Academic journal article Atlantic Economic Journal

Some Model-Based Tests of Expectational Rationality

Academic journal article Atlantic Economic Journal

Some Model-Based Tests of Expectational Rationality

Article excerpt

Some Model-Based Tests of Expectational Rationality

I. Introduction

The rational expectations hypothesis has generated a tremendous upheaval in the economics profession partly because, initially, it was mistakenly believed to be the only source of systematic policy ineffectiveness. It is now well understood that expectational rationality is a necessary but not a sufficient condition for systematic policy neutrality. This realization by the profession has led to a toning down of the attacks on the hypothesis.

Rational expectations differ from their main competitor in the field of expectations formation mechanisms--namely adaptive expectations--in three major respects. First, rational expectations are forward looking rather than simply being extrapolations of past trends; second, agents are acting in an optimizing fashion by processing all the relevant information; and third, the rational expectations approach provides a central role for economic theory in determining expectations.

Many economists find the rational expectations hypothesis appealing because it has replaced earlier ad hoc treatments with an approach squarely based on incentives, information, and optimization--principles widely considered as valuable macroeconomic foundations. The hypothesis has also received some criticisms that have been levelled at the rational expectations hypothesis is included in Begg [1982].

Most empirical tests of the hypothesis have employed survey data on expectations and have attempted to establish whether the statistical properties of expectational rationality--unbiasedness, efficiency and orthogonality--are satisfied by the data.

It is perhaps fair to say that, in general, empirical studies based on survey data on expectations do not support the rational expectations hypothesis [see, Brown and Maital, 1981; Figlewski and Wachtel, 1981; Gramlich, 1983]. Only some of the econometric forecasts offer some support [McNees, 1978]. The last paper employs forecasts of three prominent U.S. econometric models to test for the properties of expectational rationality. Although its results are mixed, they are mildly supportive of the rational expectations hypothesis.

It must be pointed out, however, that the costs of making large errors are very small for a respondent asked to complete a questionnaire under strict anonymity. Therefore, it is not clear that those completing surveys of expectations have any incentive to utilize all available information at the time of the forecast. This point does not explain rejections of the property of unbiasedness by survey data, but might explain rejections of efficiency and orthogonality.

In view of the limitations of survey data on price expectations and the absence of a widely accepted survey data series for Canada, this paper adopts a different approach to testing for expectational rationality. The tests of the rational expectations hypothesis that are performed in the paper are discussed in [Wallis, 1980; Revankar, 1980; Hoffman and Schmidt, 1981].

The test procedure is based on the recognition that, in principle, one can test for expectational rationality, given the validity of the macroeconomic structure, by imposing the non-linear restrictions that such rationality implies both within and across the equations of the macromodel.

The structure to be estimated is presented and discussed in Section II of the paper. The same section lists the restrictions imposed by expectational rationality.

It is shown that the rationality restrictions critically depend on the nature of the autoregressive processes of the forcing variables. Different sets of lag structrues are tried, in Section III, in an attempt to discover the nature of the autoregressive processes of the forcing variables. The results of the rationality tests are reported in Section IV and, finally, Section V offers some conclusions. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.