Expected versus Realized Income Changes: A Test of the Rational Expectations Hypothesis

We analyze answers to household survey questions on whether household income has changed in the past twelve months, and on whether respondents expect their household income to change in the next twelve months. Both questions are answered on a five points scale. The data are an unbalanced panel of eleven consecutive annual waves. Using cross-tabulations of expected and realized changes, we first test the ”bestcase” hypothesis, implying that respondents have rational expectations, that there are no common unexpected shocks, and that reported expectations are the best predictions of future outcomes. We reject the best case hypothesis: for all years, too many of those who predict an income fall, report ex post that their household’s income has not changed. We then construct a dynamic bivariate ordered probit random effects panel data model, in which we explain expected and realized income changes from background variables and the one year lagged expectation and realization. In this model, the hypothesis of rational expectations implies certain restrictions on the parameters. The model is estimated by simulated maximum likelihood using the Geweke-Hajivassilou-Keane (GHK) method. The hypothesis of rational expectations is rejected. The hypotheses that expectations are adaptive or naive can be tested in a similar way, and are also rejected. Statistics Netherlands is acknowledged for providing the data. We are grateful to Bas Donkers for valuable comments. Please address correspondence to Arthur van Soest, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands (e-mail: [email protected])