Trade in Ideal Varieties: Theory and Evidence

Models with constant-elasticity of substitution (CES) preferences are commonly employed in the international trade literature because they provide a tractable way to handle product differentiation in general equilibrium. However this tractability comes at the cost of generating a set of counter-factual predictions regarding cross-country variation in export and import variety, output per variety, and prices. We examine whether a generalized version of Lancaster's 'ideal variety' model can better match facts. In this model, entry causes crowding in variety space, so that the marginal utility of new varieties falls as market size grows. Crowding is partially offset by income effects, as richer consumers will pay more for varieties closer matched to their ideal types. We show theoretically and confirm empirically that declining marginal utility of new varieties results in: a higher own-price elasticity of demand (and lower prices) in large countries and a lower own-price elasticity of demand (and higher prices) in rich countries. Model predictions about cross-country differences in the number and size of establishments are also empirically confirmed.