Files
Abstract
This paper addresses a key gap in the spatial analysis of sub-national conflict: the under utilization of spatial heterogeneity methods. Using data on the Kurdish conflict from 1980 to 2015, I demonstrate how standard Ordinary Least Squares (OLS) and spatial regression techniques often fail to uncover meaningful relationships due to unaccounted for spatial nonstationarity. By applying spatially constrained endogenous regime models, I identify distinct regions where conflict is significantly shaped by the presence of oil—the primary predictor in this case study. Further analysis reveals a significant spatial lag effect within one of these regions. These findings highlight the critical importance of incorporating spatial heterogeneity into conflict modeling and offer a practical methodological framework for doing so in future research.