How Strong is the Love of Variety?

Models with monopolistic competition and constant elasticity of substitution (CES) preferences have become a mainstay of theoretical and empirical work in international trade. However, the standard model yields contrafactual predictions on the number of varieties, prices and output per variety that are traded. In particular the model predicts a rate of variety growth that is faster than that observed in the data. This paper develops and tests a model with a more general, but still tractable, CES preference structure that nests Krugman (1980) and Armington (1969) style models. With limited love of variety the consumer faces a trade-off between buying more varieties or higher quantities per variety and in equilibrium the model yields a variety growth rate consistent with the data. The empirics confirm that consumer’s “love of variety” is 42 percent lower than is assumed in Krugman’s model. One implication is that existing studies overstate the variety gains from trade liberalization. Another is that the impact of product variety on economic growth and the strength of industrial agglomerations is smaller than is typically assumed. * I am especially grateful to David Hummels for guidance and support. I wish to thank Jason Abrevaya, Thomas Hertel, and Chong Xiang for excellent comments. Will Martin graciously provided me access to COMTRADE data. For helpful comments and suggestions, I thank Sirsha Chatterjee, Volodymyr Lugovskyy, Georg Schaur, and seminar participants at Purdue University. All remaining errors are mine. ** Department of Economics, Krannert School of Management, Purdue University, 403 W. State St., West Lafayette, IN, 47904-2056. E-mail: [email protected].