Testing Power-Transition Theory Using Alternative Measures of National Capabilities

There has been a long debate over whether peace is best preserved by a balance or preponderance of power. Organski and Kugler suggested that the dynamics of relative dyadic power matter most. Using GNP to measure national capabilities, they found support for their power-transition theory, but only for states considered to be contenders for dominance in the international system. Subsequently, Houweling and Siccama reported important new evidence in support of the theory. They concluded that power transitions are a potent predictor of war for all major powers, not just a small subset; but it has been unclear whether their stronger results were a consequence of genuine improvements in methods, the use of a different measure of power, or alterations to the list of major powers. The authors replicate Houweling and Siccama's analysis using the two most common, and recently revised, measures of national capabilities—the Correlates of War composite index and GDP—and investigate the effects of modifying the set of major powers. They find substantial support for the power-transition theory, but the strength of the evidence depends importantly on how power is measured and the set of cases analyzed.