OPEN-ECONOMY MARKET: Existing Arguments for NIT 4

The upshot of the foregoing discussion is that arguments pro and con regarding NIT rules relative to ones with inflation and output gap target variables depend upon details of the dynamic relationships between nominal and real variables, about which prevailing theory is not particularly helpful. In Sections 4-8 of this paper, consequently, we shall attempt to explore some quantitative aspects of the issues by simulations conducted in a model that has been carefully specified to respect both neoclassical theory and also the quantitative time-series data for the United States economy. Link Here we examine only one model in several variants so further investigation of the robustness of our findings is needed, but the present study represents a start.

Before proceeding to our theoretical model, however, we provide evidence on the empirical relevance of nominal income targeting, by comparing the fit on U.S. data of estimated Taylor rules with that of interest rate rules that respond to nominal income growth.

Nominal Income Targeting and Historical Policy

In his now-famous paper, Taylor (1993a) demonstrated that U.S. monetary policy in the period since 1987, a period considered very successful in terms of delivering low inflation alongside relatively stable and satisfactory output growth, can be well characterized as a regime in which the federal funds rate responds with positive, fixed coefficients to (proxies for) expected inflation and the output gap. In a recent paper, Clarida, Gali, and Gertler (CGG) (1997), using instrumental variables estimation over the period 1979:3-1996:4 (a sample period beginning with the onset of Chairman Volcker’s incumbency), provide formal econometric corroboration of Taylor’s finding.

Their regression for the federal funds rate, which employs a partial adjustment specification to allow for interest rate smoothing, indicates that expected inflation enters with a long-run coefficient of approximately 1.5, in keeping with Taylor (1993a); the estimated coefficient on the output gap is, however, smaller than the 0.5 value that Taylor used.

Tags: , ,