Abstract
The choice of the best monetary aggregate for use as the target variable in conducting monetary policy has been an unsettled question for many years in economics. The problem has recently become more important for two reasons; first, the Federal Reserve has been setting objectives in growth rates for three aggregates, and is moving away from an interest rate target in favor of a money supply target. Secondly, the increase in number and importance of "near monies" in the last two decades has increased the number of monetary aggregates which could serve as the target. This study compares seven potential target aggregates, the monetary base and Ml through M6. Three criteria are posited as desirable properties of a monetary target and each aggregate is compared against these criteria. The first is the ability of the aggregate to accurately forecast GNP; the second is causality from the aggregate to the base; and the third is controllability of the aggregate by the Federal Reserve. The study uses quarterly data from 1947-1 through 1976-1. To compare GNP forecast ability, the relative stability of the income velocity function of each aggregate is calculated, using root mean squared error terms as the test statistics. Six different velocity function formulations are used for each aggregate, and the Orcutt method for eliminating first order serial correlation is employed. Overall, Ml has the most stable income velocity function regardless of the specific functional form used..
Rivard, Richard Joseph (1978). A comparison of monetary aggregates as monetary policy targets. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -638240.