Bias in U.S. Import Prices and Demand
Robert C. Feenstra, Clinton R. Shiells
NBER Working Paper No. 4841
The purpose of the paper is to measure the potential bias in the U.S. import price index due to the appearance of new product varieties, or new foreign suppliers, and determine the effect of this bias on the estimated income elasticity of import demand. Existing import price indexes are based on a sample of products from importing firms. We argue that if the share of import expenditure on the sampled products is falling over time, this will lead to an upward bias in the measured index. Using a correction based on the falling expenditure share on sampled countries, we find that the income elasticity of aggregate U.S. import demand is reduced from 2.5 to 1.7, or about halfway to unity. Our estimates suggest that the aggregate import price index is upward biased by about one and one-half percentage points annually.