Monetary policy, real economic activity, and new-classical macroeconomics

The dissertation studies the effect of money on real output. It consists of two interrelated major sections, one theoretical and one empirical.;Previous rational expectations models (both new-classical and new-Keynesian) are based on an arbitrary assumption of either all auction markets (market clearing) or all customer markets (non-market clearing). Therefore, the first section develops rational expectations models which incorporate both types of markets. We allow critical parameter values, such as the weight of forward looking expectations, the price elasticity of customer demand, and the share of auction markets, to vary. Major findings are as follows:;First, when the wage is a function of the real cost of living, auction markets and expectations have a similar effect on the pattern of price and output determination. Second, the new-classical's proposition that only unanticipated money affects output holds only when agents are completely forward looking. Under realistic parameter values, our model shows a significant effect of anticipated money on output.;The empirical section tests the new-classical's hypothesis of the ineffectiveness of anticipated money. Since unanticipated money is not directly observable, it must be estimated. The conventional method of estimating anticipated money, as in Barro (1977), is to take residuals from a money equation regression over a certain sample period using data published at a later date. Since this method uses unavailable future information, it violates the underlying assumption of rational expectations. We therefore estimated a money equation recursively confining the sample only to the period of expectation. In addition, we used the actual historical data that agents would have had available, obtained from various issues of the Economic Report of the President (1949 through 1974). Using the unanticipated money series thus obtained, the new-classical's hypothesis can not be accepted.