Research

We investigate how a private market deregulation affected the incentives for monitoring gold transactions in Brazil, ultimately leading to an increase in illegal gold mining and violence. Employing a Difference-in-Differences design and a unique database that combines the geological occurrence of gold deposits and protected areas, where mining is forbidden, we first show that the deregulation encouraged illegal gold mining. We use high-resolution data on deforestation as a proxy for illegal mining activity. Then, we demonstrate that municipalities more exposed to illegal gold mining experienced almost eleven additional homicides per 100,000 people - roughly 30% more - after the deregulation. 

We study a Force-down/Shoot-down intervention in Brazil that led cocaine traffickers to shift from air to river routes. Using data on cocaine production, homicides, and the network of rivers in the Amazon, we provide evidence that violence increased in municipalities along river routes originating from Andean producing countries after the policy. We also show that, during the same period, violence in these municipalities became more responsive to cocaine production in origin countries. We document an instance of crime displacement over the three-dimensional space, involving sophisticated adaptations from criminals regarding transportation technologies, with dramatic side-effects for local populations.

Incumbents’ reputations acquired during their time in office influence voters’ choices. Some studies suggest that, when designing their campaign messages, incumbents might benefit from playing to their strengths and emphasizing fewer issues than challengers. In this paper, I provide causal empirical evidence to support this hypothesis using information from mayoral elections in Brazil and a regression discontinuity design. The findings indicate that incumbency causes campaign manifestos to be more concentrated. Moreover, incumbents who belong to a cohesive and programmatic party tend to emphasize party-specific issues more. This suggests that reputation matters for issue selection. To measure issue emphasis and manifesto concentration, I leverage a unique database of more than 28 thousand documents containing political manifestos and use a Latent Dirichlet Allocation (LDA) algorithm to estimate a vector of topics – or the issues emphasized by each candidate. With this information, I calculate Herfindahl Hirschman Indices (HHI) for each manifesto to assess their relative issue concentration.

We examine the impact of short-term temperature changes on voting behavior in the Brazilian presidential elections from 2010 to 2018, focusing on the candidacy of Marina Silva, a prominent pro-environment politician. Using data on municipality-level voting and daily wet bulb temperature, we find that an increase in temperature on election day increases votes for Marina Silva. We argue that this effect is driven by the salience of climate change issues, which influences last-minute voting decisions. Our analysis also shows that temperature variations in the days preceding the election have a limited impact on voting outcomes, highlighting the transient nature of climate salience effects.

In this work, I empirically investigate aspects of the electoral competition theories by taking advantage of computational linguistic methods that transform texts into data (i.e., LDA model and Wordscores) and employing those methods to read the political platforms of Brazilian mayoral candidates. I start by building a unique dataset containing the topics and the partisan-scores of those candidates and describing its contents.  Then, I use this dataset to investigate, from a Downsian location model perspective, the relationship between campaign spending and platform choice. I test whether a candidate’s platform positioning strategy, compared to her opponent, is influenced by the amount of money she can spend in an election. Finally, I also empirically test the theoretical relationship between reputation and platform choice, proposed by the Salience Theory, by estimating the incumbency effect on a candidate’s platform issue concentration.

In a nutshell, I find that candidates make strategic decisions concerning their platform’s contents and partisan leaning. The data’s descriptive analysis shows a systematic correlation between the candidate’s characteristics and her platform’s contents. Moreover, the empirical evidence supports the theories that connect platform choice to reputation but not the theories that establish a link between platform choice and campaign spending. In particular, I find that opposing candidates’ strategies to differentiate their platforms do not change significantly depending on the size of their campaign expenditures. Furthermore, I show that incumbents tend to produce more concentrated platforms than challengers.