[This paper introduces a new approach to the empirical testing of the Lucas-Sargent-Wallace (LSW) "policy ineffectiveness proposition," which compares the LSW hypothesis with an alternative that states that prices respond fully in the long run, but only gradually in the short run, to nominal aggregate demand disturbances. The empirical equations, estimated for a new set of quarterly data extending back to 1890, exhibit uniformly high responses of real output and low responses of price changes to anticipated changes in nominal GNP. The paper compares and tests three alternative methods of introducing "persistence effects" into the LSW framework.]