UC Berkeley

This dissertation studies redistributive policies that have the potential to reduce poverty and inequality. To investigate the economics behind such policies, I link large administrative datasets and exploit experimental and quasi-experimental variation in exposures to various treatments. This allows for identification of causal effects on a comprehensive set of margins associated with individual welfare and development.

The first chapter studies the impacts of reparations on well-being across the life cycle. Together with Juliana Londoño-Vélez and Christian Posso, I leverage variation from a reparations program for victims of the internal armed conflict of Colombia. The reparation consists of a one-off, lump-sum, non-means-tested, unconditional payment of up to 10,000 USD (PPP 26,000 USD) and represents, on average, three times recipients’ annual household income. We construct a novel and comprehensive administrative panel microdata that allows us to estimate impacts on work and living standards, health, and human capital investments up to four years after receiving the reparation. To identify causal effects, we exploit the staggered rollout of the transfers and their unanticipated receipt using event study and before/after approaches. We document three main findings. First, reparations induce an economically small shift of formal workers out of high-risk and low-earning jobs. Some of these workers spend more time non-employed and end up in higher-paying jobs, consistent with improved outside options. Others create firms or make their existing firms more sustainable, reflecting imperfect financial markets. Second, reparations induce meaningful reductions in health care utilization, likely due to improvements in health. Third, reparations enable victims to invest in the next generation’s human capital, raising test scores and postsecondary attendance. A stylized cost-benefit analysis suggests that reparations are cost-beneficial, with the large money transfers enabling victims to make critical investments to improve their long-term well-being.

In the second chapter of my dissertation, co-authored with Christian Posso, Estefania Saravia, and Jorge Tamayo, we study how more skilled physicians affect children’s health outcomes at birth. Identifying the effect of physicians on health outcomes is a challenging task due to the nonrandom sorting between physicians and hospitals. We overcome this challenge by exploiting a Colombian government program that randomly assigned 2,126 newly graduated physicians to 618 small hospitals. We estimate impacts on the 256,806 children whose mothers received care in those hospitals during pregnancy using administrative data from the program, vital statistics records, and individual records from mandatory health-specific college graduation exams. We find that more skilled physicians improve birth outcomes: a one standard deviation increase in the health graduation exam scores of physicians decreases the probability of giving birth to an unhealthy baby by 6.31 percent. Finally, we present evidence that one potential underlying mechanism is that more skilled physicians better target care toward more vulnerable mothers.

In the third chapter, co-authored with Christian Posso, Estefania Saravia, and Jorge Tamayo, we study the gender pay gap among physicians in Colombia. We do this by leveraging random selection into a nationwide public program with two notable ex-ante characteristics: guaranteed employment for one year, and pre-defined wages that cannot be bargained. We find that providing men and women with similar jobs at early stages of their careers reduces gender gaps in the probability of formal employment and earnings up to five years after the start of the program. One mechanism through which this occurs is by increasing women’s access to specialty training.