Gender Inequality | César Garro-Marín

Rooted Decisions: How Birthplace Shapes Women's Work

Thu, 04 Jun 2026 00:00:00 +0000

When we ask why women work more in some places than others, the instinct is to look at present-day conditions: available jobs, wages, childcare access, cultural attitudes today. These matter. But a growing body of research suggests that where a person grew up—and what they observed around them as a child—can be just as consequential as where they live now. Research on second-generation immigrants, for instance, finds that women’s work choices in their new country are shaped by the cultural norms of the country their parents left behind (Fernandez and Fogli, 2009).

This paper puts that idea to a precise test within a single country, using Indonesian data. Indonesia is a useful setting: it is a large, diverse country with 22 percentage points of spread between high- and low-employment districts, and a history of internal migration that makes it possible to separate childhood environment from adult environment.

The identification challenge

Studying the effect of childhood exposure is harder than it sounds. A woman living in a high-employment city today might work more simply because more jobs are available where she is—not because of where she grew up. To isolate the childhood effect, the paper compares women who moved away from their birthplace at different ages but ended up living in the same location as adults (Chetty and Hendren, 2018).

Women who left as teenagers had more childhood exposure to their origin’s labor market environment; women who left early in childhood had less. By comparing these groups—all observed in the same current location—the study separates where you grew up from where you are now.

The critical window: ages 6 to 14

Not all childhood years matter equally. The effects of early exposure to high-female-employment places are strongest when that exposure happens between ages 6 and 14—roughly primary school through early secondary. This is the period when children are most actively learning social roles and forming expectations about what the adults around them do. Evidence from psychology and economics identifies this as a key window for preference formation: adolescents are mature enough to form their own opinions but still receptive to external influences (Dhar et al., 2022; Olivetti et al., 2020).

Exposure concentrated before age 6, or primarily during adulthood, shows weaker effects. The formative years of late childhood and early adolescence appear to be when gender norms around work take deepest root.

How large is the effect?

Comparing a woman who grew up in a district at the 75th percentile of female employment to one from a district at the 25th percentile—a gap of 22 percentage points in local female employment rates—the paper finds that the first woman is about 5 percentage points more likely to work as an adult. This is a meaningful difference: the average female employment rate in Indonesia is around 42%, so the effect represents a roughly 12% increase.

Crucially, this effect holds even after women move to a different location. It reflects something internalized, not just a response to local conditions. Taken together, the patterns suggest that roughly 23% of the spatial inequality in women’s employment across Indonesia is transmitted from one generation to the next through childhood exposure to local gender norms.

A substantial share of today’s gender gaps in employment are, in part, yesterday’s gender gaps carried forward.

The evidence points to internalized gender norms—not skills or formal education—as the main channel (Jayachandran, 2021). The effects are not explained by differences in schooling or marriage patterns: high-FLFP areas actually have worse schooling completion rates, and women with more birthplace exposure do not choose husbands with different characteristics. The pattern is consistent with women who grew up seeing other women work updating their expectations about what women do—and those updated expectations staying with them.

What this means

Gender policies today create multiplier effects tomorrow. Raising women’s employment in a place does not just help current workers—it changes the environment in which the next generation of girls grows up. Some of that effect will carry forward even if those girls later move elsewhere.
The 6–14 window is a high-leverage target. Interventions that shape gender norms—role models in schools, community programs, representations of working women—may be most effective when aimed at this age group, when norms appear most malleable.
Spatial gender gaps are partly self-reinforcing. Low-employment areas raise daughters who are less likely to work, which perpetuates the gap across generations even as individuals migrate. Closing the gap requires changing the social environment where girls grow up, not just expanding economic opportunities.
Skills programs alone may miss the mechanism. If the channel is norms rather than human capital, policies focused only on job training or education are targeting the wrong lever. Complementary investments in changing social expectations may be necessary.

References

Chetty, R and N Hendren. 2018. “The impacts of neighborhoods on intergenerational mobility I: Childhood exposure effects.” Quarterly Journal of Economics, 133(3): 1107–1162.
Dhar, D, T Jain, and S Jayachandran. 2022. “Reshaping Adolescents’ Gender Attitudes: Evidence from a School-Based Experiment in India.” American Economic Review, 112(3): 899–927.
Fernandez, R and A Fogli. 2009. “Culture: An empirical investigation of beliefs, work, and fertility.” American Economic Journal: Macroeconomics, 1(1): 146–177.
Jayachandran, S. 2021. “Social Norms as a Barrier to Women’s Employment in Developing Countries.” NBER Technical Report 3.
Olivetti, C, E Patacchini, and Y Zenou. 2020. “Mothers, Peers, and Gender-Role Identity.” Journal of the European Economic Association, 18(1): 266–301.

Status: Working paper

Learn more: Full Paper

Who Benefits When Governments Build Schools?

Thu, 04 Jun 2026 00:00:00 +0000

When a government builds 6,000 new schools in 15 years, adding millions of seats for teenagers who could not previously access high school, it seems impossible for the effort to fall short. But “fall short” is not the right question. The right question is: who actually shows up?

In the late 1990s, only about 40% of Indonesian teenagers were enrolled in high school—well below the 60% rate in East Asia and the Pacific and 49% among middle-income countries (The World Bank, 2025). Starting in the early 2000s, the government launched an ambitious construction drive, building over 6,000 public high schools and adding more than 2.4 million seats over 15 years. The expansion targeted districts with the lowest enrollment rates, where the need was greatest.

The standard case for building more schools

The logic behind school construction programs is intuitive. Where schools are scarce and travel distances are long, young people don’t enroll. Build the school, lower the costs, and enrollment rises. If you target the most underserved districts, the gains should flow to those who need them most.

In mixed public-private education markets, there is an additional concern: will new public schools simply pull students away from private providers, leaving overall enrollment unchanged and damaging a sector that many families depend on?

What happened in Indonesia

Our paper compares districts that received new schools earlier against those that received them later, tracking enrollment and private school outcomes over time.

On enrollment, the expansion worked. Within five years of a district receiving new schools, upper-secondary enrollment rose by roughly 4 percentage points—about a 12% increase relative to the baseline rate. There were also positive spillover effects: middle school enrollment rose too, as students became more likely to stay in school when they could see a path forward to high school.

The new construction did not crowd out private schools. In the United States, increased public school funding has been shown to drive private school closures (Dinerstein and Smith, 2021). In Indonesia, private school construction in expansion districts did not slow down—if anything, it ticked up slightly in the years after the public expansion began. Private school students did not score worse on exit exams either. In settings where secondary enrollment is low to begin with, expanding public supply appears to grow the overall market rather than just redistribute existing students.

Enrollment gains concentrated among boys

The enrollment gains were deeply unequal across groups:

Boys from more educated families benefited most: high school attendance in this group rose by over 6 percentage points.
Girls saw smaller gains of about 3 percentage points, concentrated among girls from less-educated households.
Students from disadvantaged backgrounds gained far less at the high school level.

This pattern inverts what happens at the middle school level, where school expansions tend to reach the most disadvantaged students. The difference suggests that at the high school level, barriers go beyond physical access.

Building a school nearby likely reduces costs. But it does not address the other reasons families keep teenagers—especially girls—out of school.

Why the gap?

When secondary enrollment is already low, the students closest to enrolling are not the most disadvantaged—they are the ones who face the fewest other obstacles. Boys from educated families already have the demand; the new school removes the final constraint. Girls and children from poorer households face additional barriers: family economic pressure, safety concerns about longer journeys, and social norms that deprioritize education beyond a certain level. A new building does not change any of those.

What this means

School construction is necessary but not sufficient. Reducing physical and financial costs helps—but other barriers remain for girls and disadvantaged students. Scholarships, transport support, and community engagement are needed alongside new buildings to make sure they reach the students who need them most.
Public expansion need not hurt private schools. Policymakers worried that public construction will undermine private providers can take some reassurance. In low-enrollment contexts, there is room for both sectors to grow simultaneously.
Who benefits matters as much as how many benefit. A program that raises average enrollment while widening gaps between advantaged and disadvantaged students is not fulfilling its equity mandate. Monitoring enrollment by gender, parental education, and household income should be standard practice for any school-building program.
The case for demand-side policies is stronger than it looks. Conditional cash transfers, girls’ scholarships, and efforts to shift norms around girls’ education are not nice-to-haves—they are what determines whether supply-side investment reaches the intended beneficiaries.

References

Dinerstein, M and T D Smith. 2021. “Quantifying the supply response of private schools to public policies.” American Economic Review, 111(10): 3376–3417.
The World Bank. 2025. “Education Statistics – All Indicators (EdStats) DataBank.”

Co-authored with: Masyhur Hilmy (University of New South Wales)

Status: Working paper

Learn more: Full Paper

Work Hours and Amenity Trade-offs

Wed, 03 Jun 2026 00:00:00 +0000

When you accept a job, you’re not just agreeing to a salary. You’re agreeing to a schedule, a commute, a set of expectations about your time, and a package of benefits. Some jobs pay well but demand long hours. Others offer lower hours but provide fewer perks. This trade-off seems familiar.

But what exactly are workers trading against what?

The standard story

The classic economic answer is clean: workers who value flexible hours are willing to accept lower pay to get them. Firms respond by offering lower wages in exchange for better working conditions. Workers sort into jobs that match their preferences, and the wage gap between demanding and flexible jobs reveals how much people value those perks.

This is the theory of compensating differentials (Rosen, 1986)—and it has shaped how economists and policymakers think about salaries and job characteristics for decades. For example, if workers accept lower wages in exchange for lower hours, salary differences across jobs partly reflect workers preferences.

But jobs don’t offer only one perk, they often package several of them together. Moreover, some perks are not offered when others are. For example, jobs with flexible hours often do not offer health insurance, or the ability to work from home.

When firms package multiple perks together, workers often trade lower hours for other benefits—not just for wages—and this reshapes how we should understand salary differences in the labour market.

The bundling problem

Think of a job as a package deal. A job with flexible hours might also come with the option to work from home but a smaller health insurance contribution. A high-paying job might come with long hours but generous paid leave. Workers don’t pick individual features, they choose from a menu of bundles.

In our paper, we develop a model that accounts for how firms combine amenities (perks) and how workers decide whether a bundle suits their preferences. The key insight is that whether you end up with the amenity mix you want depends not just on what you value, but on what bundles firms offer and how those bundles are put together.

Flexible hours don’t just cost you wages

Using data tracking 6,755 Americans from their teens through adulthood (Bureau of Labor Statistics, 2024), we find that:

American workers who value lower hours often must give up other amenities.
Shorter work hours are frequently bundled with fewer other benefits, such as less paid leave or health insurance.
The simple trade-off between wages and work hours that standard theory predicts is far less common than the trade-off between one amenity and another

Workers are not just choosing between money and work hours. They’re navigating a menu of bundles where getting one benefit often means giving up another.

Who ends up where, and the gender gap

These patterns matter especially for understanding why men and women end up in different jobs with different pay.

Women are more likely than men to value flexible hours, particularly those with caregiving responsibilities:

Women are more likely to end up in jobs with flexible hours—but these jobs also tend to bundle in fewer other benefits
As a result, women face a compounded disadvantage: lower wages and a thinner non-wage benefits package
Thus gender gap in total compensation—wages plus the full value of benefits—is wider than the gender wage gap alone suggests.

What this means

Wage statistics understate inequality. If men receive better non-wage benefits on top of higher wages, standard pay comparisons miss a significant share of the gap. Policies and audits that focus only on wages may give a misleadingly narrow picture.
Flexible work alone is not enough. The remaining gender gap in labour market outcomes is closely tied to how flexibility in hours is packaged alongside other benefits (Goldin, 2014). Expanding part-time or reduced-hour arrangements only helps if those arrangements don’t strip away other perks in the process.

References

Bureau of Labor Statistics, U.S. Department of Labor (2024), “National Longitudinal Survey of Youth 1997 cohort, 1997–2021 (rounds 1–20)”, produced and distributed by the Center for Human Resource Research (CHRR), The Ohio State University.
Goldin, C (2014), “A grand gender convergence: Its last chapter”, American Economic Review 104(4): 1091–1119.
Rosen, S (1986), “The theory of equalizing differences”, Handbook of Labor Economics 1: 641–692.

Co-authored with: Neil Thakral (Brown University) and Linh Tô (Boston University)

Published in: AEA Papers & Proceedings, Vol. 115, May 2025

Learn more: Full Paper