OPEN-ECONOMY MARKET: Nominal Income Targeting 2

In the third column of Table 1, expected next-period nominal income growth is replaced by expected current nominal income growth, EM Ax,. The improvement in fit over CGG’s specification becomes more noticeable, with the residual standard deviation about a fifth lower than that of (3.1) (0.94% vs. 1.14%). Source Compared to the results with Em Ax/+1, the 1979-1982 dummy increases further in significance, although the estimated magnitude of the output gap term declines.

In the final column of Table 1, we add EM 1 to the preceding regression, and find that it contributes no explanatory power, has a wrongly-signed coefficient, and leaves the remaining coefficient estimates virtually unaltered.

The regressions in Table 1 using nominal income growth agree with CGG’s result that, at least from 1979, the long-run response coefficient of the funds rate to the nominal aggregate exceeded unity: the estimates of (3.1) indicate a value of фр = 1.53, while the response coefficient when expected Axt+J is the nominal aggregate is фх = 1.20 (j = 0) or 1.04 (j = 1)9 All regressions in Table 1 are also uniform in implying a high degree of interest rate smoothing by the Fed, as indicated by the estimates of pR in the 0.75-0.85 range.

Our results suggest therefore that actual U.S. monetary policy since 1979 is well approximated empirically as a rule that reacts to expected nominal income growth. The success of U.S. monetary policy in recent years can then perhaps be regarded as evidence in favor of monetary policies directed towards the stabilization of nominal income growth.

We now turn to specification of the model to be used in our simulations. The aggregate demand side is an open-economy extension of the optimizing IS-LM specification used in our earlier work (McCallum and Nelson [1997, 1998]);10 our incorporation of a foreign sector draws on recent work on international versions of the Sidrauski-Brock model by Obstfeld and Rogoff (1996) and Kollmann (1996).

Tags: , ,