DescriptionThis dissertation examines higher education and social policies in developing countries. The first chapter evaluates how attending college offers a wage premium in the long term while also impacting financial decision-making. The second and third chapters examines the extent to which safety net benefits provide incentives inducing socially desirable behavior and reallocate resources within households. To study these questions, I apply modern causal inference techniques to exploit the plausibly exogenous variation from these policies' designs and construct new datasets from administrative information systems and survey data.
Chapter 1 examines how college attendance affects financial decisions in the credit market, providing the first such estimate in a developing country setting. This chapter is coauthored with Cristian Posso. I follow the applicants to a selective college in Colombia by constructing a novel administrative dataset of labor market outcomes from social security records combined with banks’ lending operations reported to the Financial Regulation Office in Colombia. The resulting dataset covers 18 years after college entry for applicants to this public flagship university, allowing us to compare the short- and long-term effects. Using the discontinuity in the admission test scores of this flagship public university in Colombia, I find that attending selective public universities has no effect two years after graduation but boosts credit card usage and earnings eight years after college entry. In the long term, students attending the selective public university are more likely to be homeowners by the time they are 30 to 35 years old. The college education impact on earnings growth and job quality likely explains the gains in credit market access.
In the next chapter, I shift attention to the Covid-19 pandemic and the emergency policies implemented to help disadvantaged households. This research, coauthored with Juan Mogollon and Catalina Villamil, delves into whether receiving food assistance could incentivize households to remain at home and increase social distancing during the critical months under lockdown in Colombia. The impact is estimated using the quasi-exogenous rollout of the food distribution program within municipalities. We construct a high-frequency data set linking detailed daily deliveries with georeferenced food recipients' location and mobility indicators measuring out-of-home events. The main findings indicate that receiving food assistance delivered to the home reduced out-of-home mobility by 1.6 percent during the first two weeks after delivery. The program reduced out-of-home activities by reducing visits to grocery stores.
Social safety net policies can also affect labor market participation for beneficiaries and their families. In Chapter 3, I seek to understand how a social pension program destined for disadvantaged elderly individuals affects labor market participation of working-age mothers living in the same household. To do this, I use the eligibility for the Colombia Mayor Program (Elderly Adults Program), a social pension program transferring income to poor adults of retirement age. I use a regression discontinuity approach based on the beneficiaries scores in the social assistance index qualifying for the pension. I show that grandparents who narrowly qualified for the pension and live in extended families have lower labor force participation than their otherwise similar counterparts who narrowly failed to qualify for the pension. This paper hypothesizes that the additional grandparent time available for childcare increases labor market participation for mothers with children living in these households. Consistent with this, I find a positive estimate of 5.4 percentage points on labor force participation for mothers, with a higher effect for mothers with children between 0 and 5 years old. Grandparental support is still an essential source of low-cost childcare in developing economies.