Elsevier

Journal of Health Economics

Volume 43, September 2015, Pages 85-102
Journal of Health Economics

The effects of paid maternity leave: Evidence from Temporary Disability Insurance

https://doi.org/10.1016/j.jhealeco.2015.04.005Get rights and content

Abstract

This paper investigates the effects of a large-scale paid maternity leave program on birth outcomes in the United States. In 1978, states with Temporary Disability Insurance (TDI) programs were required to start providing wage replacement benefits to pregnant women, substantially increasing access to antenatal and postnatal paid leave for working mothers. Using natality data, I find that TDI paid maternity leave reduces the share of low birth weight births by 3.2 percent, and the estimated treatment-on-the-treated effect is over 10 percent. It also decreases the likelihood of early term birth by 6.6 percent. Paid maternity leave has particularly large impacts on the children of unmarried and black mothers.

Introduction

In many countries, maternity leave is thought to be essential to both child and parent well-being. Maternity leave programs are designed to provide compensated and job-protected time off from work so that mothers can prepare for and recover from childbirth and new parents can stay home to care for their infants. The United States is the only industrialized nation in the world without a national policy granting new mothers access to paid time off from work. However, several states offer paid parental leave to qualified workers through state-run insurance programs or state-mandated employer coverage. In this paper, I examine the impact of the introduction of statewide access to paid maternity leave on birth outcomes. Importantly, this leave can be taken in the period before birth. Despite the popular belief that both antenatal and postnatal leave are beneficial to the infant and mother, there is little causal evidence from the United States suggesting that expanding access to antenatal paid leave improves infant outcomes. I find that the availability of statewide paid maternity leave reduces the share of low birth weight births by 0.22 percentage points, or approximately 3.2 percent.

Paid maternity leave can affect infant birth outcomes through several channels. The availability of paid time off may reduce physical and mental stress during pregnancy. Stress has been shown to have adverse effects on infant endowments of human capital, as measured by birth outcomes such as birth weight and gestational age (Aizer et al., 2012, Copper et al., 1996). Maternal prenatal stress is correlated with cognitive development in young children as well (Bergman et al., 2007). Access to paid antenatal leave may be especially important for single or low-educated mothers who are less likely to be able to afford to take unpaid time off and may have more physically demanding jobs. Jobs that involve standing, lifting, long hours, shift work, or noisy environments are all associated with negative birth outcomes (Mozurkewich et al., 2000). Additionally, while the income associated with paid leave may reduce mental stress, there are potentially also direct income effects for some families that could lead to improved birth outcomes. Finally, time off through paid maternity leave may impact choices regarding prenatal care visits or other types of medical care.

On October 30, 1978, Congress enacted the Pregnancy Discrimination Act (PDA). An amendment to the 1964 Civil Rights Act, the PDA prohibits discrimination in the labor force on the basis of pregnancy. This act requires that employers treat pregnancy and childbirth like any other temporary disability. Consequently, states with Temporary Disability Insurance (TDI) programs were required to start providing maternity leave benefits to pregnant workers. These benefits can be used in the period immediately before and after birth. This study evaluates the effects of the availability of this paid leave in the five states with TDI programs: California, Hawaii, New Jersey, New York, and Rhode Island. TDI covered approximately 50 percent of all women and over 80 percent of working women in these states in 1978, so access to paid maternity leave expanded substantially under this policy.

This paper uses a difference-in-differences (DD) framework to examine the impact of statewide access to short-term paid leave through TDI on the share of low birth weight births in the state. Low birth weight is defined as weighing less than 2500 g (approximately 5.5 pounds) at birth and is a strong predictor of newborn health and survival probability (Goldenberg and Culhane, 2007). To estimate this effect, I use Vital Statistics natality data from 1972 to 1985. Because these TDI programs all predate 1978, the PDA is plausibly an exogenous shock to the ability of pregnant women to claim disability insurance. This identification strategy allows for the estimation of intent-to-treat (ITT) effects of widespread access to paid maternity leave for working mothers. Doing so is possible because workers in non-TDI states do not have any rights to paid leave at the state level. Because it is likely that access to paid leave has heterogeneous effects across demographic groups, I also estimate separate effects for white/Hispanic, black, married, and unmarried mothers.

The results show that access to paid leave reduces the percentage of low birth weight births in a state and that this effect is strongest for unmarried mothers, who are more likely to be employed and thus benefit from TDI. While the ITT effects are modest, analysis of TDI maternity leave take-up in California suggests that the treatment-on-the-treated effects are quite substantial. Furthermore, I demonstrate that TDI increases both gestational age and birth weight at other points in the distribution. Paid leave reduces the proportion of early term births by over 1 percentage point, or 6.6 percent. Additionally, it reduces the share of births in the bottom quintile of the birth weight distribution by 0.39 percentage points, or 2.6 percent. These results are all consistent with the policy increasing antenatal leave taking. Data from the Survey of Income and Program Participation (SIPP) indicate that almost 70 percent of women in TDI states who take paid postnatal maternity leave also take time off before birth during this period. Therefore, it is plausible that the policy affects birth outcomes. While this is the first study to evaluate paid leave specifically, the results are in line with research on the effects of other types of leave-taking on birth outcomes in the United States. In a sample of U.S. mothers, Del Bono et al. (2012) find statistically significant effects of stopping work up to one month before giving birth on birth weight. However, they do not distinguish between different reasons for the work absence or whether the leave is paid or unpaid. Rossin (2011) shows that the availability of unpaid maternity leave in the U.S. increases birth weight and decreases the likelihood of premature birth, primarily for college-educated and married mothers. However, in contrast to studies of unpaid leave, I show that access to short-term paid maternity leave has substantial impacts on the children of unmarried and black mothers.

This paper is organized as follows: Section 2 provides background on maternity leave through TDI and other family leave programs in the United States. Section 3 reviews the relevant literature, and Section 4 describes the data used to evaluate the effects of this policy. The fifth section introduces the empirical strategy, and results are discussed in Section 6. Section 7 contains a discussion of possible mechanisms and Section 8 presents placebo tests. Finally, Section 9 provides concluding remarks.

Section snippets

Maternity leave in the United States

Women with young children make up a significant proportion of the labor force in the United States today. In 1975, 34 percent of mothers with children under the age of three were in the labor force; this number rose to 50 percent by 1985 and 61 percent by 2011 (U.S. Bureau of Labor Statistics, 2013). There are several explanations for this sharp rise in the percentage of working mothers. First, there has been an overall trend of increasing female labor force participation as gender roles have

Literature

The literature on paid maternity leave in other countries has produced mixed results as to its effectiveness in improving child outcomes. Studies that examine expansions of existing leave policies generally do not find significant effects on infants at birth or later in life. Liu and Skans (2010) show that increasing paid leave from twelve to fifteen months in Sweden has had no significant effects on either child or mother outcomes. Baker and Milligan, 2010, Baker and Milligan, 2011 find

Data

This study uses data on births in the United States from 1972 to 1985 to estimate the effects of TDI maternity leave coverage on birth outcomes. The natality data comes from the National Vital Statistics System (NVSS) of the National Center for Health Statistics. These individual-level records contain demographic information about the mother including her age, race, marital status, state and county of residence, and educational attainment. Mother's education is sufficiently poorly reported in

Empirical strategy

To identify the impact of statewide paid maternity leave coverage on birth outcomes, I employ a difference-in-differences framework. I compare birth outcomes within states before and after the implementation of the Pregnancy Discrimination Act in November 1978. I estimate the following equation:Ysmy=α+βTreatPostsmy+Wsmyϕ+δs+γmy+ɛsmyThe main outcome, Y, is the share of low birth weight (less than 2500 g) babies born in state s, month m, and year y. The variable Post is an indicator for

Effects of paid maternity leave on birth outcomes

Panel A of Table 3 presents the DD estimates of the intent-to-treat effect of TDI maternity leave coverage on the share of low birth weight births for each of the five TDI states separately as well as overall. To get the overall estimate in column (6), I combine the samples in the five previous columns and apply the same two-step approach. In the second stage, πˆmy are the estimated differences in the share of low birth weight births between the average of the five TDI states and the average of

Potential mechanisms and subsequent effects

The effects of TDI on birth outcomes seem large, but are consistent with other estimated effects of access to leave on the percentage of low birth weight births in the United States. These effects may work through several channels. The literature on the effects of physical and psychological stress on birth outcomes suggests that reductions in stress levels may be an important mechanism through which paid maternity leave improves birth outcomes. Antenatal leave may directly reduce job-related

Placebo tests

In this section, I do a placebo test using the synthetic control method to show that my identification strategy does not falsely identify treatment effects in non-TDI states. For each of the 46 states without TDI programs, I create a synthetic control group using the same selection criteria as above. I then falsely label the state as a treatment state and estimate the placebo treatment effect. Because none of these states actually have TDI programs, the treatment effect estimates should be

Conclusion

Leave policies that help working mothers balance family and career have many possible benefits. Paid family leave laws have been shown to positively impact maternal labor force participation rates as well as maternal health and well-being. However, evidence on the effectiveness of maternity leave at improving child outcomes is less conclusive. This study is the first to estimate the impact of state-based access to paid maternity leave on infant outcomes in the United States. I find that the

References (59)

  • Christopher J. Ruhm

    Parental leave and child health

    J. Health Econ.

    (2000)
  • A. Abadie et al.

    Synthetic control methods for comparative case studies: estimating the effect of California's Tobacco Control Program

    J. Am. Stat. Assoc.

    (2010)
  • A. Aizer et al.

    Maternal Stress and Child Outcomes: Evidence from Siblings

    (2012)
  • D. Almond et al.

    The costs of low birth weight

    Q. J. Econ.

    (2005)
  • D. Almond et al.

    Inside the war on poverty: the impact of food stamps on birth outcomes

    Rev. Econ. Stat.

    (2011)
  • M. Baker et al.

    Evidence from maternity leave expansions of the impact of maternal care on early child development

    J. Hum. Resour.

    (2010)
  • M. Baker et al.

    Maternity Leave and Children's Cognitive and Behavioral Development

    (2011)
  • J.R. Behrman et al.

    Returns to birthweight

    Rev. Econ. Stat.

    (2004)
  • L.M. Berger et al.

    Maternity leave, early maternal employment and child health and development in the US

    Econ. J.

    (2005)
  • S.E. Black et al.

    From the cradle to the labor market? The effect of birth weight on adult outcomes

    Q. J. Econ.

    (2007)
  • V. Brusentsev et al.

    Compensating for birth and adoption

    (2007)
  • A. Camacho

    Stress and birth weight: evidence from terrorist attacks

    Am. Econ. Rev.: Pap. Proc.

    (2008)
  • P.M. Carneiro et al.

    A Flying Start? Long-Term Consequences of Maternal Time Investments in Children During their First Year of Life

    (2010)
  • P. Cerón-Mireles et al.

    The risk of prematurity and small-for-gestational-age birth in Mexico City: the effects of working conditions and antenatal leave

    Am. J. Public Health

    (1996)
  • T.G. Conley et al.

    Inference with “difference in differences” with a small number of policy changes

    Rev. Econ. Stat.

    (2011)
  • C.C. Justin et al.

    Understanding heterogeneity in the effects of birth weight on adult cognition and wages

    J. Health Econ.

    (2015)
  • C. Crump et al.

    Early-term birth (3738 weeks) and mortality in young adulthood

    Epidemiology

    (2013)
  • J. Currie et al.

    Child health and young adult outcomes

    J. Hum. Resour.

    (2010)
  • G.B. Dahl et al.

    What Is the Case for Paid Maternity Leave?

    (2013)
  • Cited by (110)

    • War on Families: Limiting access to reproductive health

      2023, Children and Youth Services Review
    • Unequal use of social insurance benefits: The role of employers

      2023, Journal of Econometrics
      Citation Excerpt :

      A growing body of evidence demonstrates that access to temporary leave-taking social insurance programs, which allow individuals to take partially paid leave for their own medical issues or to care for new children or ill family members, has beneficial labor market and health effects on workers and their families (e.g., Rossin-Slater, 2018; Olivetti and Petrongolo, 2017; Stearns, 2015; Carneiro et al., 2015).

    View all citing articles on Scopus
    1

    I am grateful to Kelly Bedard, Peter Kuhn, Shelly Lundberg, Maya Rossin-Slater, and participants at the UCSB Economics Labor Lunch for helpful comments and suggestions.

    View full text