The WTO Trade Effect

This paper reexamines the GATT/WTO membership effect on bilateral trade flows, using nonparametric methods including pair-matching, permutation tests, and a Rosenbaum (2002) sensitivity analysis. Together, these methods provide an estimation framework that is robust to misspecification biases, allows general forms of heterogeneous treatment effects, and addresses potential hidden selection biases. This is in contrast to most conventional parametric studies on this issue. Our results suggest large GATT/WTO trade-promoting e®ects, robust to various restricted matching criteria, alternative indicators for GATT/WTO involvement, different matching methodologies, non-random incidence of positive trade flows, and inclusion of multilateral resistance terms.