I am running a fe regression on panel data and had some basic questions on how to interpret the results:
1. R-sq within group is just 0.0306 seems to be extremely small. How should I be interpreting this number?
2. There are 2 F-values i,e, F(12,1099) with value 2.89 and F(122, 1099) with value 19.64 . My understanding is that the second one is a reflection of the quality of the overall model. Would that be a correct assumption? If so, what does the first F statistic represent?
(In case my questions are too basic my apologies. Please feel free to point me to any book material where I could find the answers. Thank you.
Code:
. xtreg ROA_win05 Ln_Revenue Ln_LTDTA CoAge TPSD wGDPpc wCPI wDCF wExpgr wGDPgr wCons c.l1.G > SD##c.l1.GSD if FOREIGNSALESTOTALSALES >10 & Year_ < YearInactive, fe Fixed-effects (within) regression Number of obs = 1,234 Group variable: n_CUSIP Number of groups = 123 R-sq: Obs per group: within = 0.0306 min = 1 between = 0.4929 avg = 10.0 overall = 0.3306 max = 18 F(12,1099) = 2.89 corr(u_i, Xb) = 0.3905 Prob > F = 0.0006 ------------------------------------------------------------------------------- ROA_win05 | Coef. Std. Err. t P>|t| [95% Conf. Interval] --------------+---------------------------------------------------------------- Ln_Revenue | 2.073987 .6697842 3.10 0.002 .7597872 3.388188 Ln_LTDTA | -.7476576 .1945061 -3.84 0.000 -1.129303 -.3660123 CoAge | -.0010398 .0901846 -0.01 0.991 -.1779934 .1759137 TPSD | -.9328759 .9147456 -1.02 0.308 -2.727721 .8619692 wGDPpc | -.0000301 .0001556 -0.19 0.847 -.0003354 .0002753 wCPI | -.2623995 .2668114 -0.98 0.326 -.7859167 .2611178 wDCF | -2.71e-14 1.20e-12 -0.02 0.982 -2.39e-12 2.33e-12 wExpgr | .1684812 .0990232 1.70 0.089 -.0258147 .3627772 wGDPgr | -.3395405 .2691567 -1.26 0.207 -.8676594 .1885785 wCons | -5.39e-14 4.31e-13 -0.13 0.900 -8.99e-13 7.91e-13 | GSD | L1. | 7.670794 3.519331 2.18 0.029 .7654273 14.57616 | cL.GSD#cL.GSD | -5.572459 2.683354 -2.08 0.038 -10.83754 -.3073827 | _cons | -39.89984 13.55392 -2.94 0.003 -66.49434 -13.30535 --------------+---------------------------------------------------------------- sigma_u | 16.843882 sigma_e | 7.0694025 rho | .85023217 (fraction of variance due to u_i) ------------------------------------------------------------------------------- F test that all u_i=0: F(122, 1099) = 19.64 Prob > F = 0.0000
0 Response to Interpreting STATA FE regression output
Post a Comment