Hi everyone Im struggling a bit would apprechiate some help:
my inital fixed effects model is this:
xtreg lntobinsq lnassets Deriv IR bookleverage roa_w1 cratio_w1 rnd_rev_w1 cash_to_totalassets div_yield_w1 yeardummy if inlist(year,2015,2016)
1) How do I test for hetroskedascity for this fe model, I read around and found it to be xttest3? is groupwise hetroskedascity the same as normal hetro?
2) I am trying to perform a robust hausman on this model via the xtoverid
a) I know you can't include the time dummies in the simple hausman but can you include the time dummies in this robust hausman test:
b) also want to know how to accurately implement this test: if I used the exact command above and tried, would it be like this:
if this is wrong please let me know how to do it exactly. Thanks
and if this is right does the very low p value indicate we use FE?
3)Is the resulting test stat equal to a hausman stat or completely different?
4) I am doing this for my project so do I only run the robust Hausman, or do I run both the normal and then the robust hausman xtoverid (or is this wrong) just asking as I would like to show the comparison of the coefficients between fe and re in the appendix or something and was wondering if this is even possible in this situation)
Thanks everyone for being so helpful and insightful, any help is apprechiated. Thanks
my inital fixed effects model is this:
xtreg lntobinsq lnassets Deriv IR bookleverage roa_w1 cratio_w1 rnd_rev_w1 cash_to_totalassets div_yield_w1 yeardummy if inlist(year,2015,2016)
1) How do I test for hetroskedascity for this fe model, I read around and found it to be xttest3? is groupwise hetroskedascity the same as normal hetro?
2) I am trying to perform a robust hausman on this model via the xtoverid
a) I know you can't include the time dummies in the simple hausman but can you include the time dummies in this robust hausman test:
b) also want to know how to accurately implement this test: if I used the exact command above and tried, would it be like this:
Code:
. xtreg lntobinsq lnassets Deriv IR bookleverage roa_w1 cratio_w1 rnd_rev_w1 cash_to_totalassets div_yield_w1 year2016 if inlist(year,2015,2016), fe robust
Fixed-effects (within) regression Number of obs = 586
Group variable: firmid Number of groups = 306
R-sq: Obs per group:
within = 0.3854 min = 1
between = 0.1586 avg = 1.9
overall = 0.1726 max = 2
F(10,305) = 15.70
corr(u_i, Xb) = -0.7459 Prob > F = 0.0000
(Std. Err. adjusted for 306 clusters in firmid)
-------------------------------------------------------------------------------------
| Robust
lntobinsq | Coef. Std. Err. t P>|t| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
lnassets | -.4470604 .0687788 -6.50 0.000 -.5824015 -.3117193
Deriv . | .0766741 .0925322 0.83 0.408 -.1054082 .2587564
IR . | -.0954453 .0659056 -1.45 0.149 -.2251325 .0342419
bookleverage | .1864517 .120672 1.63 0.105 -.0410034 .4339067
roa_w1 | .014447 .0077733 1.86 0.064 -.0008491 .029743
cratio_w1 | -.0616183 .0234384 -2.63 0.009 -.1077398 -.0154968
rnd_rev_w1 | -.0150043 .014466 -1.04 0.300 -.0434701 .0134615
cash_to_totalassets | .1879683 .2583643 0.69 0.491 -.3304338 .6863704
div_yield_w1 | -.053192 .0087376 -4.94 0.000 -.0603856 -.0259985
year2016 | .0303004 .0138687 1.46 0.144 -.0069901 .0475909
_cons | 3.728503 .4793996 7.99 0.000 2.885154 4.771853
--------------------+----------------------------------------------------------------
sigma_u | .76302993
sigma_e | .11779167
rho | .9767235 (fraction of variance due to u_i)
-------------------------------------------------------------------------------------
.
. est store fixed
.
. xtreg lntobinsq lnassets FXDerivatives10 IRDerivatives10 bookleverage roa_w1 cratio_w1 rnd_rev_w1 cash_to_totalass
> ets div_yield_w1 year2016 if inlist(year,2015,2016), re robust
Random-effects GLS regression Number of obs = 586
Group variable: firmid Number of groups = 306
R-sq: Obs per group:
within = 0.2040 min = 1
between = 0.6631 avg = 1.9
overall = 0.6317 max = 2
Wald chi2(10) = 203.50
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
(Std. Err. adjusted for 306 clusters in firmid)
-------------------------------------------------------------------------------------
| Robust
lntobinsq | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
lnassets | -.0670922 .0143928 -3.97 0.000 -.0853017 -.0288828
Deriv | .0849556 .0341868 2.78 0.005 .0279506 .1619606
IR | -.1269466 .038487 -3.30 0.001 -.2023798 -.0515133
bookleverage | .0960349 .0816089 1.14 0.254 -.0669155 .2529853
roa_w1 | .05175 .0101699 5.09 0.000 .0318173 .0716826
cratio_w1 | -.0658327 .0151559 -4.94 0.000 -.1045376 -.0451277
rnd_rev_w1 | .0149893 .0040661 3.69 0.000 .0070198 .0229588
cash_to_totalassets | .4445181 .1900494 2.34 0.019 .0720282 .817008
div_yield_w1 | -.0456181 .0082363 -5.90 0.000 -.0647609 -.0324753
year2016 | -.0151233 .010654 -1.51 0.130 -.0370048 .0047583
_cons | .8223877 .1466037 5.61 0.000 .5350496 1.109726
--------------------+----------------------------------------------------------------
sigma_u | .28804944
sigma_e | .11779167
rho | .85673459 (fraction of variance due to u_i)
-------------------------------------------------------------------------------------
.
.
. xtoverid
Test of overidentifying restrictions: fixed vs random effects
Cross-section time-series model: xtreg re robust cluster(firmid)
Sargan-Hansen statistic 100.089 Chi-sq(10) P-value = 0.0000
and if this is right does the very low p value indicate we use FE?
3)Is the resulting test stat equal to a hausman stat or completely different?
4) I am doing this for my project so do I only run the robust Hausman, or do I run both the normal and then the robust hausman xtoverid (or is this wrong) just asking as I would like to show the comparison of the coefficients between fe and re in the appendix or something and was wondering if this is even possible in this situation)
Thanks everyone for being so helpful and insightful, any help is apprechiated. Thanks
Comment