The Effects of Drug Enforcement on Violence in Colombia 1999-2010: A Spatial Econometric Approach

In this paper paper, I use Mejia and Restrepo's(2011d} strategy to disentangle the causal relationship between drug enforcement and violence. To test this relationship, I use information on Colombian municipalities during the period 1999-2010. Due to technological reasons related to the quality of terrain, climate, and locational characteristics of the Colombian territory, cocaine production is more productive at low altitudes. Using the altitude of each municipality and distance from capital cities as sources of exogenous variation, I estimate the effect of drug enforcement on violence in Colombia. To control for a possible omitted-variable bias in the estimations, I run a Panel Data Spatial Durbin Model (SDM). Additionally, I construct a set of indices with comparable units of measure which allows me to determine which percentage of the Colombian violence data is explained by drug enforcement. The results indicate that the Colombian government's enforcement activities increased in 0.98% the homicide rate and in 1.24% the displacement rate and the war among drug dealers increased in 4.00% the homicide rate and 0.16% the displacement rate in the period 1999-2010.