How big is the elasticity of substitution between goods from different countries—the Armington elasticity? Estimates of the macroelasticity between home and imported goods are often smaller than the microelasticity between foreign sources of imports. Using new, highly disaggregate U.S. production data matched to imports and simulated data from a Melitz-style model with nested CES preferences, we explore estimation techniques for the two elasticities. For between two-thirds and three-quarters of sample goods, there is no significant difference between the macro- and microelasticities, but for the rest, the microelasticity is significantly higher, even at the same level of disaggregation.