Return to search

An empirical comparison of autoregressive and rational models of price expectations

This dissertation presents several empirical tests to measure the relative abilities of alternative models in capturing the unobservable process by which economic individuals may form expectations of future inflation. Three empirical representations of the inflation expectations process are tested: an autoregressive model which uses only past inflation data; a rational expectations model which utilizes the structural economic relationships in the economy (excluding past inflation); and a general model which exploits both the information sets just described.

These competing approaches are each subjected to tests for rationality and predictive accuracy. The rationality tests employed in this study are the breakpoint test suggested by Sargent and the incorporation of each model's inflation predictions into an analysis of the Fisher equation. To gauge the predictive accuracy of each model, post-sample extrapolations were generated and compared by means of the root-mean-squared error and Theil inequality coefficient.

The outcome of these various tests provides support to the contention that, for an individual attempting to obtain optimal (error minimizing) forecasts of future inflation would select the relatively simple autoregressive model over the rational expectations or general approaches. In three out of four tests presented, the autoregressive model performed as well if not better than its more informational intensive competitors. / Ph. D.

Identiferoai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/76551
Date January 1979
CreatorsHafer, R. W.
ContributorsEconomics
PublisherVirginia Polytechnic Institute and State University
Source SetsVirginia Tech Theses and Dissertation
Languageen_US
Detected LanguageEnglish
TypeDissertation, Text
Formativ, 126 leaves, application/pdf, application/pdf
RightsIn Copyright, http://rightsstatements.org/vocab/InC/1.0/
RelationOCLC# 5201117

Page generated in 0.0016 seconds