Elasticities for U.S. Imports

Abstract

According to traditional thinking, it takes an estimate to beat an estimate—a view that aptly describes three decades of econometric analyses of U.S. imports. The estimate to beat is that of Houthakker and Magee (1969) who, modeling U.S. imports in terms of income and relative prices, reported an estimate of 1.5 for the income elasticity (see references on p. 5). Interest in beating this estimate stems from its puzzling implications. First, from a statistical standpoint, the elasticity estimate exceeds 1, which is puzzling given that it is assumed to be constant. Second, from an economic standpoint, the GDP share of imports will exceed 1 in the absence of price increases, which is puzzling because the GDP shares of consumption and investment are constant in comparison. Indeed, appendix 3.7.2 shows that, with prices fixed, U.S. imports will equal U.S. GDP between 60 years and 100 years from now depending on the estimated income elasticity.