Abstract: We study an expansion in social protection to middle-income households that also induced beneficiaries to open digital accounts to receive their benefits. Using a regression discontinuity design and credit-bureau data, we study the impacts of the program by baseline account ownership. Households without prior accounts were more likely to receive the transfers in digital accounts. For these households, the program reduced financial distress by reducing delinquent utility and retail debt, increasing income, and preventing spending cuts. Among households without traditional bank loans or accounts at baseline, the program modestly increased credit access. We find no impacts among households with preexisting accounts.

Abstract: This paper studies how changes in family income influence domestic violence against women. I study municipalities in Colombia with high levels of coffee production and exploit negative shocks to coffee prices as an exogenous source of variation in household income. I find that municipalities with more intensive coffee production experience significant increases in domestic violence following a negative coffee price shock. Using individual-level data, I find similar effects for rural women residing in coffee-producing municipalities. Interestingly, women with decision-making power within the household are less likely to experience domestic violence in response to a negative income shock. This result highlights the role of female empowerment in mitigating the consequences of economic distress. 

Abstract: This paper studies how increasing short-term benefits from social protection affects insurance enrollment decisions and labor market behavior. I focus on a 2022 reform in Peru that expanded immediate access to contributory health insurance for pregnant women. During the period I study, the reform primarily changed incentives to affiliate during pregnancy by broadening access to health services and a one-time benefit at birth. Using data from the Peruvian household survey (ENAHO) and a dynamic difference-in-differences design comparing women at the same stage relative to childbirth before and after the reform, I find that the reform increases contributory enrollment around pregnancy and childbirth, especially through self-affiliation and dependent coverage. I also find little change in overall labor force participation, but some evidence of reallocation across employment categories around childbirth. Overall, the results suggest that expanding immediate access to short-term social protection benefits can shift the timing and form of insurance enrollment. More broadly, the patterns point to the importance of benefit timing in shaping behavioral responses in critical periods, such as giving birth. Future work will examine a subsequent reform that reinforced these incentives by extending immediate access to paid maternity leave.

Abstract: We document that lower income individuals in Santiago, Chile are exposed to higher concentrations of PM10 and PM2.5 while at home and when taking into account their locations throughout the entire day. The exposure gradient across income deciles when assuming individuals are at their residence 24 hours a day is similar to that when considering their location at each hour of the day as reported on the reference day in an origin-destination survey. We use a fuzzy regression discontinuity design to explore the effect of the declaration of critical air quality episodes, which activate city-wide restrictions on private vehicles among others, on air pollution concentrations. We find  that the declaration of critical air quality episodes leads to small and mostly statistically insignificant decreases in air pollution overall but the effects vary considerably across neighborhoods and hours of the day.      

Abstract: We study inequality in monitoring and exposure to particulate matter air pollution in four metropolitan areas of Latin America, Bogota, Colombia, Mexico City, Mexico, Gran Santiago, Chile, and Sao Paulo, Brazil. We find that the population residing in close proximity to at least one monitoring station in Bogota, Mexico City, and Sao Paulo generally have higher educational attainment and income. In contrast, in Gran Santiago, education levels are generally higher further from monitoring stations. Considering only census geographic units that contain a monitoring station, we find that areas where individuals with lower educational attainment reside tend to be exposed to higher pollution levels. While we find small and mostly insignificant disparities in mean annual concentrations of particulate matter, we find that lower education quintiles experience significantly more hours of extreme pollution relative to the highest education quintile. Non-linear effects of pollution imply that the small disparity in mean concentrations likely masks large disparities in the negative impacts of air pollution. Our findings indicate that in Bogota, Mexico City, and Sao Paulo, air pollution exposure is likely to be better monitored for those with higher educational attainment and income, and in all four cities, lower income and education groups have greater exposure to extreme levels of air pollution.