Working Paper: NBER ID: w2150
Authors: Ray C. Fair
Abstract: This paper uses stochastic simulation and my U.S. econometric model to examine the optimal choice of monetary policy instruments. Are the variances, covariances, and parameters in the model such as to favor one instrument over the other, in particular the interest rate over the money supply? The results show that the interest rate and the money supply are about equally good as policy instruments in terms of minimizing the variance of real GNP. The variances of some of the components of GNP are, however, much larger when the money supply is the policy instrument, as is the variance of the change in stock prices. Therefore, if one's loss function is expanded beyond simply the variance of real GNP to variances of other variables, the interest rate policy does better. The results thus provide some support for what seems to be the Fed's current choice of using the interest rate as its primary instrument. Stochastic simulation is also used to estimate how much of the variance of real GNP is due to the error terms in the demand for money equations. The results show that the contribution is not very great even when the money supply is the policy instrument .
Keywords: Monetary Policy; Econometric Models; Stochastic Simulation
JEL Codes: E52; C15
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Choice of policy instrument (E64) | Variance of real GNP (E20) |
| Choice of policy instrument (E64) | Variances of consumption of durables (E21) |
| Choice of policy instrument (E64) | Variances of housing investment (R21) |
| Expanded loss function (C51) | Choice of policy instrument performance (E61) |
| Interest rate policy (E43) | Broader economic outcomes (F69) |