Exchange Rate Pass-Through in Turkish E-Commerce Marketplace
Abstract
This paper examines exchange rate pass-through (ERPT) in a large Turkish e-commerce marketplace using high-frequency weekly micro price data. The empirical analysis relies on a balanced panel of 13,433 products continuously observed over a 26-week period following a substantial currency depreciation episode. We estimate a dynamic single-equation panel model in first differences of logarithmic variables using a recursive expanding-window framework, allowing ERPT coefficients to evolve over time. The results indicate that short-run pass-through is consistently negative and statistically significant across recursive windows, suggesting the presence of price smoothing behavior and nominal rigidities even in a digital retail environment characterized by minimal menu costs. In contrast, long-run pass-through exhibits marked state dependence, rising to approximately 30-35 percent during periods of sustained exchange rate pressure and declining during intensive discount weeks. After implementing product-level corrections to account for temporary promotional distortions, cumulative pass-through becomes more stable but remains incomplete. Furthermore, percentile-based heterogeneity analysis reveals a monotonic increase in long-run pass-through from low to high-priced product segments, indicating substantial variation across pricing tiers. Overall, the findings demonstrate that exchange rate transmission in Türkiye’s e-commerce market is gradual, incomplete, heterogeneous, strongly state-dependent, and declines over time following exchange rate shocks, while also varying across different price groups, despite the high degree of digital price flexibility.