UC Berkeley

This dissertation studies redistributive tax and transfer policies in developing countries. It examines how these policies can be shaped to level the playing field, promote equal opportunity, and reduce inequalities. To tackle these questions, it exploits large administrative datasets and original survey data, and focuses on empirical settings where plausibly exogenous variation can be leveraged using program evaluation methods and quasi-experimental techniques.

In recent years, wealth taxation has received renewed interest as a way to raise revenue and address inequality. The first chapter of my dissertation studies how wealthy individuals respond to personal wealth taxes and wealth tax enforcement. This chapter is co-authored with Javier Avila-Mahecha. We use Colombian tax return microdata from 1993 to 2016 linked with the leaked "Panama Papers." We exploit discrete jumps in wealth tax liability and reforms varying exemption cutoffs and tax rates. We find clear evidence that individuals lower their reported wealth to reduce their tax burden, especially by misreporting items subject to less third-party reporting. We complement this analysis by studying offshore sheltering in Colombia's most relevant tax havens. Offshore entities have become more popular since the reintroduction of wealth taxes in Colombia. These entities are predominantly used by the wealthiest taxpayers, at least in part to hide assets from the tax authority. Finally, better enforcement helps recover tax on offshore wealth. A voluntary disclosure scheme revealed wealth worth 1.7 percent of GDP, most of it having been hidden offshore. The wealthiest individuals disproportionately disclosed under the scheme and, as a result, pay more taxes. Halfway through the scheme, the Panama Papers news story broke, raising disclosures by more than 800 percent. The higher perceived detection probabilities, coupled with harsher noncompliance sanctions, improved wealth tax collection and enhanced tax progressivity at the top.

In addition to progressive taxation, governments may reduce inequality by leveling access to colleges through financial aid. The second chapter, co-authored with Fabio Sanchez and Catherine Rodriguez, exploits a large-scale financial aid program in Colombia for low-income high-achievers to study how aid affects postsecondary enrollment, college choice, and student composition. Our regression discontinuity estimates show aid eligibility raised immediate enrollment by 56.5 to 86.5 percent, depending on the complier population. This rise, driven by matriculation at private high-quality colleges, closed the SES enrollment gap among high-achievers. Moreover, a difference-in-differences approach suggests enrollment of aid-ineligible students also improved because college supply expanded in response to heightened demand. With ability stratification largely replacing SES stratification, diversity increased 46 percent at private high-quality colleges.

This large expansion in socioeconomic diversity has the potential to deeply affect interactions between low- and high-income students at elite private colleges. In the third chapter, I shift attention to the high-income classmates of aid recipients and ask, does socioeconomic diversity shape their perceptions of inequality and their preferences for redistribution? I collect original survey data and conduct survey experiments to measure these outcomes, which are not available in the administrative records. I exploit the plausibly exogenous timing of the aforementioned financial aid program and test the effect of this shock in peer characteristics. Specifically, I test the effect of exposure to low-income classmates on high-income students' interactions with low-income peers, and their perceptions of inequality and poverty, beliefs of social justice, and redistributive preferences. I find significant positive effects of exposure to low-income peers on interactions among students with heterogeneous family backgrounds, and an increase in perceived inequality, poverty and upward social mobility, and meritocracy in college admissions. I also find a stronger support for redistribution. Lastly, I rule out any negative effects on academic performance in college.