We empirically examine the heterogeneity in the effects of multiple dimensions of distance on trade across detailed product groups. Using finite mixture modeling on bilateral trade data at the 3-digit SITC level, we endogenously group product categories into an, a priori unknown, number of segments based on estimated coefficients of multiple dimensions of distance in the gravity equation. We find that institutional distance, whether countries belong to the same trade block and especially geographical distance are crucial and distinct factors to classify commodities in homogeneous groups. Classification-JEL: F14, F21, F23 Keywords: bilateral trade, gravity models, distance, institutions, product heterogeneity, finite mixture modeling File-Url: https://papers.tinbergen.nl/12065.pdf File-Format: application/pdf File-Size: 394753 bytes Handle: RePEc:tin:wpaper:20120065