THE PURPOSE OF THIS DISSERTATION is to develop and estimate for the post-Accord, 1952-65 period, a policy action function applicable to the Federal Reserve System. The policy action function relates a preferred indicator of monetary policy action, as a dependent variable, to variables affecting monetary policy and to forecasts of exogenous variables, which affect these variables but are not affected by monetary policy action. This problem has received inadequate attention in most theoretical and policy-oriented work. The policy action function is derived from assumptions about lags and exogenous variables affecting policy and from assumptions about policymaker behavior. As suggested by a review of earlier studies, without these assumptions, movements in the chosen indicator of monetary policy could be misinterpreted. For example, changes in a popular indicator, the level of free reserves, could be induced either by policy action, by movements in the asset demand functions of the banks, or by feedbacks from earlier policy action. Total reserves adjusted for legal reserve requirement changes is selected as a preferred indicator of policy action. It is considered, for purposes of this study, superior to all other indicators because feedback from the real sector to it is most readily known and compensated for by the policymaker, because it reflects changes in the three conventional tools of policy (open market operations, changes in the discount rate, and reserve requirement changes) and because it is immediately measurable. This indicator enters into the money supply function in a Keynesian-type model of the economy. The transmission of policy is postulated as flowing from the indicator to the money supply to the long-term interest rate to investment to aggregate demand to unemployment, the price level, and other real variables affecting policy. The lag from a change in the long-term interest rate until an effect is observed on the real variables is assumed to be one quarter. After substituting endogenous solutions for the real variables into a simple utility function for the policymaker, a policy action function is mathematically derived. It contains the indicator as the dependent variable with both endogenous and several forecasted exogenous variables as independent variables. This function is tested by standard single-equation least squares regression analysis. The results suggest that the policymaker responded systematically to the real variables: unemployment less a 4 percent goal, the squared difference between the general price level and its goal, a balance of payments measure (first differences in foreign-held, short-term liabilities as a percentage of gold reserves), and the forecasted level of income. In addition, foreign economic activity (represented by a proxy for foreign, long-term interest rate levels) evoked a policymaker response, as it affects the payment's balance. Also shifts in the balance of payments situation in the 1961-65 period, five years of heavy short-term capital flows, induced responses by the policymaker. In the conceptual framework of this study, these were the actual determinants of monetary policy action.