The arsenal of democracy: Production and politics during WWII

Abstract

We study the geographic distribution of military supply contracts during World War II. This is a unique case, where $3 trillion current day dollars was spent. We find robust evidence consistent with the hypothesis that economic factors dominated the allocation of supply contracts, and that political factors—or at least winning the 1944 presidential election—were at best of secondary importance. General industrial capacity in 1939, as well as specialized industrial capacity for aircraft production, are strong predictors of contract spending across states. Electoral college pivot probabilities are weak predictors of contract (and new facilities) spending, and under the most plausible assumptions they are essentially unrelated to spending. This is true over the entire period 1940–1944, and also for shorter periods leading up to the election in November 1944. That is, we find no evidence of an electoral cycle in the distribution of funds.