Power System Membership and Patterns of War

Abstract

This study presents national-level findings pertaining to the basic stochastic patterns of war duration and severity, and determines the effects of power system membership on those patterns. The results have theoretical and empirical relevance to general contagion hypotheses of war and specific import to the negative addiction (or war-weariness) thesis. Among the principal findings are indications that irrespective of the characteristics of past or current war behavior, in the long run: (1) a nation that fights a war has more than a one-in-three probability of fighting for over two years and sustaining over 15,000 battle fatalities; (2) major powers are more likely to fight severe wars and less likely to fight moderate wars than are minor powers, whereas they are equally likely to fight small wars; (3) major powers and minor powers have roughly equiprobable chances of fighting wars at short, moderate, and long duration levels. It is concluded that power distinctions among nations have significant effects on long-term probability patterns of war severity but not on war duration. Regarding war-to-war linkages, the probable duration or severity of a nation's next war is unaffected by the duration or severity of its last war. In other words, there is no evidence indicating either a systematic increase or decrease in war-to-war cost-tolerance levels. Therefore, decisions on allocations of resources in current or future wars appear to be unaffected by the past: a nation's aggregate capability (i.e. power status)—not its antecedent experience—is a determinant of the scale of its wars.