The Role of Beliefs in Inference for Rational Expectations Models
This paper discusses inference for rational expectations models estimated via minimum distance methods by characterizing the probability beliefs regarding the data generating process (DGP) that are compatible with given moment conditions. The null hypothesis is taken to be rational expectations and the alternative hypothesis to be distorted beliefs. This distorted beliefs alternative is analyzed from the perspective of a hypothetical semiparametric Bayesian who believes the model and uses it to learn about the DGP. This interpretation provides a different perspective on estimates, test statistics, and confidence regions in large samples, particularly regarding the economic significance of rejections in rational expectations models.
Published Versions
Lehmann, Bruce N. "The Role of Beliefs in Inference for Rational Expectations Models." Journal of Econometrics 150, 2(June 2009): 322-331.