UC Davis

This dissertation studies potential causes for questions of particular relevance in developing countries.

In the first chapter, I use a Regression Discontinuity Design on close elections in Brazil to analyze how the election of a female mayor affects the gender composition gap of top municipal executives. I show that electing a female mayor in a close race increases the share of female managers by 17 %. This increase doesn't come at the cost of observed quality of employees. I also present evidence that public sector-specific and supply-side channels are unlikely to explain my findings; instead, an increase in gender-inclusive policies within female-led governments suggests that homophily may be a driver of the results. \

In the second chapter, focusing on the 2016 Brazilian election, my coauthors and I perform a similar analysis to chapter one but using race as the subject of analysis. We estimate the impact of electing a non-white mayor on a given municipality over the share of non-white municipal managers. We find that electing a non-mayor is not associated with any significant increase in the share of non-white managers. \

In the third chapter, we use a random corruption audit program in Brazil to cast light on the relationship between corruption and political participation. Different from other studies, we analyze corruption impacts at a different electoral level from where the corruption act took place. Empirical analysis shows that while being a standard deviation away from the mean of corruption violations and having random audits released before the election is not associated with a decrease in null voting on the local level, being in the same position of the corruption violation distribution and having random audits released prior to the election is associated with a 4% decrease in null voting in gubernatorial elections.