Do Deep Trade Agreements’ Provisions Actually Increase – or Decrease – Trade and/or FDI?

Over the past 30 years, deep trade agreements (DTAs) have expanded, influencing trade and multinational enterprises (MNEs) through numerous provisions. However, estimating the individual effects of these provisions remains underdeveloped. Using the World Bank’s DTA database, this paper applies the Shapley Value approach from cooperative game theory to generate unbiased estimates of the signs of all individual provisions’ partial effects. We find that DTAs’ provisions can have both positive and negative effects on trade. Our method addresses biases from omitted variables, over- aggregation, and multicollinearity. Beyond trade, we introduce a new dataset on MNEs (the Multilateral Revenue, Employment, and Investment Database, or MREID) examining provisions’ effects on bilateral FDI, costs, employment, revenues, and assets. We provide strong evidence that trade-boosting provisions tend to reduce FDI, and vice versa, indicating a substitution relationship with respect to DTAs’ provisions. Moreover, we find novel evidence using MREID that FDI-enhancing provisions are associated with lower costs per employee. Finally, we present computable general equilibrium welfare estimates for various policy counterfactuals.