A Suboptimal Method for Feedback Control of the St. Louis Econometric Model

Abstract

In this study, the econometric model used by the Federal Reserve Bank of St. Louis to provide information on the most likely movement of strategic economic variables in response to monetary and fiscal actions is viewed as a control system. Government expenditure is viewed as an uncontrolled disturbance, while change in the money stock is viewed as the input control variable. Quarterly change in the money stock is treated as a linear function of errors in economic variables from established target values. A numerical gradient scheme is then applied to the proportionality constants of the function so as to minimize a quadratic performance index. Throughout the study, the nonlinear system equations are retained. The results of this study indicate that automatic control of the money stock by this method would have improved the nation’s economy over the last fifteen years.