Policy Brief Issue 1 | March 2017 Could Childcare Services Improve Women’s Labor Market Outcomes in Indonesia? EAST ASIA AND PACIFIC GENDER By: Daniel Halim, Hillary Johnson and Elizaveta Perova INNOVATION LAB The East Asia and KEY FINDINGS Pacific Gender • Low female labor force participation is linked with unsatisfied childcare needs. Innovation Lab • Urban women without access to informal childcare forego approximately (EAPGIL) carries out US$1,300 in earnings due to prolonged absence from the labor market. impact evaluations and inferential research to • After childbirth, women without access to informal childcare are more likely to switch into, and remain in, unpaid family work. generate evidence on what works in closing • For women who return to work, childcare constraints are gender gaps in assets, associated with a switch into less lucrative occupations. economic opportunities and agency, and how closing these gaps CONTEXT can help achieve Increasing female labor force participation (LFP) is essential for growth and productivity other development gains1 and could offset the decline in the working population of a rapidly aging economy.2 outcomes. Ultimately, At 53.5 percent, female LFP in Indonesia is far below the regional average of 67.7 EAPGIL seeks to percent.3 increase the welfare of Studies from other regions show that supporting access to childcare increases female women and men in East LFP; however, the magnitude of these effects varies by country and by program type.4 To Asia and the Pacific by know whether similar programs might be appropriate for Indonesia, policy makers need to promoting the uptake understand whether low levels of female LFP are driven by preferences or constraints. of effective policies and programs identified 1 World Development Report 2012 – Gender Equality and Development, World Bank Washington, DC, 2011. based on evidence. 2 Live Long and Prosper – Aging in East Asia and Pacific, World Bank, Washington, DC, 2016. 3 World Development Indicators for 2014; http://wdi.worldbank.org/tables. 4 Cashing in on Education: Women, Childcare, and Prosperity in Latin America and the Caribbean, Mercedes Mateo- Diaz and Lourdes Rodriguez-Chamussy, World Bank and Inter-American Development Bank, Washington, DC, 2016. WHAT DID WE DO? We use household survey data5 to shed light on the following questions: • Do women’s labor market decisions in Indonesia reflect their preferences, or are they a consequence of childcare constraints? • What is the maximum amount that households may be willing to pay for childcare services? • If childcare constrains women’s choices, what is the economic cost in terms of foregone earnings? To answer these questions, we compared the labor market decisions of women with children who live in households with elderly household members and without elderly household members. The elderly may provide informal childcare services that alleviate the care burden. If labor market decisions are driven by preferences, we should not observe dramatic differences between women who have access to informal childcare and Figure 1a those who do not. PERCENT OF WOMEN WHO WORK / HAVE Due to data constraints, we can only describe trends and A CHILD UNDER 2 (URBAN AREAS) correlations; we cannot establish a causal relationship. 80% 60% WHAT DID WE FIND? 40% The childcare burden seems to pose a constraint, at least for some women. 20% Labour force participation patterns of women who live in rural and 0% 20 25 30 35 40 45 urban households with and without elderly members are strikingly Age different. Figures 1a and 1b show the percent of urban women No elderly Any elderly who report work or housekeeping as their main activity. Between Has child under 2 ages 26 and 28, when fertility peaks, the percent of urban women who work is higher among those who live with elderly by a margin of 10 to 19 percentage points. The same trend does not appear Figure 1b in rural areas. Rural women are more likely to be employed in PERCENT OF WOMEN WHO HOUSEKEEP / agriculture and less likely to be in formal employment, which may HAVE A CHILD UNDER 2 (URBAN AREAS) facilitate the combination of work and childcare. 100% In addition, between 2000 and 2014, the increase in female 80% LFP was commensurate with an increase in households with a 60% grandmother as the children’s primary caretaker. The share of 40% grandmothers as primary caregivers increased almost sevenfold 20% from a nearly negligible 0.8 percent to 5.7 percent, or 4.9 percentage points. During the same period, female LFP almost 0% 20 25 30 35 40 45 mirrored this growth, increasing by 3.9 percentage points. Age No elderly Any elderly These trends suggest that LFP decisions are not made purely Has child under 2 based on preferences. 5 Indonesian Family Life Survey (IFLS) rounds 1–5 (1993–2014). Unmet childcare needs have an economic cost in Figure 2 terms of foregone earnings. PROBABILITY OF WORKING IN PAST We conducted an event study to explore how the WEEK COMPARED TO YEAR BEFORE probability of being in the labor force and being in CHILDBIRTH (URBAN AREAS) different types of employment changes in response to .2 women’s first childbirth. We compared five years before .1 and eight years after childbirth to the year before giving 0 Probability birth. Again, we find striking differences between women –.1 in households with and without elderly and between rural and urban areas. –.2 –.3 In urban areas, the probability of working returns to −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 pre-pregnancy levels two years after childbirth for Years from childbirth No elderly (N=27,383) Any elderly (N=9,205) women in households with elderly members (figure 2).For those without elderly members, this period is Note: Dashed lines show 95% Confidence Interval.6 twice as long, only regaining pre-pregnancy levels four years after childbirth. Using data from the Labor Force Figure 3a Survey, we estimate the cost of lost earnings. Urban PROBABILITY OF BEING AN UNPAID women without access to informal childcare forego FAMILY WORKER COMPARED TO YEAR approximately US$1,300 in earnings due to prolonged BEFORE CHILDBIRTH (RURAL AREAS) absence from the labor market.7 .1 While informal childcare appears to accelerate the return .05 to work only in urban areas, the type of work women Probability 0 return to differs by the presence of elderly in both –.05 urban and rural areas. Figures 3a and 3b show that the –.1 probability of working as an unpaid family worker increases during the year of giving birth for women –.15 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 without the presence of elderly. However, for those Years from childbirth with elderly support, it remains unchanged in urban areas No elderly (N=13,810) Any elderly (N=5,150) and returns to pre-pregnancy levels within a year in rural Note: Dashed lines show 95% Confidence Interval.6 areas. Without access to informal childcare, women who transition into unpaid family work remain in it at least eight Figure 3b years after birth in urban areas, and in rural areas they only return to other activities after six years, which is the time PROBABILITY OF BEING AN UNPAID FAMILY WORKER COMPARED TO YEAR of school entry for their child. BEFORE CHILDBIRTH (URBAN AREAS) Women in households without elderly support are also .05 more likely to move out of manufacturing occupations in both rural and urban areas (figures 4a and 4b). The 0 Probability destination occupation varies by area. In rural areas, the likelihood of working in agriculture increases, while in –.05 urban areas women move into sales. These transitions are associated with US$319 and US$255 in foregone –.1 earnings in rural and urban areas, respectively.8 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 Years from childbirth Notably, the switch of sector is not temporary; the No elderly (N=16,877) Any elderly (N=6,195) likelihood of working in a specific sector does not return to Note: Dashed lines show 95% Confidence Interval.6 pre-pregnancy levels in either rural or urban areas. The solid lines show the probability estimate for the individuals interviewed. The actual probability for the population may be a little more or less - the 6  confidence interval shows the range in which we expect the actual probability to be with 95% confidence. For example, one year after childbirth urban women without elderly in the survey sample are 13 percentage points less likely to be employed than before giving birth, as indicated by the solid line in Figure 2. There is a 95% likelihood that the probability for all Indonesian women lies between -0.11 and -0.15 (shown by dashed lines). Labor Force Survey data show that the median income of urban women of peak age for first childbirth (22 to 24) is US$650 per year. 7  We estimate foregone earnings as the difference in median annual income in manufacturing and median income in agriculture or sales for women of peak age 8  for first childbirth (22 to 24) using the Labor Force Survey (LFS). Figure 4a PROBABILITY OF BEING IN MANUFACTURING / AGRICULTURE COMPARED TO YEAR BEFORE CHILDBIRTH (RURAL AREAS) Manufacturing Agriculture .05 .1 .05 0 Probability Probability –.05 0 –.1 –.05 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 ACKNOWLEDGMENTS Years from childbirth −5 −4 −3 −2 −1 0 1 2 Years from childbirth 3 4 5 6 7 8 We gratefully acknowledge Note: Dashed lines show 95% Confidence Interval.6 No elderly (N=18,246) Any elderly (N=6,595) comments and inputs of colleagues from the Social, Figure 4b Urban, Rural & Resilience PROBABILITY OF BEING IN MANUFACTURING / SALES Global Practice, the Poverty COMPARED TO YEAR BEFORE CHILDBIRTH (URBAN AREAS) & Equity Global Practice, Manufacturing Sales and the Africa Gender .05 .1 Innovation Lab. EAPGIL is supported through the World .05 Bank Group’s Umbrella Facility for Gender Equality Probability Probability 0 (UFGE) in partnership with 0 the Australian Department of Foreign Affairs and Trade. The UFGE has received –.05 –.05 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8 generous contributions from Years from childbirth Years from childbirth Australia, Canada, Denmark, Note: Dashed lines show 95% Confidence Interval.6 No elderly (N=23,863) Any elderly (N=8,149) Finland, Germany, Iceland, the Netherlands, Norway, WHAT ARE THE POLICY IMPLICATIONS? Spain, Sweden, Switzerland, First, differences in labor market decisions of women with and without immediate access the United Kingdom, to informal childcare suggest that constraints rather than preferences drive some women’s and the United States. decisions. There is likely to be a demand for childcare services among women in households without elderly members (67 percent of all Indonesian households with young children). Second, we estimate the maximum willingness to pay for childcare using data on earnings. If able to continue working after childbirth, urban women would likely earn FOR MORE INFORMATION US$650 on average per year. Public or private childcare services provided at a similar or higher cost would not provide a viable alternative to reducing LFP. Elizaveta Perova eperova@worldbank.org Third, childcare constraints have economic costs due to foregone earnings and switches into lower-paying occupations. Losses due to changing occupations are not temporary Helle Buchhave and highlight the need to explore constraints to combining market and household hbuchhave@worldbank.org roles in more lucrative professions. http://www.worldbank.org/eapgil Finally, future research should explore the causal relationship between childcare constraints and labor market choices and determine which public policies are most appropriate to address women’s childcare needs.