Revisiting the Effects of Regional Trade Agreements Using Staggered Adoption Designs

Studies examining the effect of regional trade agreements (RTAs) on trade find rather heterogeneous effects depending on the terms of the agreements, sectors and products covered, and the characteristics of the countries involved. The standard approach to estimating RTA effects does not explicitly account for variations in the initial periods the RTAs enter into force and the number of periods these RTAs have been effective. Such variations have important implications on the choice of proper estimation methods as well as the magnitude and interpretation of the RTA effect. This paper builds on recent advances in the difference-in-differences (DD) estimation literature to explicitly account for time-varying and multiple periods of RTAs and quantify their impact on trade. Using bilateral trade flows between 1989 and 2019, we adopt a staggered adoption design to allow for heterogeneous timing of RTA effects. Our estimates are larger than previous studies because the failure to account for dynamic treatments biases the estimates toward zero. Furthermore, we find evidence of anticipation and phasing-in effects