THE INVESTIGATIONS REPORTED in this paper are directed to the question of the type of monetary rule which should be used if a rule is to be adopted. The rule usually considered in discussions of "rules versus discretion" is that of a constant growth rate of the money supply [4,7,8]. However, results in control theory (introduced to economics by Tustin [14], Phillips [12] and Holt [9], and investigated more recently by others) suggest that rules in which policy variables are automatically adjusted in response to deviations of target variables from desired levels (proportional controls) and in response to the rate of change of target variables controls) can be stabilizing relative to a constant growth rate rule. Whether such controls produce significant gains in stability in models of the economy is not, however, a trivial question, for these models may have characteristics-such as very long lags-which make the gain from use of adaptive controls of little significance. For instance, it will be seen, in our simulations, that proportional controls do not lead to sizeable increases in stability. In this paper, we report the results of simulations of some monetary rules in the FRB-MIT-Penn (FMP) Econometric Model for the period 1956.I through 1968.IV. The simulations are directed to the question of whether, in this model, for the period covered, the variability of the rates of inflation and unemployment produced under a constant growth rate rule can be reduced by following some simple rule of