Edward Wiles

MIT · Economics · 2026–27 job market · Postdoc

Fields

International, Development

Job market paper

Aggregate Consequences of Relationship Dynamics in Trade

Most trade takes place within long-term buyer-seller relationships whose value reflects past experience and, when contract enforcement is limited, the expectation of future business. We study how these relationship dynamics shape aggregate adjustment to changes in trade costs. We develop a dynamic multi-location, multi-sector Ricardian model with relationship-specific learning and relational contracts. Accumulated relationship value slows adjustment as some firms retain incumbent suppliers despite cheaper alternatives, while improved outside options can offset this by undermining the incentives needed to sustain those relationships. A property of the optimal contract that we call scale equivariance permits tractable aggregation of these micro-level dynamics. Using rich firm-to-firm transaction data from India, we establish that proportional growth within relationships is remarkably similar across firm characteristics and experience---supportive of the scale equivariance property---and varies with contract enforcement. We use these empirical results as moments to estimate the model's learning and contracting parameters, and study a reduction in interstate trade costs calibrated to a major Indian domestic integration reform. Both dynamic mechanisms are quantitatively important: it takes 16 quarters for the economy to reach 90% of its long-run trade elasticity, compared with 26 quarters when relationships can continue even if their contracts violate the post-shock incentive constraint.