The Decline of War Question: A Comparative Regional Perspective

Abstract The decline of war question has generated sustained debate, with global analyses reaching conflicting conclusions. We argue that framing matters: a comparative regional perspective finds substantial evidence of decline in some parts of the world even as bellicosity has held steady or increased elsewhere. Using Lyall’s Project Mars data on conventional war from 1800 to 2011, we test for changes in participation rates across eight world regions at a 1950 break point, using methods that account for overdispersion in annual counts. Participations are assigned to the belligerent’s home region rather than to the location of the fighting, so we measure who has been fighting rather than where. Europe, Latin America and East and Southeast Asia decline significantly; the Middle East and North Africa increases; four regions return nulls. The global decline is attributable to European belligerents: excluding them, it is not significant in any specification. This is not the familiar observation that Europe has been peaceful since 1945: European states have also cut their participation in conventional war outside Europe by roughly 80 percent, a decline more securely established than the near-disappearance of fighting on European soil. Two limits qualify these findings. Outside Europe no region has the statistical power to detect a 30 percent change, so our nulls are undetermined rather than evidence of stability. And three regional verdicts turn on a disputed coding question: whether a group of civil wars omitted from Project Mars, which critics say belong in it, count as conventional warfare.