Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse

A large body of scholarship finds that there is a relationship between economic dependence on oil or minerals and authoritarianism. This finding is based, however, on pooled, time-series crosssectional regressions without country fixed effects run on datasets that are longitudinally truncated. This is not an effective strategy to uncover causal associations. We therefore develop unique historical datasets, and employ time-series centric techniques, that allow us test for long-run relationships between resource reliance and regime type within countries over time. Our results indicate that increases in resource dependence are not associated with the undermining of democracy or less complete transitions from authoritarianism to democracy. We suggest that when the theory in question is not about static, cross-sectional differences between countries, but about changes that take place within countries over time, assembling and properly using historical datasets that operationalize explicitly specified counterfactuals provides a better fit between theory and evidence.