Predispositions and foreign policy surprises: Assessing the impact of rational and biased beliefs on strategic decision-making

This project analyzes the relationship between foreign-policy decision-making, surprise and misjudgment using a multi-method approach. In recent years, high profile foreign policy and intelligence failures have captured the attention of policymakers, the media and academics alike. Many retrospective studies of policy failure identify, implicitly or explicitly, psychological biases of pertinent decision-makers as contributing to poor outcomes. By focusing on the internal mechanics of information usage and decision processes, these studies draw from bureaucratic politics theory. An alternative paradigm used to explain state behavior, rational actor theory, focuses on the competition between states and often assumes that the bureaucratic process works perfect. Since each explanation stresses different aspects of the decision-making process, it is difficult to know whether conclusions derived with one paradigm are transportable to the other. The bureaucratic politics explanation often stresses the negative relationship between psychological biases and outcomes, whereas rational actor theory illustrated the benefits of feigning irrationality. This thesis improves our understanding of psychological biases in competitive interstate interactions by incorporating the possibility of misjudgment into a formal model of surprise attack. The project compares an efficient form of learning, Bayesian updating, to an inefficient learning mechanism whereby actors irrationally overweigh their prior beliefs relative to new information. This project then identifies conditions where irrationally entrenched beliefs may be overemphasized as a reason for policy failure, and situations where irrational over-suspicion can offer competitive advantage by altering another player's behavior. The deductive analysis suggests that psychological biases may be over-used as an explanatory factor in retrospective studies of failure, that psychological bias is neither a necessary or sufficient condition for surprise, and that irrational over-suspicion can have a stabilizing impact on a hostile interaction. These deductive propositions are examined quantitatively using a dataset of international crises from 1918-2002, which suggests that irrationally entrenched beliefs may reduce the incentive to attempt surprise attack. This is followed by a qualitative analysis of the 1962 Cuban Missile Crisis and 1967 Arab-Israeli War, showing that the model with a potentially biased actor seems to outperform the model with perfectly rational actors.