I have a panel of N = 3810 & T = 6, and after reading forum posts have run areg, absorb(company) along with xtreg, fe vce(cluster company) and the results are as below. Why is there such a large discrepancy in my R2? My model aims to establish the link between a company's environmental score over time vs its financial performance through its Tobin's Q ratio


Code:
. xtreg tobinq env_score de lnta, fe vce(cluster company)

Fixed-effects (within) regression               Number of obs     =      2,634
Group variable: company                         Number of groups  =        611

R-sq:                                           Obs per group:
     within  = 0.0114                                         min =          1
     between = 0.0602                                         avg =        4.3
     overall = 0.0423                                         max =          6

                                                F(3,610)          =       7.13
corr(u_i, Xb)  = -0.0341                        Prob > F          =     0.0001

                              (Std. Err. adjusted for 611 clusters in company)
------------------------------------------------------------------------------
             |               Robust
      tobinq |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   env_score |   .0046432   .0015731     2.95   0.003     .0015539    .0077326
          de |   3.04e-06   1.17e-06     2.60   0.010     7.39e-07    5.33e-06
        lnta |  -.1758607   .0747755    -2.35   0.019    -.3227093    -.029012
       _cons |   4.310994   1.341992     3.21   0.001     1.675509    6.946479
-------------+----------------------------------------------------------------
     sigma_u |  1.6744399
     sigma_e |  .48804015
         rho |  .92170004   (fraction of variance due to u_i)
------------------------------------------------------------------------------

. areg tobinq env_score de lnta, absorb(company)

Linear regression, absorbing indicators         Number of obs     =      2,634
Absorbed variable: company                      No. of categories =        611
                                                F(   3,   2020)   =       7.73
                                                Prob > F          =     0.0000
                                                R-squared         =     0.9173
                                                Adj R-squared     =     0.8922
                                                Root MSE          =     0.4880

------------------------------------------------------------------------------
      tobinq |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   env_score |   .0046432   .0012264     3.79   0.000     .0022381    .0070484
          de |   3.04e-06   3.78e-06     0.80   0.422    -4.38e-06    .0000104
        lnta |  -.1758607    .043828    -4.01   0.000    -.2618134    -.089908
       _cons |   4.310994   .7840253     5.50   0.000     2.773412    5.848577
------------------------------------------------------------------------------
F test of absorbed indicators: F(610, 2020) = 34.830          Prob > F = 0.000

.