Thirty Years of the Audience Cost Theory: Correlations, Mechanisms, and Scientific Realism

This article offers a comprehensive review of Audience Cost Theory (ACT) over the past 30 years, analyzing it through the lens of scientific realism. Originally proposed by James Fearon (1994), ACT posits that leaders face domestic political costs if they make threats in international crises and subsequently back down. The theory has sparked significant debate, with empirical support being divided. While survey experiments generally confirm public disapproval of empty threats, historical case studies and large-N quantitative analyses provide mixed or limited evidence. This article identifies three key research agendas: the existence of audience costs (theoretical entity), how leaders act with audience cost awareness (theoretical mechanism), and whether ACT’s predictions reflect real-world outcomes (logical consequences). This article argues that disagreement over ACT has been hitherto the result of each researcher’s different analysis of the three agendas. From a scientific realism perspective, ACT’s core entity—audience costs—may exist but often remains unobservable in real-world data owing to the idealization of experimental scenarios and selection effects. The study concludes that future research must link audience cost entities with mechanisms and outcomes, enhance data reproducibility, and conduct meta-analyses. Oral histories and additional survey experiments conducted under various political conditions are recommended to enhance the theory’s empirical foundation and practical significance in international relations.