Hufbauer, Schott and Elliott have written probably the most comprehensive empirical study of economic sanctions in their volume Economic Sanctions Reconsidered: History and Current Policy. They conclude their analysis with several policy recommendations. However, there are problems with their empirical analysis that significantly affect these recommendations. To overcome these problems, I reanalyze their data using ordered logit estimation. In addition to modeling their policy recommendations, I add three hypotheses derived from the sanction literature. The results show that most of the relationships between the variables in their recommendations are insignificant, calling their accuracy and importance into question. In two cases, the original recommendations are accurate only when conditioned by other variables. Cooperation only has a negative effect on success when international organizations are not involved, and nations trying to subvert the sanctions only succeed when the target was originally dependent on the sender for its imports. I conclude by discussing the policy implications these findings have for the future use of economic sanctions.
Replication datasets available