Rational Financial Expectations? An Empirical Analysis of British Household Panel Data

Rational Financial Expectations? An Empirical Analysis of British Household Panel Data

This paper tests the validity of the rational expectations hypothesis (REH) using twelve waves of the British Household Panel Survey, which gives a very large cross section of individual units over a 12-year period. Most previous test of the REH have used either aggregated macroeconomic data or micro-data for short sample periods only. Both of these approaches are problematic since aggregated data can lead to spurious rejections of rationality when agents’ information sets differ, whilst the micro-data cannot efficiently average out forecast errors over a short period of time even if individual forecasts are perfectly rational. It is therefore important to test rationality using micro-data on expectations over long sample periods. To this end we undertake two kinds of test. First, we test whether expectations errors accord with the REH or whether they contain systematic components; that is, whether the forecasts are unbiased and efficient. Second, we use a random effects ordered probit model to investigate the characteristics of the subjective data by describing the relationship between individual households’ financial expectations and a set of explanatory variables, including realised financial changes in the past, a set of demographic variables, household characteristics and a set of time dummies. The preliminary empirical results suggest that households’ financial expectations are significantly biased and that there is evidence of a significant heterogeneity effect. In terms of the efficiency of the change in financial expectations, the expectations errors are largely unsystematic, although the demographic variables are jointly significant, which is counter to the assumption of efficiency. Finally, there are number of interesting results from the subsequent examination of the relationship between the change in financial expectations and the subsequent financial change outcome. For example, people whose financial situation has improved in the past tend to be more sensitive than those whose financial  position has deteriorated; the expectations of men, the married and those with only a secondary level of education are more sensitive to their financial outcomes than others; finally, comparing the expected and realised financial outcomes over the same time period suggests that the married, women, the paid employed and people living in smaller households are more pessimistic than others, whilst smokers and the unemployed are over-optimistic.