In this paper I explore the potential link between Plan Colombia and violence with a new perspective. I focus the analysis only on the first three running years of the program (2000-2002) in order to avoid the overlapping effect with a security policy started in 2002/2003. This paper exploits the differential in the success of the program among the different regions to identify the potential side effects on homicides and violent deaths. Results show that, although consistently negative estimates, no-significant effects are observed on homicides. On the other hand, I found evidence of increases in the number of violent deaths for women living in urban areas, and an opposite negative effect for men living in rural areas. These findings are fully consistent across different specifications of the model, the cut-off end of the program, and the classification of the regions’ criteria.