Estimating a Rational Expectations Business Cycle Model: Methodology and Application (Lucas Supply, New Classical).

Macroeconomic events of the 1970's have led to dissatisfaction with the Keynesian theory of business cycles. An alternative theory based on intertemporal labor substitution, market clearing, and rational expectations, and associated with the name of Robert Lucas, is at the center of current debate in macroeconomics. Yet little empirical work has been done on "New Classical" theory, and existing empirical evidence is inconclusive. The approach of this thesis is to construct a small macroeconometric model in which a Lucas supply function determines aggregate supply. This requires that the supply function be estimated in a way consistent with rational expectations. A synthesis of the literature on estimating rational expectations models, focusing on single equation techniques, is provided. A single equation estimator is proposed for a class of models of which the Lucas supply function is a special case. Its properties are rigorously established under a number of alternative assumptions concerning the specification of the equation, the generation of the exogenous variables of the model, and the generation of the disturbance. A new result connected with inference is presented: it is shown that the estimated coefficient variances are downward biased asymptotically. This methodology is applied to an empirical Lucas supply equation. The estimated model in which the supply function is embedded is an empirical version of the inventory dynamics model of A. Blinder and S. Fischer. Inventories serve, in combination with the rest of the model, to propagate the serially uncorrelated nominal price "impluses" into serially correlated fluctuations of output about its natural level. Of the propagation mechanisms discussed in the literature this solution to the "persistence problem" in New Classical models would appear to be both the most realistic, in terms of the plausibility of its role in the business cycle, and the most easily implemented econometrically. The estimation results for the supply equation fail to support the intertemporal labor substitution hypothesis. However the hypothesis is not rejected in favor of any well defined alternative. Versions of the model excluding the supply equation are simulated in order to establish that its dynamic behavior is on the whole reasonable.PhDEconomic theoryUniversity of Michiganhttp://deepblue.lib.umich.edu/bitstream/2027.42/161031/1/8612629.pd