Promises or Policies? An Experimental Analysis of International Agreements and Audience Reactions

Abstract A key assumption of audience costs theories of crisis bargaining and international cooperation is that audience members have strong preferences for consistency between their leader's commitments and actual policy choices. However, audiences also have strong preferences over the policy choices themselves, regardless of their consistency with past commitments. I conducted a randomized survey experiment to evaluate the magnitude of consistency and policy effects in the context of international agreements over trade policy. Respondents with expressed policy preferences, whether supporting or opposing free trade, have muted reactions to learning that their leader has broken an agreement. Only respondents with no opinion on trade policy are affected by learning that their leader's policy is inconsistent with prior commitments. This suggests that constituents' underlying preferences limit the degree to which audience costs influence policymakers' calculations.