The sectoral shifts hypothesis, advanced by Lilien (1982) and Davis (1987), suggests that unemployment is, in part, the result of resources being reallocated from declining to expanding sectors of the economy. Using U.S. data from 1931 to 1987, we test this hypothesis by constructing an index measuring the dispersion among stock prices from different industries. We find that lagged values of this index significantly affect unemployment. We show that the stock market dispersion index is less contaminated by aggregate demand influences than Lilien's employment dispersion index, which makes our test less vulnerable to the Abraham and Katz (1986) critique. © 1990.