Dynamic Properties. We now turn to the basic dynamic properties of output and inflation—as measured by the unconditional autocorrelations~in the four models. Because output and inflation dynamics are sensitive to the specification of monetary policy, we begin by specifying a baseline policy rule used in each model for comparison purposes. For this purpose, we use the following interest rate reaction function, which was estimated using quarterly U.S. data over the sample period 1980Q1 – 1996Q4:

where rt is the federal funds rate, n(4)t is the four-quarter moving average of the inflation rate, andjt is the current output gap. Clarida, Gali and Gertler (1997a) and others argue that this period was characterized by a fairly stable policy regime that differed substantially from that of the 1960s and 1970s. This estimated policy rule features a relatively large coefficient on the lagged funds rate and a fairly aggressive response to increases in inflation and output gaps. Furthermore, the pattern of coefficients on the output gap and its lag suggest that policy not only responded to the level of output but also its recent growth. read only

We use the Anderson and Moore (1985) implementation of the Blanchard and Kahn (1980) method to solve for the VAR representation of the stable solution of the models, and we compute the unconditional moments of model variables analytically. The inflation autocorrelogram of each model under the estimated policy rule is depicted in the upper panel of Figure l. Not surprisingly, the FM and MSR models—which share the Fuhrer-Moore contracting specification known to deliver a high degree of inflation persistence—are characterized by highly persistent inflation. Inflation is somewhat less persistent in the FRB model. As seen clearly in the figure, TAYMCM displays the least degree of inflation persistence, with autocorrelations that fall below zero after only four quarters. Even when combined with some inertia in price markups, the staggered nominal wage contract specification in TAYMCM delivers relatively low inflation persistence.

Dynamic Properties. We now turn to the basic dynamic properties of output and inflation—as measured by the unconditional autocorrelations~in the four models. Because output and inflation dynamics are sensitive to the specification of monetary policy, we begin by specifying a baseline policy rule used in each model for comparison purposes. For this purpose, we use the following interest rate reaction function, which was estimated using quarterly U.S. data over the sample period 1980Q1 – 1996Q4: where rt is the federal funds rate, n(4)t is the four-quarter moving average of the inflation rate, andjt is the current output gap. Clarida, Gali and Gertler (1997a) and others argue that this period was characterized by a fairly stable policy regime that differed substantially