Terrorism Determinants, Model Uncertainty and Space in Colombia.

This paper studies the determinants of terrorism at the sub-national level in Colombia during 2001–2014. In order to establish robust relationships, a Bayesian model averaging framework has been implemented using departmental data. We find that the violence suffered by this country is linked to economic factors, especially labor market outcomes. The results obtained are not significantly altered by the use of relative measures of terror, the specification of alternative parameters and model priors or the presence of spatial dependence. The main conclusion drawn from our analysis is that an appropriate strategy to fight against terrorism in similar contexts is to increase its opportunity cost. This might be achieved through the promotion of inclusive socioeconomic development, primarily in rural areas. [ABSTRACT FROM AUTHOR]