Incidencia de la inversión pública en los indicadores de criminalidad en la ciudad de Medellín, Colombia de 2012 a 2017

Abstract

This paper studies the effects of public investment on crime rates, using the case of Medellín, Colombia from 2012 to 2017. Considering that public investment and crime have an endogeneity problem, a longitudinal panel of comunas estimated by fixed effects, is employed. Although the endogeneity problem is not eliminated, regressions are estimated under different assumptions that allow an approximation to identify causal effects. The results suggest that the concurrent investment does not have a significant relationship on crime. However, investment in security in one year has a negative and statistically significant relationship with crime in the following year. Both results are mainly explained by ordinary investment in security, and not by investment in participatory budgeting. Also, the effect by type of project is studied. The results suggest that concurrent investment does not have a significant relationship on crime, but there is a negative and statistically significant relationship of cohabitation projects with crime in the following two years.