FDR, U.S. Entry into World War II, and Selection Effects Theory

To the Editors (Dan Reiter writes)

In “The Deception Dividend: FDR’s Undeclared War,” John Schuessler argues that in 1941 President Franklin Roosevelt took actions against Germany and Japan to increase the likelihood that the United States would enter World War II, despite public hesitance to do so.¹ Schuessler writes, “The ‘undeclared war’ in the Atlantic and the oil embargo on Japan should be understood as designed, at least in part, to invite an ‘incident’ that could be used to justify hostilities” (p. 145). The broader implication is that if elected leaders can manipulate their countries into war despite public hesitation, then public opinion does not affect democratic foreign policy choices. More specifically, the ability of elected leaders to circumvent public opinion would damage the selection effects explanation of why democracies win their wars. The selection effects argument, which Allan Stam and I have made, claims that democracies win the wars they initiate because public opinion constrains them to initiate only those wars they are highly confident they will go on to win, and the marketplace of ideas better informs their policy choices.²

In this letter, I make three main points. First, even if one concedes the entirety of Schuessler’s historical interpretation, the United States in World War II is not a case of a democracy initiating a war it went on to lose, and therefore not evidence against the proposition that democracies are likely to win the wars they initiate. Second, the evidence does not clearly indicate that Roosevelt sought to provoke war, and there is evidence to the contrary. Third, it is hard to think of a better demonstration of selection effects theory than Roosevelt’s actions in 1941. He was deeply aware of public opinion and recognized that major foreign policy actions must be popular. Despite his grave concern about the imperative of U.S. aid to Britain, Roosevelt was careful not to get ahead of public opinion, taking only those actions the public supported and no more. He worked within the marketplace of ideas to influence public opinion through persuasion and public speeches. His public claims about the seriousness of the Nazi and Japanese threat and the importance that the Soviet Union and Britain not lose the war reflected his genuine beliefs, and in hindsight appear to be right. The marketplace of ideas corrected his largest deviation from the facts, his biased portrayal of the Greer incident.

Schuessler’s empirical argument does not provide evidence against selection effects theory. His argument would damage the theory if Roosevelt duped the public into supporting the initiation of a war that the United States went on to lose, but the United States defeated the Axis. And if one holds that war with Japan and Germany was avoidable and occurred because of U.S. provocation, then this is a case of a democracy choosing war and winning, the prediction of selection effects theory.

Some scholars might claim that the very attempt to deceive is evidence against selection effects theory. The theoretical assumption that elected leaders wish to manipulate public opinion to give them more foreign policy latitude is, however, not inconsistent with selection effects theory. Elected leaders pay attention to public opinion not because they share a normative commitment to democratic processes, but because they worry about the electoral consequences of unpopular policies. Stam and I argued that elected leaders engage in actions inconsistent with normative visions of democratic foreign policymaking, such as covertly undermining elected governments.³ Credibly blaming the other side for initiation of the war certainly affects public opinion. I noted fifteen years ago in this journal that one reason preemptive wars are so rare is that leaders understand the political advantages of being attacked rather than of attacking.⁴

The historical record indicates that public opinion is not endogenous to the machinations of elected leaders. The ability of elected leaders to form public opinion is highly limited, and the constraint of public opinion dissuades them from initiating wars they will lose. The quantitative evidence demonstrating different implications of the assumption...