Selection Effects and Deterrence

The empirical question of how often deterrent threats issued during international disputes succeed has been hotly debated for years, with some researchers arguing that virtually no robust cases of success can be identified. I argue that what appears to be an empirical and methodological debate actually arises from the inadequacy of classical rational deterrence theory, which fails to comprehend the implications of states' strategic self-selection into international disputes. Rational self-selection is shown to imply that in a sample of crises, deterrent threats issued after an initial challenge will tend to fail in precisely those cases where they are relatively most credible signals of an intent to resist with force. The product of a selection effect, this paradoxical implication allows a resolution of the debate on the efficacy of deterrence in crises. And because selection effects can arise whenever a historical "case" is the product of choices by actors who also influence the outcome in question, this example from the study of deterrence has broad relevance for empirical research.