Citation

Basedau M, Lay J. J. Peace Res. 2009; 46(6): 757-776.

Copyright

(Copyright © 2009, SAGE Publishing)

DOI

10.1177/0022343309340500

PMID

unavailable

Abstract

The ‘resource curse’ hypothesis claims that abundance in natural resources, particularly oil, encourages especially civil war. Natural resources provide both motive and opportunity for conflict and create indirect institutional and economic causes of instability. Contrarily, the theory of the rentier state — largely neglected in the study of peace and war in this respect — suggests that regimes use revenue from abundant resources to buy off peace through patronage, large-scale distributive policies and effective repression. Consequently, such rentier states would tend to be more stable politically and less prone to conflict. These two theories thus imply ambivalent effects of resource abundance on conflict proneness. This article presents part of a solution to this apparent puzzle for the case of oil-producing countries. The key argument is that both resource wealth per capita and resource dependence need to be taken into account, since only the availability of very high per capita revenues from oil allows governments to achieve internal stability. The empirical analysis supports this hypothesis. The findings of multivariate cross-country regressions indicate a U-shaped relationship between oil dependence and civil war onset, while high resource wealth per capita tends to be associated with less violence. The results of a macro-qualitative comparison for a reduced sample of highly dependent oil exporters are even more clearcut. Using the same reduced sample, we find that oil-wealthy countries apparently manage to maintain political stability by a combination of large-scale distribution, high spending on the security apparatus and protection by outsiders. Compared to oil-poor countries and in contradiction to the rentier state theory, the institutions of oil-wealthy countries do not seem to be particularly characterized by patronage and clientelism.