I want to know if I can run a Hausman test for Random Effects and First-Difference models with different regressors? For Example:
MODre = X + Y + Z
MODfd = X + Y + C
Where X, Y, Z, C are some explanatory variables.
Thank You.
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
The Hausman test only takes the common variables of both models (and compares them). The test does not tell you anything about the non-common variables as those are not compared and dropped in the test (for most statistic packages).
