| Scope and method of study. This study utilizes time series data from 1869 to 1998 to examine the statistical relationship between the monetary aggregates and economic activity. The vector autoregression (VAR) technique is employed for the examination. Forecast error variances from variance decompositions and impulse response functions are produced in the analysis. Four time periods are analyzed in the study: The Friedman and Schwartz era over 1869–1960, the St. Louis era over 1952–1968, the era 1968–1981 in which velocity's trend remained fairly stable, and the period 1981–1998 in which the monetary aggregates fell out of favor.; Findings and conclusions. The monetary aggregates maintain a fairly close statistical relationship to economic activity in the Friedman and Schwartz era over 1869–1960, the St. Louis era over 1952–1968, and the era 1968–1981. However, the monetary aggregates do not maintain a close statistical relationship with economic activity in the period 1981–1998. In addition, the Federal funds rate, a competing monetary indicator, also shows a less close relationship to economic activity in the period 1981–1998. Of the monetary aggregates tested, the M1 money supply displays a marked deterioration in its predictive ability from the period 1968–1981 to 1981–1998. The MZM money supply displays some decline in forecast ability but maintains a closer relationship to GDP than does M1. The aggregate L, a broad measure of total liquidity, declines substantially in its predictive ability. |
