| Summary/Abstract |
Audience cost theory posits that concern over the nation’s reputation pushes voters to sanction leaders who make empty threats because they tarnish the nation’s honor. We question the empirical support for that theory. We show that survey vignettes in the previous experimental literature conflate audience costs generated by inconsistency and belligerence with approval losses arising from the perception that the leader is incompetent. These ‘incompetence costs’ are due to leaders not achieving audiences’ preferred outcomes. Our article contributes to the literature on audience costs by disentangling inconsistency and belligerence costs from incompetence costs, which we find are the larger component of audience costs. We also make a methodological contribution: we show that experimental designs in previous studies cannot test the different mechanisms; that previous estimates of audience costs are biased because treatments affect respondents’ beliefs about the likely outcome of policy actions; and we suggest a new experimental framework to estimate audience costs. Our results are consistent with arguments that audiences care more about policy outcomes than about leaders’ inconsistency or belligerence during a crisis.
|