Hausman test

Quick Reference

A common name for model specification tests in which one proposed estimator of a parameter is both consistent and efficient under the null hypothesis and inconsistent under the alternative, whereas the other proposed estimator of the same parameter is consistent both under the null and under the alternative hypotheses but inefficient under the null. Important applications include the test for random effects against fixed effects, and the test for exogeneity (instrumental variables against ordinary least squares.)

Subjects: Economics.

Reference entries