Nuclear Proliferation and the Risk of Nuclear War: Experimental Tests

Abstract:     
The causes and prevention of nuclear war are critical to human survival but difficult to study. Causal inference requires evidence of positive effects on nuclear war, but observations of nuclear war are neither available nor desirable. I circumvent the observational constraint with an experimental design that evaluates the relationship between the number of nuclear actors and the risk of nuclear conflict. I isolate the strategic aspect of the problem, using simple N-player nuclear-option games in laboratory experiments that manipulate the number of players with nuclear option, the alliance structure and the retaliation technology. Results show that decisions are largely peaceful at N = 2 despite the existence of nuclear options with a relative first-strike advantage; but a jump in the number of nuclear actors makes one more likely to choose the nuclear option. The findings highlight an avenue of research that evaluates theoretical mechanisms of nuclear war experimentally, moving a highly important topic beyond pure theory.

Number of Pages in PDF File: 34

working papers series