R-squared 0.653838 Mean dependent var
-0.118947 Adjusted R-squared
0.641798 S.D. dependent var 0.898646
S.E. of regression 0.537840 Akaike info criterion
1.638261 Sum squared resid
33.26620 Schwarz criterion 1.754406
Log likelihood -93.29564 Hannan-Quinn criter.
1.685428 F-statistic
54.30370 Durbin-Watson stat 1.091972
ProbF-statistic 0.000000
Sumber: Data Diolah
b. Hasil Estimasi FEM
Tabel. 4.9 Dependent Variable: ROA
Dependent Variable: ROA Method: Panel Least Squares
Date: 053016 Time: 19:34 Sample: 2011Q1 2015Q4
Periods included: 20 Cross-sections included: 6
Total panel balanced observations: 120 Variable
Coefficient Std. Error
t-Statistic Prob.
C 34.22488
4.305576 7.948966
0.0000 CAR
-0.753718 0.320949
-2.348403 0.0206
NPF -0.170614
0.132419 -1.288440
0.2003 FDR
-0.415437 0.702405
-0.591450 0.5554
BOPO -6.762593
0.648100 -10.43449
0.0000 Effects Specification
Cross-section fixed dummy variables R-squared
0.730577 Mean dependent var -0.118947
Adjusted R-squared 0.708533 S.D. dependent var
0.898646 S.E. of regression
0.485158 Akaike info criterion 1.470972
Sum squared resid 25.89164 Schwarz criterion
1.703263 Log likelihood
-78.25834 Hannan-Quinn criter. 1.565307
F-statistic 33.14215 Durbin-Watson stat
1.356573 ProbF-statistic
0.000000
Sumber: Data Diolah
c. Hasil Estimasi REM
Tabel. 4.10 Dependent Variable: ROA
Dependent Variable: ROA Method: Panel EGLS Cross-section random effects
Date: 053016 Time: 19:35 Sample: 2011Q1 2015Q4
Periods included: 20 Cross-sections included: 6
Total panel balanced observations: 120 Swamy and Arora estimator of component variances
Variable Coefficient
Std. Error t-Statistic
Prob. C
33.91741 3.921129
8.649911 0.0000
CAR -0.601522
0.301863 -1.992698
0.0487 NPF
-0.274144 0.127244
-2.154475 0.0333
FDR -0.617903
0.651596 -0.948291
0.3450 BOPO
-6.561755 0.620080
-10.58211 0.0000
Effects Specification S.D.
Rho Cross-section random
0.129179 0.0662
Idiosyncratic random 0.485158
0.9338 Weighted Statistics
R-squared 0.644165 Mean dependent var
-0.076495 Adjusted R-squared
0.631788 S.D. dependent var 0.830373
S.E. of regression 0.503874 Sum squared resid
29.19728 F-statistic
52.04580 Durbin-Watson stat 1.218767
ProbF-statistic 0.000000
Unweighted Statistics R-squared
0.650764 Mean dependent var -0.118947
Sum squared resid 33.56163 Durbin-Watson stat
1.060279
Sumber: Data Diolah
2. Uji Pemilihan Model Regresi Panel
Untuk membandingan model mana yang terbaik antara ke tiga model di atas maka kita dapat mengujinya dengan tiga cara, yaitu: Uji Chow, Uji Hausman
dan Uji Langrange Multiplier LM.
a. Uji Chow
Uji Chow dilakukan untuk membandingkan model mana yang terbaik antara CE dan FE. Pada tabel uji chow,