-taking, but the evidence on labor market effects for both mothers and fathers is mixed. In most cases, mothers’ labor supply and earnings rise with the increase in paternity leave. Studies of the earliest introduction of earmarked leave in Norway and Sweden have found a small reduction in fathers’ earnings, but more recent introductions of earmarked leave found no effect on fathers’ earnings. Moreover, there are important spillovers to other aspects of family life—most notably, couple stability and fertility. Few studies have investigated the effect of paternity leave on child outcomes, but the findings on paternity leave suggest that there are important complementarities between maternal and paternal care.

Parental Leave: Firms’ Perspective

Whereas a large body of work explores how parental leave affects households, less is known about their consequences for employers. In most countries, employers do not have to pay for the wages of workers on leave because these are typically funded through the social insurance system. However, employers may bear more indirect costs. Specifically, a worker’s absence due to parental leave leads to a decrease in the firm’s labor input. The costs of parental leave for the firm thus depend on its ability to effectively replace this lost labor input. This in turn hinges on the availability of substitutes for the absent worker within the firm or in local labor markets. Recently, a number of studies have examined the implications of parental leave for employers in different settings. In general, studies have focused on how employers are affected by (a) the introduction of short periods of paid parental leave, (b) reforms that extend the duration of paid leave, and (c) employee leave-taking.

Introduction of Paid Leave

Although the United States is the only OECD country with no national paid parental leave, several U.S. states have recently introduced PFL—which provided researchers the opportunity to evaluate how the introduction of leave affects employers. In 2004, California became the first state to give employees the right to take up to 6 weeks of partially paid leave. Other states followed, with New Jersey mandating up to 6 weeks of paid leave in 2009, and Rhode Island and New York introducing 4 and 8 weeks of leave in 2014 and 2018, respectively.

Using firm-level data from 2010 to 2018, Goldin et al. (2020) provide descriptive evidence on the type of U.S. firms that offer paid parental leave. They find that firms with generous paid parental leave tend to hire more workers who invest in firm-specific human capital and that they tend to be larger and have a younger workforce compared to other firms. Other studies focus on how employers are affected by the introduction of paid leave. Using surveys and in-depth interviews with employers, early descriptive evidence indicates that businesses in California and New Jersey experienced either positive or no noticeable changes in profitability, turnover, employee productivity, and morale (Appelbaum & Milkman, 2011; Lerner & Appelbaum, 2014; Milkman & Appelbaum, 2013). These results align with those of studies that place more emphasis on identifying causal effects. Bedard and Rossin-Slater (2016) use an employer fixed effects model along with administrative panel data from California and find that employees’ leave-taking slightly reduces firms’ wage bill and increases turnover. Several other studies use a difference-in-differences design that compares the change in employer outcomes in a state in which PFL was introduced with that of neighboring states, before and after PFL enactment. Bartel et al. (2016) surveyed small- and medium-sized food services and manufacturing businesses in Rhode Island, Connecticut, and Massachusetts in 2013 and 2015. They find that Rhode Island’s PFL enactment had no significant effect on businesses’ turnover, as well as employee productivity and morale. Bartel et al. (2021) also conducted a survey among employers in Pennsylvania and New York from 2016 to 2019. Their results show that New York’s PFL did not change turnover, employee performance, or the characteristics of the firms’ workforce. Furthermore, firms with more than 50 employees reported an increase in the ease of handling employee absences.

A study by Goodman et al. (2020) further shows that short periods of paid leave do not negatively affect employers, even when they have to pay for part of the wages of workers on leave. Goodman et al. focus on the 2017 introduction of the San Francisco Paid Parental Leave Ordinance, which requires employers to supplement California’s 6-week partial wage replacement. This guarantees employees access to fully paid leave—the first such program in the United States. Despite an increase in availability of paid leave, San Francisco employers report no changes in their performance or employees’ well-being.

Overall, these studies indicate that the introduction of short periods of paid leave does not significantly alter how businesses rate the performance and well-being of their workers.

Extensions in Duration of Paid Leave

Other work examines how employers are affected by reforms that increase the length of paid parental leave. Studies typically leverage reforms that unexpectedly extended the duration of paid leave. The unexpected nature of these reforms implies that firms are unable to plan in advance for worker absence, which can in turn limit their ability to efficiently compensate for lost labor input and can therefore negatively affect their performance. Ginja, Karimi, et al. (2020) focus on such a reform in Sweden, which extended the duration of paid leave from 12 to 15 months. The researchers find that the reform induces mothers to take an additional 2½ months of leave, but it also raises their likelihood of switching to another firm. Firms then make costly adjustments to compensate for the sudden increases in turnover and employees’ leave duration: They hire more temporary and permanent workers and increase work hours of co-workers of women on leave—resulting in a significant increase in the total wage bill. Following these adjustments, manufacturing firms experience decreases in revenues, sales, and value added by labor input, suggesting that replacement workers are less productive than women on leave.

Another study by Gallen (2019) examines a 2002 unexpected expansion in the length of paid parental leave from 10 to 32 weeks in Denmark. Gallen shows that small firms are more likely to shut down within 5 years of being exposed to the leave extension. Co-workers of women on leave see no significant changes in their earnings or employment rate, but they delay the timing of their own leave-taking and take more sick days due to the reform.

Finally, Huebener et al. (2021) evaluate a German reform that increased the amount of wage replacement for employees on leave from 3 to 12 months. They demonstrate that firms experience decreases in their employment and wage bill while an employee is on leave, suggesting that they are unable to fully compensate for the lost labor input. Furthermore, the researchers find suggestive evidence that firms are more likely to discriminate in their hiring against women of childbearing age. Taken together, these studies indicate that firms cannot effectively compensate for the lost labor input and experience a deterioration in their performance when exposed to unexpected extensions in the duration of parental leave.

Leave-Taking Events

Although reforms that change parental leave entitlements can have significant effects on employer outcomes, the event of having an employee take leave (versus not take leave) is equally consequential. Furthermore, the influence of leave-taking on firms can be different than the effect of the introduction or extension in the duration of paid leave. In the absence of changes in parental leave regulations, firms anticipate the timing and length of employee leave-taking, which can allow them to better plan for worker absence.

Using Danish administrative data from 2001 to 2013, Brenøe et al. (2020) examine how small firms cope with having a female employee give birth and take parental leave. They use a dynamic difference-in-differences design that compares employers of women who give birth to employers of women who do not give birth over the next few years. They first document that employers of women who give birth are exposed to an average of 9½ months of leave. Firms adjust to this leave-taking by hiring temporary workers and increasing retention and work hours of employees in the same occupation as the women on leave. As a result, firms’ total work hours remain unchanged, which indicates that these adjustments were effective at compensating for employee absence. The costs of these adjustments appear minimal. Despite increasing earnings of co-workers of women on leave, firms see no significant changes in their total wage bill. They also do not experience significant changes in their overall performance as measured by their output, profits, and likelihood of survival.

Replaceability of Workers on Leave

The costs of parental leave for firms depend on how well they can replace the absent workers. This in turn could be determined by labor market conditions and constraints facing the firm at the time of leave-taking. A study from Denmark highlights how labor market conditions may affect firms’ ability to adjust to leave-taking. Friedrich and Hackmann (2021) investigate how a 1-year extension in the duration of paid leave in 1994 (from 28 weeks of paid leave) affected hospitals and nursing homes. They show that the program led to a significant increase in nurses’ leave-taking. Because of stringent labor market regulations, employers were unable to replace nurses on leave, which led to a decrease in nurse employment. This in turn resulted in a significant decline in the quality of care provided by hospitals and nursing homes.

Another study suggests that the high costs of hiring and dismissing workers may limit firms’ ability to replace leave-takers. Schmutte and Skira (2020) find that in Brazil—a country with rigid labor laws—firms exposed to leave-taking only slightly increased their hiring and could not replace the absent worker at a one-to-one rate. Ginja, Karimi, et al. (2020) also provide evidence that labor market conditions affect how firms replace workers on leave. They show that in thick local labor markets, employers mainly increase hiring of new workers and do not change their existing workers’ work hours—with the opposite effects occurring in thin labor markets. Huebener et al. (2021) show that firms adjust by using both internal and external substitutes. Their findings reveal that firms use replacement hires more often when they have few internal substitutes (i.e., workers in the same occupation). They show that workers postpone their return from leave when internal substitutes are available. By exploiting an increase in the duration of paid leave, they demonstrate that this relationship between internal substitution and leave duration is greatly reduced, suggesting that coordination between workers and firms grows distorted by the increase in leave duration.

Finally, the substitutability of a firm’s employees affects how they fare with a co-worker’s leave-taking. Ginja, Karimi, et al. (2020) show that firms with a high fraction of same-occupation employees primarily increase work hours of their employees in response to leave-taking, whereas other firms rely more heavily on new hires. Brenøe et al. (2020) further find that firms with no other workers in the same occupation as the absent employee cannot fully adjust to leave-taking, despite having anticipated the leave. Specifically, the researchers show that these firms experience declines in their total work hours, wage bill, sales, gross profits, and survival.

Summary

In general, firms are able to adjust to worker absence due to parental leave. They compensate for the lost labor input by hiring new employees and/or increasing the work hours of existing employees. These adjustments are not costly and prevent firms from incurring losses in terms of their overall performance. However, certain factors—such as unexpected leave-taking, lack of substitutes for the worker on leave within the firm or in local labor markets, and high costs of hiring or dismissing new workers—may limit employers’ ability to replace workers on leave. This can result in high replacement costs and negative effects on firms’ performance.

Much progress has recently been made in studying firms’ response to parental leave, but some questions remain open. First, several U.S. studies show that the introduction of paid leave does not change employers’ rating of their performance. However, the lack of administrative data in these settings prevents understanding how employers adjust to the introduction of leave and how this affects the labor supply and well-being of other workers. Second, there remains no conclusive evidence on whether parental leave-taking results in statistical discrimination against women. Indeed, employers may limit the hiring and promotion of women of childbearing age in order to reduce their exposure to any consequences of parental leave-taking.

Conclusion and Future Research

Parental leave policies have evolved significantly since the mid-20th century. Following women’s entrance into the labor market, the focus of parental leave policies has changed from mother and child survival to parental labor market outcomes, family welfare, and child development. It is therefore important to evaluate various outcomes when examining parental leave policies.

Overall, parental leave policies prove highly important in helping parents balance job and family welfare responsibilities upon having children. Women are still the primary caregivers of newborns; thus, most leave policies remain targeted toward mothers. In general, there is an inverted U-shaped relationship between length of parental leave and most of the outcome variables. First, the introduction of short paid leave improves mothers’ labor market outcomes as well as their own health and the health of their children. Second, the findings show that extending the leave beyond 6 months has negligible effects on child development and the health of both mothers and children, and long-term leave affects mothers’ wages and employment negatively. Few studies focus on the heterogeneous effects of a long parental leave. Interestingly, the existing studies find that a parental leave that extends beyond 6 months negatively affects the income of highly educated women with specialized jobs the most. Long parental leave brings health benefits for women in lower SES groups but not for women in higher SES groups. Furthermore, it appears that a long parental leave may prove beneficial in terms of schooling outcomes for children born in higher SES families, whereas it adversely affects children in low SES families. The heterogeneous effects of family leave policies hold critical importance, and ample need remains for more studies on this topic.

The evidence on the effects of paternity leave is more mixed. Overall, studies show that introducing earmarked paternity leave proves effective in increasing fathers’ uptake rates and child care involvement. The evidence on mothers’ and fathers’ labor market outcomes varies across countries and policies. Findings show no to small positive effects of paternity leave on mothers’ earnings and no to small negative effects on fathers’ earnings. Paternity leave is also found to increase family stability and fertility, but again, the small amount of existing literature provides mixed findings on these topics. The nonmonetary effects of the different types of earmarked paternity leave are in general understudied, and more studies are required to make an overall conclusion.

Economists have only recently begun to study the effects of parental leave on firm performance, and many effects remain vastly understudied. In general, recent studies find that firms are able to compensate for the lost labor input from leave-taking employees, usually through hiring and increasing the workload for the remaining workers. However, firms that need to replace leave-taking employees in highly educated and specialized positions face higher replacement costs, which in turn can negatively affect productivity and firm performance.

Most studies thus far focus on the length and inter-parental distribution of parental leave. A demand exists for more studies focusing on the compensation rate and eligibility of the leave-taking workers. Over time, various reforms in different countries have changed the compensation rate for the entire duration of paid parental leave or parts thereof. It would be fruitful to gain a deeper understanding of how compensation rates affect uptake rates and the division of leave between the parents as well as the effects on parental and child welfare. Similarly, it would also prove interesting to study leave-taking behavior and fertility when eligibility for receiving parental leave benefits changes. Eligibility rules have largely changed over time and vary across countries. Whereas some countries have no eligibility requirements, in others, the parents need to have been in a full-time position for 1 year before becoming eligible for parental leave benefits. As future studies explore such research directions, decision-makers, firms, and individual workers will have a more well-rounded body of evidence to draw from in making decisions regarding parental leave.

Acknowledgments

We thank the two constructive reviewers for their valuable comments. Financial support from the Independent Research Fund Denmark is gratefully acknowledged.

References

Notes

  • 1. A few papers have also focused on early reforms in European countries to study the effect of introducing a relatively short leave. Bütikofer et al. (2021) study the influence of introducing paid parental leave in Norway in 1977, and Gregg et al. (2007) examine the introduction of job protection and 6 weeks of wage compensation in the United Kingdom in 1979.

  • 2. See Blanden and Rabe (2021) for an overview of the heteregoneity of the impact of child care, depending on the family’s SES, the quality of the child care, and whether the child care is replacing informal or parental care.

  • 3. They only observe a significant effect on marriages in communities where nurseries are available.

  • 4. For more on this mechanism, see the section on “Paternity Leave.”

  • 5. Very few articles have studied the effect of simultaneous leave. Persson and Rossin-Slater (2019) and Fontenay and Tojerow (2020) find that simultaneous leave improves maternal health in Sweden and Belgium, respectively. Andersen (2018) uses a series of reforms in Denmark, including the introduction of paternity leave around childbirth, and finds a positive effect on mothers’ income.