Trade Integration And Trade Agreements:Resolving The Endogeneity Problem Through A Qualitative Var

While the endogeneity of trade and regional integration agreements was established early on, this issue has only been addressed explicitly in gravity models during the last decade and a half. Initial attempts using instrumental variables proved unreliable, causing authors to look for alternative solutions. This paper brings together the literature on both gravity equations explaining trade and probit regressions explaining the probability of an integration agreement. This is done by estimating them simultaneously in a qualitative vector autoregression model. The qualitative VAR allows us to estimate their interdependence without having to resort to instrumental variables. In addition, the endogenous nature of other control variables like the GDP or the capital labor ratio can be taken into account. Our preliminary findings confirm that an increase in trade raises the probability of an agreement and vice versa, although the response can di er over specific continents. We find a relatively small average treatment effect of RIAs: trade increases with 10% after one year and 40% after five years whereafter it slowly rises to 80% after 35 years.