Hausman vs Sargan-Hansen test

[Darcy Hill]

I would like to test whether FE or RE model is more appropriate, however I suspect clustered errors and heteroskedasticity, Is it correct to run Sargan-Hansen test rather than Hausman?

Thank You

[Carlo Lazzaro]

Darcy:
yes, it is.
If you impose cluster-robust standard errors, -hausman- is out of debate and you should consider the user-written command -xtoverid- (unfortunately, being a bit old-fashioned, the glorious -xtoverid- does not accept -fvvarlist- notation; see -help xi:-), as you can see from this following toy-example:

. use "http://www.stata-press.com/data/r15/nlswork.dta"
(National Longitudinal Survey.  Young Women 14-26 years of age in 1968)

. quietly xtreg ln_wage age, fe vce(cluster idcode)

. estimates store fe

. quietly xtreg ln_wage age, re vce(cluster idcode)

. estimates store re

. hausman fe re
hausman cannot be used with vce(robust), vce(cluster cvar), or p-weighted data
r(198);

. xtoverid

Test of overidentifying restrictions: fixed vs random effects
Cross-section time-series model: xtreg re  robust cluster(idcode)
Sargan-Hansen statistic  14.529  Chi-sq(1)    P-value = 0.0001

*-xtoverid-  outcome points to -fe- specification*

[Darcy Hill]

Thank You