Timeline for Hausman test FE vs. RE in case that FE is not consistent
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 29, 2014 at 12:04 | comment | added | Andy | That's right, IV estimation is more inefficient. If you have good theoretical reason to believe that the fixed effects are correlated with some of the other explanatory variables I would go with that. | |
| Dec 29, 2014 at 10:46 | vote | accept | František Kaláb | ||
| Dec 29, 2014 at 10:46 | comment | added | František Kaláb | This make sense! There is only one problem that as the IV estimation is not efficient, the power of the test is significantly lowered. In my case the standard errors are high enough to not reject any difference, but still FE are probably much more justifiable by economic reasoning.. | |
| Dec 16, 2014 at 15:47 | history | edited | Andy | CC BY-SA 3.0 |
added 72 characters in body
|
| Dec 16, 2014 at 15:42 | history | answered | Andy | CC BY-SA 3.0 |