Hausman Test Failed

[Dewi Karina]

Hi! my name is Karina. I run a Hausman test on my stata using command:

quietly xtreg ecgrowth jubgrowth sukubunga inflation, fe
estimates store fe
quietly xtreg ecgrowth jubgrowth sukubunga inflation, re
estimates store re
hausman fe re

and it come up with this:

chi2(3) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= -2.36 chi2<0 ==> model fitted on these
data fails to meet the asymptotic
assumptions of the Hausman test;
see suest for a generalized test

after searching for a little while, i found another option for doing the hausman test; it's either using xtoverid or sigmaless/sigmamore. Later, i try to use this as my command:

xtreg ecgrowth jubgrowth sukubunga inflation if t>=2004 & t<=2021, re
xtoverid
di r(j)
est store re
xtreg ecgrowth jubgrowth sukubunga inflation if t>=2004 & t<=2021, fe
hausman fe re, sigmaless

is this the right command to run the hausman test? if not, then would anyone help me to give the right command? im so confused and on the edge of a thesis deadline. Thank you for anyone replying!

[Carlo Lazzaro]

Dewi:
welcome to this forum.
Since your -xtreg,re- code does not show -fvvarlist- notation, the community-contributed module -xtoverid- (that, unlike -hausman-, supports cluster-robust standard errors but does not allow -fvvarlist- notation) comes in handy:

xtreg ecgrowth jubgrowth sukubunga inflation if t>=2004 & t<=2021, re
xtoverid

The null of -xtoverid- is: -re- is the way to go.