We model the war on drugs in source countries as a conflict over scarce inputs of successive levels of the production and trafficking chain. We explicitly model the vertical structure of the drug trade as being composed of several stages, and study how different policies aimed at different stages affect the supply, prices and input markets. We use the model to study Plan Colombia, a large scale intervention in Colombia aimed at reducing the supply of cocaine by targeting illicit crops and illegal armed groups’ control of the routes used to transport drugs outside of the country - two of the main inputs of the production and trafficking chain. The model fits many of the patterns found in the data and sheds light on certain puzzling findings. For a reasonable set of parameters that match well the data on the war on drugs under Plan Colombia, our model predicts that the marginal cost to the U.S. of reducing the amount of cocaine transacted in retail markets by one kilogram is $1’631.900 if resources are allocated to eradication efforts; and $267.450 per kilogram if resources are allocated to interdiction efforts.

This paper proposes a new identification strategy to estimate the causal impact of illicit drug markets on violence using a panel of Colombian municipalities covering the period 1994-2008. Using a UNODC survey of Colombian rural households involved in coca cultivation, we estimate the determinants of land suitability for coca cultivation. With these results we create a suitability index that depends on the altitude, erosion, soil aptitude, and precipitation of a municipality. Our exogenous suitability index predicts the presence of coca crops cross sectionally and its expansion between 1994-2000. We show that following an increase in the demand for Colombian cocaine, coca cultivation increases disproportionately in municipalities with a high suitability index. This provides an exogenous source of variation in the extent of coca cultivation within municipalities that we use as an instrument to uncover the causal effect of illegal cocaine markets on violence. We find that a 10% increase in the value of coca cultivation in a municipality increases homicides by about 1.25%, forced displacement by about 3%, attacks by insurgent groups by about 2%, and incidents involving the explosion of land mines by about 1%. Our evidence is consistent with the view suggesting that prohibition creates rents for suppliers in illegal markets, and these rents cause violence as different armed groups fight each other, the government and the civil population for their control and extraction.