Weak Cycles,Length and Magnitude of War: Duration Dependence in International Conflict

Abstract

Standard statistical tests have provided little evidence for the existence of war cycles. Cycle theorists argue that these tests are inappropriate because such cycles are inherently irregular. Duration analysis provides a means of testing for the presence of somewhat irregular weak cycles, by determining the extent to which measures of war duration cluster around a mean value, a phenomenon known as positive duration dependence or increasing hazard. Tests are performed on a 500-year data set. Turning points in long cycles of war intensity do manifest a weak cycle of 50 years, when the cycle is measured from peak to peak. The length and magnitude of individual wars exhibit no common durations and appear prone to an anti-cyclical process of negative duration dependence, whereby the probability of war termination declines with the time and magnitude duration of the war.