Abstract:
Over the past 30 years, "deep trade agreements" (DTAs) have proliferated. DTAs go beyond traditional preferential trade agreements focused historically on reducing tariff rates on goods trade, influencing the behavior of exporting firms and multinational enterprises (MNEs) via a broad swath of "provisions." However, estimation of individual effects of several hundred provisions in DTAs on trade using the World Bank's DTA database remains in its infancy. While this database categorizes substantive provisions between liberalizations vs. obligations, this paper is the first to use the Shapley Value approach from cooperative game theory to generate unbiased estimates of the signs of all individual provisions' partial effects. First, we find ex post evidence that (the World Bank's) "liberalizations" can have positive or negative effects on trade and "obligations" can have negative or positive effects. Our approach generates precise estimates of the positive-versus-negative effects of narrow sets of provisions, while addressing challenges posed by omitted-variables, over-aggregation, and multi-collinearity biases. Second, in contrast to most studies that focus exclusively on trade effects of DTAs, we introduce a new data set on MNEs to inform us of DTA provisions' effects on MNEs' bilateral FDI, costs, employment, revenues, and assets – alongside provisions' effects on trade – which are of importance to MNEs' managements and governments' policymakers. Third, we find convincing evidence that provisions that positively (negatively) affect trade flows also negatively (positively) affect FDI flows, suggesting that trade and FDI are predominantly substitutes with respect to DTAs' provisions. Finally, we provide computable general equilibrium welfare estimates associated with several policy counterfactuals.