|
|
|
|
|
|
|
|
|
|
|
Soo-Hwa Lee
Dr. Ronald A. Ratti, Dissertation Supervisor
ABSTRACT
|
This paper investigates the neutrality proposition of money focusing on the roles played by various definitions of the natural rate of real economic variables. Using postwar U.S. GNP data, this paper's tests reject Mishkin's trend stationary process(TSP) hypothesis of the natural rate in the GNP models in which the first two lagged money terms are included, but does not reject Mishkin's hypothesis in the models beyond two lagged money terms. This finding is consistent with Nelson and Plosser's(1982) results of the presence of a unit root in GNP.
Empirical investigations assuming different natural rate definitions also provide contradictory results of neutrality tests, the ARCH effect, Friedman's short run tradeoff, and the forecastability of recessions. For example, the model based on a TSP rejects the neutrality proposition, while the model based on a difference stationary process(DSP) cannot reject the proposition in the case of the money forecasting model proposed in this paper. In addition, it is found that TSP model has a tendency to exaggerate the effect of anticipated money on real economic activity, while DSP model tends to make unanticipated money more important than anticipated money (see, Gochoco 1986; Eichenbaum and Singleton 1986). In order to avoid definition-dependent results about the length of the short run, a series of filtered data are used as proxies for the short run components. The results show that the one-year component of TSP model demonstrates Friedman's short run tradeoff, while the entire DSP model does not.
To the extent that TSP and DSP hypotheses affect the empirical results of the testing of the natural rate hypothesis, this paper warns that studies that have failed to carefully consider an appropriate model for the natural growth rate may have reached misleading conclusions about the neutrality proposition of money.
 
|
|
|
|
|
 |
|
