chapter  13
16 Pages

A subnational study of insurgency: FARC violence in the 1990s: Jennifer S. Holmes, Sheila Amin Gutiérrez de Piñeres and Kevin M. Curtin

ByJENNIFER S. HOLMES, SHEILA AMIN GUTIÉRREZ DE PIÑERES AND KEVIN M. CURTIN

Many factors are theorized to be important to explain insurgency, including geography and history, the economy, government, and demography. Although Colombia is one of Latin America’s oldest democracies, with a history of unusually consistent economic growth for the region, it is also home to one of the most entrenched leftist insurgencies in the world. Colombia is an ideal case to test these factors, due to the ability to analyze the issues at the subnational level and the variability of violence within the country. Although most of the research on Colombia is conducted at the national level, a subnational study of Colombia offers a unique opportunity to incorporate many of the theoretically important factors into a model of FARC violence. In this article, the analysis at the subnational level is based on unique dataset constructed from CINEP (Centro de Investigación y Educación Popular) and Colombian government sources. The subnational-level analysis can assist in the determination of whether the cross-national or national level results are supported at a disaggregated level. After extensive specifi cation testing, the zero-infl ated negative binomial regression was selected to model guerilla violence. In this analysis, a zero-infl ated negative binomial regression model is utilized because the dependent variable (FARC human rights violations) is a count, with excess zeros and overdispersion. This model allows for different factors to account for the absence of violence (zeros) and the presence of violence (non-zeros). In this case, the forces that contribute to an absence of guerilla violence may be different from forces that explain the intensity of guerilla violence. Additionally, the model allows different processes to determine the absence of violence (zero counts). Based on the theories of guerilla violence, there could be multiple reasons why there is an absence of violence in a department-no exports to loot, high level of GDP or development, strong state presence, or other, unspecifi ed factors.