"This paper re-examines aggregate and disaggregate import and export demand functions for the United States. This re-examination is warranted because (1) income elasticities are too high to be warranted by standard theories, and (2) remain high even when it is assumed that supply factors are important. These findings suggest that the standard models omit important factors. An empirical investigation indicates that the rising importance of vertical specialization combined with decreasing tariffs rates explains some of results. Accounting for these factors yields more plausible estimates of income elasticities, as well as smaller prediction errors"--National Bureau of Economic Research web site.