Explaining Armington: what determines substitutability between home and foreign goods?

For decades trade economists have modelled imperfect substitution between home and foreign goods in consumption (often called the Armington assumption) with little analysis of what explains the wide variation in these substitution elasticities across sectors. Using a varying coefficients model, we estimate Armington elasticities between U.S. domestic and foreign goods across over 100 industrial sectors from 1980-88 and examine the role of product, industry, political, and 'home bias' factors as determinants. We find strong support that the presence of foreign-owned affiliates affects Armington elasticities in important ways, and some support that entry barriers and union presence have an effect. JEL Classification: F23, D12.