Replication, realism, and robustness: Analyzing political regimes and international trade

Our earlier article established that pairs of democracies trade more freely than country-pairs composed of a democracy and an autocracy (Mansfield, Milner, and Rosendorff2000). Xinyuan Dai (2002) incorrectly asserts that our conclusion depends on the preferences of the decision makers who formulate trade policy. We show that Dai fails to accurately replicate our model, and hence erroneously claims that the new equilibria she deduces are consistent with it. In addition, we demonstrate that in altering one of our assumptions, Dai offers a model that is less realistic as well as inconsistent with the substantive literature on international bargaining. Finally, we question the robustness of her approach. Due to these problems of replication, realism, and robustness, we conclude that Dai's model is of limited utility.