R&R Journal of Public Economics.
Joint with Alberto Chong.
We provide evidence of the Peltzman effect, which predicts that individuals who are required to wear protective gear offset the benefits by taking additional risks. The Ice Hockey visor is a protective devise whose use is mandatory in some leagues and not in others. We compare players who always wear a visor with players who only wear one when required to. Due to differences in league regulations, as the latter group of players change leagues their use of the visor changes too, which allow us to estimate the effect of mandatory visor use on behavior. We find that mandatory visors substantially increase risky behavior. Due to their risky and aggressive behavior on the ice rink, players obliged to wear a visor are penalized 0.18 additional minutes per game. This is a large effect. In our sample the average number of penalty minutes per player per game is 0.8. The estimated effects are not driven by characteristics of players, playing style, or differences in the penalization standards of different leagues.
Joint with Daniel Mejia.
This paper explores the effects of cocaine booms on violence in Colombia during the period of 1990 to 2011. We exploit variation across municipalities in their suitability for cocaine production and variation over time in the demand for Colombian cocaine. Our suitability, based on a survey of coca farmers, depends on the altitude, erosion, soil aptitude, and precipitation of a municipality. The measure of external demand combines proxies for consumption in the U.S. with other forces that shape the demand for Colombian cocaine, including changes in enforcement in transit countries and other source countries. We find that increases in the external demand for Colombian cocaine creates cocaine booms in municipalities that have a high suitability for coca cultivation. In these communities homicides increase significantly both in the short and the long run. We argue that, on the whole, these cocaine booms would not have caused violence if cocaine was legal. To support this claim, we show that cocaine booms bring violence independently of state presence; while booms in other commodities (cocoa, sugar cane and palm oil) typically reduce violence, except in a few communities where state institutions are too weak.
Joint with Daniel Mejia and Juan Camilo Castillo.
This paper explores if scarcity increases violence in markets without a centralized authority. We construct a model in which temporal supply shortages increase contested revenues and induce more violence. To test our model we examine the cocaine trade in Mexico. Scarcity created by cocaine seizures in Colombia – Mexico’s main supplier – increases violence in Mexico, especially in municipalities near the U.S. border, in municipalities with multiple cartels, and where crackdowns on the cocaine trade are more frequent. We conclude that between 2006 and 2009, a sharp decline in the cocaine supply from Colombian could account for 10%-14% of the increase in violence in Mexico and 25% of the differential increase in Mexico’s North.