Question

Hausman statistic is negative. What to do?

Answer 1

The Hausman test shows the choice between random effects and fixed effects regressions. Thus we take the absolute value of the statistic. The null hypothesis means that the random effects regression is better while the fixed effects regression is better under the alternative hypothesis. This statistic follows the chi squared distribution and so we look at the calculated value against the tabulated value. Ther statistic being negative does not matter. We take the absolute value and check with the chi square tables.