Have Regional Trade Agreements in Developing Countries Been a Success or a Failure? Evidence from the COMESA Free Trade Agreement and MERCOSUR

Abstract

This paper models the trade creating and trade diverting effects of the COMESA free trade agreement (FTA) and MERCOSUR over time using several theoretically consistent gravity equations. The extensive sector-specific TradeProd panel data set from 1980 to 2006 is employed throughout. From the most preferred specification, both agreements are found to have large trade creating effects relative to estimates of other trade agreements in the literature. Six years after their respective implementations, intra-COMESA FTA trade and intra-MERCOSUR trade are shown to have grown to 171% and 116% above their gravity equation predicted levels. There is some evidence that MERCOSUR diverted modest levels of trade from non-members but there is no evidence of trade diversion as a result of the COMESA FTA. A key finding generally is the gradual effects of trade agreements over time which has important consequences for the appropriate modelling specification. Trade created in the COMESA FTA is shown to have occurred largely in sectors in which countries have varying comparative advantages, indicating that new links are efficient. Welfare changes in accordance with Arkolakis et al. (2012) are also calculated to show both the heterogeneity of gains across countries and the somewhat limited impact of the COMESA FTA due to low natural levels of absolute trade among members. Import penetration ratios therefore saw only small changes. Robustness is provided by Pseudo Poisson Maximum Likelihood estimation.