What matters for agricultural trade? Assessing the role of trade deal provisions using machine learning

Abstract

This paper employs machine learning to determine which preferential trade agreement (PTA) provisions are relevant to agricultural trade patterns and the factors that may influence their adoption. Utilizing the three-way gravity model, we apply plug-in Lasso regularized regression to pinpoint predictive PTA provisions for agricultural trade. Our findings underscore the importance of competition policies, export taxes, intellectual property rights, capital movement, state enterprises, and technical trade barriers. Subsequently, we use Random Forests to reveal the economic, political, social, and geographic factors associated with the inclusion of those provisions in PTAs. The findings highlight the roles of contagion, governance quality, energy use, and geographic proximity. Our analysis provides new insights that can aid in formulating strategies to support agricultural trade.