Illegal markets, violence, and local economic impacts: evidence from gold mining in the Brazilian Amazon

Abstract

Using a rich set of data covering more than 540 municipalities in the Brazilian Amazon biome over 20 years, we investigate the effects of economic and institutional changes on violent crime associated with illegal gold mining. To identify causal effects, we exploit the exogeneity of a rare episode of market deregulation and a boom in international gold prices, combined with the random distribution of geological deposits in protected areas, where mining is strictly prohibited. Employing a difference-in-differences design, we document that areas more exposed to illegal mining experienced increases in violent crime following market deregulation and the gold price boom. In our analysis of mechanisms, we find stronger support for the rapacity hypothesis, which posits that higher returns from illegal extraction incentivize the plundering of gold deposits and increase local violence. Consistent with this hypothesis, we observe a significant increase in land use for gold mining (garimpo) in areas more exposed to illegal mining. We also document a deterioration in labor market conditions, reflected in declines in formal employment, as well as a relative reduction in the importance of royalty revenues in municipal budgets, providing evidence of a complementary role for the opportunity-cost mechanism. In policy terms, our theoretical framework suggests that violence associated with illegal gold mining can be mitigated through both centralized and decentralized monitoring. Although both are theoretically effective, the comparison between the two institutional designs suggests that decentralized monitoring may involve relatively lower implementation costs, as it relies on private incentives along the production chain.