| The Brazilian Amazon is one of world's most valued natural treasures, but has been subject to extraordinary deforestation over the past half-century. Despite expansive research on Amazonian deforestation and its drivers, the role of local politics in deforestation is not well understood. Using various applications of the political agency model, I examine the extent to which political processes influence deforestation in the Brazilian Amazon, specifically exploring the effect of electoral incentives.;The first chapter of my dissertation identifies a deforestation cycle that coincides with elections. Political economy theory, supported by empirical evidence, demonstrates that incumbent politicians tend to distort monetary or fiscal policy or public goods provision in election years to boost their popularity and chances of re-election. Few, however, have explored election cycles in natural resource use. Using a panel dataset that combines municipal-level deforestation and election data for 807 counties in the Brazilian Amazon from 2000 to 2013, I estimate the effect of an incumbent mayor running for re-election on deforestation rates in and around election years. I find that deforestation rates increase 8-10% in election years, and this effect is stronger among corrupt mayors.;The second and third chapters of my dissertation explore the effect of term-limits on deforestation in the Amazon. Absent re-election incentives, politicians no longer cater to the desires of the voting population and engage in self-interested behavior. This self-interested behavior can manifest as a politician implementing his preferred policies, or can manifest as an increase in rent-seeking behavior. Using fixed effects model and regression discontinuity design approaches, I find that term-limits effects on deforestation are not driven by agricultural policy distortions, and limited evidence that term limits effects are driven by rent-seeking behavior.;Several interesting findings emerge from this work. Second-term mayors serving in communities with many agricultural interest groups demonstrate significantly higher deforestation rates than second-term mayors in areas with few agricultural interest groups. These results suggest that, as they leave office, mayors may cater to interest groups in exchange for future political favors. I present evidence that corrupt mayors deforest more in the second term compared with the first in non-election years. I also find that corrupt mayors who are affiliated with the governing coalition party deforest significantly more than corrupt, not affiliated mayors. These results suggest interesting dynamics between corruption and political connectedness that merit further investigation.;This research shows that local elections create incentives to manipulate forest resources, and that corruption exacerbates these effects. These findings suggest that democratic processes may be in conflict with conservation goals, particularly in areas with weak governance. Compared with other forest-rich countries, Brazil's forests are relatively well-managed, with centralized conservation policy and enforcement, implying that local political processes could have an even stronger impact on forests located in areas with weaker resource governance and decentralized forest management. This work also demonstrates that forest resources are susceptible to political manipulation, and may be especially vulnerable since forest exploitation is more easily hidden than other rent-seeking activities, and difficult to attribute to a single actor. |
