Job Market Paper
Abstract: The aging of the population is increasing the demand for adult caregiving. In most of the world, care for the elderly and sick is provided almost exclusively by families and, within families, by women. This paper studies the impact of adult caregiving on gender inequality in the labor market. Using administrative data from Chile, we leverage variation in a parental health shock —the first cancer hospitalization of a parent— to examine who bears the burden of adult caregiving. After a parental health shock, daughters but not sons experience a reduction in employment and earnings. A parental health shock creates a caregiving penalty —the effect of the shock on daughters relative to sons— of 11% on earnings, increasing the overall gender pay gap by 9%. These penalties affect women even if they earn more than their partners or brothers, suggesting that gender norms influence the distribution of adult caregiving. Additionally, penalties are concentrated among women who are mothers, suggesting a correlation across the life cycle between care given to children and then to aging parents.
Published and Forthcoming Papers
Abstract: We leverage staggered implementation of lockdown across Chile’s 346 municipalities, identifying dynamic impacts on domestic violence (DV). Using administrative data, we find lockdown imposition increases indicators of DV-related distress, while decreasing DV reports to the police. We identify male job loss as a mechanism driving distress, and female job loss as driving decreased reporting. Stimulus payments to poor households act on both margins, their impacts partially differentiated by lockdown status. Once lockdown is lifted, police reports surge but we see a ratchet effect in distress. Our findings accentuate the controversy around welfare impacts of lockdown mandates.
Is there a Critical Mass? Gender Composition and Behavior in US City Councils [link]
Presented at NBER SI 2023 Political Economy, ASSA 2023 Annual Meeting (scheduled)
Abstract: How does gender composition influence individual and group behavior? To study this question empirically, we assembled a new, national sample of United States city council elections. Within this sample, we digitized information from the minutes of over 40,000 city-council meetings. These documents describe the number and type of motions proposed by each councilor as well as the outcomes of consequential votes pertaining to the operation of city government. The resulting data includes rich measures of group behavior, individual behavior, city council gender composition, and municipal expenditures. Using a close-election regression discontinuity design, we find that replacing a male councilor with a female councilor results in a 25p.p. increase in the share of motions proposed by women. This is despite causing only a 20p.p. increase in the council female share. We find that the discrepancy is driven, in part, by behavioral changes by isolated female councilors. When a lone woman is joined by a new female colleague, she participates more actively in council discussion by proposing more motions. However, these apparent changes in behavior do not translate into clear differences in spending. Taken together, our results highlight the importance of ``tokenism'' for inhibiting female participation in a real-world, high-stakes setting, and provide mixed evidence on the importance of having a ``critical mass'' of female representation for affecting outcomes in political decision-making bodies.
Can Early Interventions in Children and Families Reduce Child Maltreatment? [Draft pending upon disclosure approval]
Presented at ASSA 2023 Annual Meeting (scheduled)
Abstract: Children with a disability are three and a half times more likely to be maltreated. We study whether the risk of future maltreatment declines when children with a disability receive federal Early Intervention (EI) services. Early intervention’s model is built on the premise of parental cooperation, with the goal being to teach parents how best to care for their children with disabilities. Given that maltreatment often initiates in infancy, we study whether receipt of EI services before 12 months of age results in reduced maltreatment later, after age 3 when children have aged out of EI. Children referred earlier to EI, however, are more disadvantaged and are at higher underlying risk of maltreatment. To address the endogeneity of early receipt of services, we construct a control group based on children who were referred early to EI, but found to be ineligible. Using this control group to address selection on unobservables, we find that if a child receives services in the first year of life, the probability of maltreatment after age 3 declines by 2.5 percentage points, or 34%. This is driven by reductions in neglect, not abuse. Given that child maltreatment usually originates in infancy, these results underscore the importance of influencing parental behavior early, before maltreatment begins or is firmly established. It also underscores the efficacy of targeting children (not necessarily parents) at high risk of maltreatment and of using a cooperative model of engagement with parents.
Work in Progress
Labor Market Returns to Graduate Degrees: Evidence from a Scholarship Program (analysis in progress)
Family Health Shocks and Gender Inequality in Educational Investments (data access secured)
The Labor Market Effects of Having a Family Member with Dementia (data access secured)
Mental Health Consequences of Informal Caregiving (data access secured)
What Makes Childcare Effective at Fostering Mothers' Employment? Evidence from Chile's Public Spots Assignment
Women in Male-dominated Fields in Economics: Can Female Faculty Attract More Female Students? (analysis in progress)