Heterogeneous economic integration agreements' effects, gravity, and welfare

It is now widely accepted that economic integration agreements (EIAs) and other trade-policy liberalizations contribute to nations' economic growth and development and help alleviate poverty. However, the economic effects of such policies vary across countries' economic structures; importing developing countries face higher market-entry costs (partly due to poorer international networks). In this paper, we address how the trade and welfare effects of EIAs are sensitive to the levels of country-pairs' variable and fixed trade costs. It is now well established that the (variable-cost) 'trade elasticity' typically estimated using gravity equations of international trade flows is central to computing general equilibrium impacts on trade and economic welfare of trade-policy liberalizations using the new quantitative trade models. However, this trade elasticity is generally assumed to be an exogenous parameter, such as the elasticity of substitution in consumption or an index of heterogeneous productivities; moreover, most studies have ignored the role of the fixed-export-cost trade elasticity for trade-policy liberalizations (when not ignored, assumed parametric). This paper offers three potential contributions. First, we extend a standard Melitz general equilibrium trade model with firm heterogeneity to show how variable-cost and fixed -cost trade elasticities associated with trade liberalizations are heterogeneous and endogenous to levels of country-pairs' bilateral policy and non-policy, variable and fixed trade costs even allowing for CES preferences and an untruncated Pareto distribution of productivities. Using associated comparative statics, we provide explicit predictions of the heterogeneous (variable- and fixed-cost) bilateral extensive-margin, intensive-margin, and trade elasticities. Second, a modification of the state-of-the-art panel-data methodology to estimate consistent average treatment effects of economic integration agreements (that liberalize variable and fixed trade-policy costs) provides empirical support for the theoretical hypotheses. Consistent with a growing empirical literature, trade elasticities vary across particular settings; such variation is especially pronounced for North-South EIAs. Third, we demonstrate the relevance of these theoretical and empirical results for welfare calculations using the new quantitative trade models. We show empirically that 83-94 percent of the welfare (or probability) estimates of economic integration agreement liberalizations between 2,266 North-North, North-South, and South-South country-pairs can be explained by our heterogeneous economic integration agreement partial treatment effects.