Abstract: In 1957, Robert Solow published a paper that provided the theoretical foundation for almost all subsequent work on productivity measurement. Although most applications of Solow’s method have measured trends over fairly long time periods, the method also has important uses at higher frequencies. Under constant returns to scale and competition, the Solow residual measures the pure shift of the production function. Shifts in product demand and factor supplies should have no effect on the residual. Tests of this invariance property show that it fails in a great many industries. Though other explanations may deserve some weight, it appears that the leading cause of the failure of invariance is increasing returns and market power. The empirical findings give some support to the theory of monopolistic competition.