Northwestern University and Federal Reserve Bank of Chicago
This paper suggests a new procedure for evaluating the fit of a dynamic structural economic model. The procedure begins by augmenting the variables in the model with just enough stochastic error so that the model can exactly match the second moments of the actual data. Measures of fit for the model can then be constructed on the basis of the size of this error. The procedure is applied to a standard real business cycle model. Over the business cycle frequencies, the model must be augmented with a substantial error to match data for the postwar U. S. economy. Lower bounds on the variance of the error range from 40 percent to 60 percent of the variance in the actual data.
Economists have long debated appropriate methods for assessing the empirical relevance of economic models. The standard econometric approach can be traced back to Haavelmo (1944), who argued that an economic model should be embedded within a complete probability model and analyzed using statistical methods designed for conducting inference about unknown probability distributions. In the modern literature, this approach is clearly exemplified in work such as that of L. Hansen and Sargent (1980) or McFadden (1981). However, many economic models do not provide a realistic and complete
This paper has benefited from constructive comments by many seminar participants; in particular I thank John Cochrane, Marty Eichenbaum, Jon Faust, Lars Hansen, Robert Hodrick, Robert King, and Robert Lucas. Two referees also provided valuable constructive criticism and suggestions. The first draft of this paper was written while I was visiting the University of Chicago, whose hospitality is gratefully acknowledged. This research was supported by the National Science Foundation through grants SES-89-10601 and SES-91-22463.
[Journal of Political Economy, 1993, vol. 101, no. 6]
© 1993 by The University of Chicago. All rights reserved. 0022-3808/93/0106-0007$01.50