Expectations in micro data: rationality revisited

An increasing number of longitudinal data sets collect expectations information regarding a variety of future individual level events and decisions, providing researchers with the opportunity to explore expectations over micro variables in detail. We present a theoretical framework and an econometric methodology to use that type of information to test the Rational Expectations (RE) hypothesis in models of individual behavior. This RE assumption at the micro level underlies a majority of the research in applied fields in economics, and it is the common foundation of most work in dynamic models of individual behavior. We present tests of three different types of expectations using two different panel data sets that represent two very different populations. In all three cases we cannot reject the RE hypothesis. Our results support a wide variety of models in economics, and other disciplines, that assume rational behavior. We would like to acknowledge outstanding research assistance from Huan Ni. The Michigan Retirement Research Center (MRRC) and the TIAA-CREF Institute made this research possible through their financial support of two related projects. Benítez-Silva also acknowledges the financial support from NIH grant AG1298502 on a related project, and also from the Fundación BBVA, and the Spanish Ministry of Science and Technology through project number SEJ2005-08783-C04-01, and wants to thank the Department of Economics at the University of Maryland and the Department of Economics at Universitat Pompeu Fabra for their hospitality during the completion of this paper. Three anonymous referees provided excellent comments and suggestions. Any remaining errors are the authors’.