I recently found we are able to run fixed effects to the ordered logit model. However, I don't know why the number of firms and obs in the fixed effects ordered logit model decreases compared with the random effects (default) model.
Random effects model: 206,445 obs, 25,957 firms
Code:
. ologit DebtMat_D KS_oi_5_w01 $CONTROLSF LEV_w01 $CONTROLSC, cluster(gvkey_n)
Iteration 0: log pseudolikelihood = -140510.67
Iteration 1: log pseudolikelihood = -113303.82
Iteration 2: log pseudolikelihood = -105523.71
Iteration 3: log pseudolikelihood = -104558.41
Iteration 4: log pseudolikelihood = -104543.3
Iteration 5: log pseudolikelihood = -104543.28
Iteration 6: log pseudolikelihood = -104543.28
Ordered logistic regression Number of obs = 206,445
Wald chi2(16) = 6620.20
Prob > chi2 = 0.0000
Log pseudolikelihood = -104543.28 Pseudo R2 = 0.2560
(Std. Err. adjusted for 25,957 clusters in gvkey_n)
------------------------------------------------------------------------------
| Robust
DebtMat_D | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
KS_oi_5_w01 | .0152759 .0009607 15.90 0.000 .0133929 .0171588
MB2_w01 | -.0993666 .0067298 -14.77 0.000 -.1125568 -.0861764
amat_w01 | .0114257 .0005626 20.31 0.000 .0103231 .0125283
fsize_w01 | -.2787021 .0088328 -31.55 0.000 -.2960141 -.2613901
abne_w01 | -.270203 .0458447 -5.89 0.000 -.360057 -.180349
tang_w01 | -.877234 .045055 -19.47 0.000 -.9655401 -.7889279
prof_w01 | .1890342 .0716216 2.64 0.008 .0486584 .32941
xrd_at_w01 | .0313527 .2107822 0.15 0.882 -.3817729 .4444783
LEV_w01 | -8.469879 .1944697 -43.55 0.000 -8.851033 -8.088726
g_gdpcap | .0346774 .0033836 10.25 0.000 .0280457 .041309
Inflation | -.0001904 .000247 -0.77 0.441 -.0006745 .0002936
lawandorder | -.0764271 .0173026 -4.42 0.000 -.1103395 -.0425146
corruption | -.1995165 .0137202 -14.54 0.000 -.2264077 -.1726254
stmktcap | -.002578 .0003136 -8.22 0.000 -.0031927 -.0019634
prbond | .0054808 .0005011 10.94 0.000 .0044986 .006463
dbagdp | .0037042 .0003146 11.77 0.000 .0030875 .0043208
-------------+----------------------------------------------------------------
/cut1 | -2.318402 .0971453 -2.508803 -2.128
/cut2 | -1.263156 .095973 -1.45126 -1.075052
------------------------------------------------------------------------------
.Code:
. feologit DebtMat_D KS_oi_5_w01 $CONTROLSF LEV_w01 $CONTROLSC, group(gvkey_n) cluster(gvke
> y_n)
note: multiple positive outcomes within groups encountered.
Iteration 0: log conditional likelihood = -43257.074
Iteration 1: log conditional likelihood = -41739.535
Iteration 2: log conditional likelihood = -41724.439
Iteration 3: log conditional likelihood = -41724.433
Iteration 4: log conditional likelihood = -41724.433
Fixed-effects ordered logistic regression
N. of obs. (inc. copies) = 128749
N. of observations = 85518
N. of panel units = 8666
Wald chi2(16) = 1951.93
Prob > chi2 = 0.0000
Log conditional likelihood = -41724.433 Pseudo R2 = 0.1958
(Std. Err. adjusted for 8,666 clusters in gvkey_n)
------------------------------------------------------------------------------
| Robust
DebtMat_D | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
KS_oi_5_w01 | .0043495 .0016068 2.71 0.007 .0012003 .0074987
MB2_w01 | -.0787541 .0096785 -8.14 0.000 -.0977236 -.0597845
amat_w01 | .0061931 .0009245 6.70 0.000 .004381 .0080052
fsize_w01 | -.5205221 .0339207 -15.35 0.000 -.5870054 -.4540388
abne_w01 | -.1495407 .0490377 -3.05 0.002 -.2456529 -.0534286
tang_w01 | -.8534134 .0999987 -8.53 0.000 -1.049407 -.6574196
prof_w01 | .0631142 .1064506 0.59 0.553 -.1455252 .2717535
xrd_at_w01 | -.9581672 .3847432 -2.49 0.013 -1.71225 -.2040845
LEV_w01 | -10.41296 .2942963 -35.38 0.000 -10.98977 -9.836153
g_gdpcap | .0104853 .0050467 2.08 0.038 .000594 .0203766
Inflation | -.0009479 .0013757 -0.69 0.491 -.0036441 .0017483
lawandorder | -.3423384 .0437399 -7.83 0.000 -.428067 -.2566099
corruption | -.135947 .0314069 -4.33 0.000 -.1975034 -.0743907
stmktcap | -.0011825 .0007712 -1.53 0.125 -.0026941 .000329
prbond | .0167317 .0017496 9.56 0.000 .0133026 .0201608
dbagdp | .0018411 .0011145 1.65 0.099 -.0003433 .0040255
------------------------------------------------------------------------------Huyen
0 Response to Number of firms and observations decreases in fixed effects ordered logit model
Post a Comment