AI-Driven Trade Analytics for Evaluating Free Trade Agreement Outcomes and Industrial Competitiveness

Abstract: This study investigates how AI-driven trade analytics can be integrated with econometric identification to evaluate Free Trade Agreement (FTA) outcomes and to quantify changes in industrial competitiveness using trade microdata. The empirical design is structured around a policy evaluation workflow that combines (i) event-study and difference-in-differences estimators for causal attribution and (ii) explainable machine-learning models for forecasting and heterogeneity discovery across products and partners. The analysis focuses on India’s goods trade with ASEAN as an empirically salient case because India–ASEAN merchandise trade is large, policy-relevant, and characterized by persistent bilateral imbalances that allow competitiveness mechanisms to be tested. To ensure transparency and reproducibility, the paper documents real-data collection pathways from official repositories (Ministry of Commerce and Industry, Government of India; World Bank WITS/UN Comtrade) and uses those data to build results tables and figures embedded within the Results section. The results indicate a sharp post-pandemic expansion in India–ASEAN total trade alongside a widening trade deficit, suggesting that liberalization and supply-chain integration may have increased import penetration faster than export expansion in several partner corridors. The findings motivate a policy-facing interpretation: AI-driven analytics can sharpen diagnostic power by decomposing aggregate effects into partner–product contributions and by identifying the features most associated with competitiveness outcomes. Keywords: Free Trade Agreements; trade microdata; industrial competitiveness; explainable AI; difference-in-differences; event study; India–ASEAN trade.