Brookings Papers on Economic Activity, 1: 1988 the models quantitative implications by calculating the real effects of shocks for a range of plausible parameter values. The results suggest that the effects of average inflation and demand variability are large Next we argue that the implications of our model are robust: they carry over to broad classes of other Keynesian models Finally, we compare the predictions of Keynesian theories with thos of models in the new classical or equilibrium tradition, focusing on Lucas's model of imperfect information. Like our model, Lucas's predicts that the size of the real effects of shocks depends negatively of the variance of aggregate demand. Since this prediction is common to Keynesian and new classical theories, testing it empirically, as lucas and others have done, is not useful for distinguishing between the two theories. Crucially, luca del differs from ours by predicting that the effects of shocks do not depend on average inflation. This difference leads to the tests of the models in the next section THE MODEL AND QUALITATIVE RESULTS Our model of price adjustment is similar in spirit to those of John Taylor and Olivier Blanchard. 25 The model is set in continuous time The economy contains imperfectly competitive firms that change prices at discre ete I costly. Price setting is staggered, with an equal proportion of firms changing prices at every instant. The crucial departure from Taylor and Blanchard is that the length of time between price changes, and hence the rate at which the price level adjusts to shocks, is endogenous. Th we can study the determinants of the speed of adjustment Consider the behavior of a representative firm, firm i. Rather than derive a profit function from specific cost and demand functions, we simply assume that firm is profits depend on three variables: aggregate spending in the economy, y; firm is relative price, Pi-p; and a firm specific shock, 0, (all variables are in logs). The aggregate price level p is defined simply as the average of prices across firms. Aggregate spending y affects firm is profits by shifting the demand curve that it faces. When aggregate spending rises, the firm sells more at a given 25. Taylor, ""Staggered Wage Setting"and"Aggregate Dynamics and Staggere Contracts"; and Blanchard, " "Price Asynchronization and "Wage Price Spiral