LAMPIRAN Langkah-Langkah Pemilihan Model Regresi Data Panel
Hasil Common Effect Dependent Variable: LOG(MISKIN) Method: Panel Least Squares Date: 12/06/11 Time: 18:16 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
12.40080 0.145885 -5.157553 -0.027868 0.054077
1.872750 0.114857 0.400902 0.005025 0.053468
6.621707 1.270151 -12.86487 -5.545917 1.011393
0.0000 0.2124 0.0000 0.0000 0.3188
R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)
0.847999 0.830628 0.189102 1.251581 12.53190 48.81555 0.000000
Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat
2.530722 0.459488 -0.376595 -0.165485 -0.300264 0.470897
71
Hasil Fixed Effect Dependent Variable: LOG(MISKIN) Method: Panel Least Squares Date: 12/06/11 Time: 18:16 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
10.98810 -0.489244 -1.040229 -0.004925 0.038281
1.465929 0.139912 0.524818 0.003915 0.025106
7.495657 -3.496797 -1.982073 -1.257805 1.524771
0.0000 0.0016 0.0574 0.2189 0.1385
Effects Specification Cross-section fixed (dummy variables) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)
0.990465 0.986719 0.052953 0.078513 67.90980 264.4081 0.000000
Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat
2.530722 0.459488 -2.795490 -2.288826 -2.612296 1.411928
72
Hasil Random Effect Dependent Variable: LOG(MISKIN) Method: Panel EGLS (Cross-section random effects) Date: 12/06/11 Time: 18:16 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Swamy and Arora estimator of component variances Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
9.498533 -0.203844 -2.052676 -0.005554 0.066210
1.311218 0.115235 0.435154 0.003581 0.023160
7.244055 -1.768952 -4.717122 -1.551038 2.858769
0.0000 0.0856 0.0000 0.1299 0.0071
Effects Specification S.D. Cross-section random Idiosyncratic random
0.233854 0.052953
Rho 0.9512 0.0488
Weighted Statistics R-squared Adjusted R-squared S.E. of regression F-statistic Prob(F-statistic)
0.562943 0.512994 0.062708 11.27028 0.000006
Mean dependent var S.D. dependent var Sum squared resid Durbin-Watson stat
0.254971 0.089857 0.137629 1.069881
Unweighted Statistics R-squared Sum squared resid
0.455705 4.481749
Mean dependent var Durbin-Watson stat
2.530722 0.032855
73
1.
Uji Common Effect dengan Fixed Effect
Signifikansi Model Fixed Effect Ho : α1= α2 =…=αi (intersep sama) H1 : sekurang-kurangnya ada 1 intersep yang berbeda Wilayah kritis : F(0,05; 7;28 ) = 2,36 Statistik Uji F: 59.76427
Keputusan : Tolak H0 Kesimpulan : Model Fixed Effect lebih baik daripada Common Effect pada α = 5 persen, intersep untuk tiap kabupaten/kota tidak sama.
74
2.
Uji Fixed Effect dengan Random Effect (Hausman Test)
H0 : model Random Effect lebih baik daripada Fixed Effect H1 : model Fixed Effect lebih baik daripada Random Effect Correlated Random Effects - Hausman Test Equation: Untitled Test cross-section random effects
Test Summary Cross-section random
Chi-Sq. Statistic
Chi-Sq. d.f.
Prob.
18.082184
4
0.0012
Random
Var(Diff.)
Prob.
-0.203844 -2.052676 -0.005554 0.066210
0.006296 0.086075 0.000003 0.000094
0.0003 0.0006 0.6912 0.0040
Cross-section random effects test comparisons: Variable LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
Fixed -0.489244 -1.040229 -0.004925 0.038281
Cross-section random effects test equation: Dependent Variable: LOG(MISKIN) Method: Panel Least Squares Date: 12/06/11 Time: 18:23 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
10.98810 -0.489244 -1.040229 -0.004925 0.038281
1.465929 0.139912 0.524818 0.003915 0.025106
7.495657 -3.496797 -1.982073 -1.257805 1.524771
0.0000 0.0016 0.0574 0.2189 0.1385
Effects Specification Cross-section fixed (dummy variables) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)
0.990465 0.986719 0.052953 0.078513 67.90980 264.4081 0.000000
Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat
2.530722 0.459488 -2.795490 -2.288826 -2.612296 1.411928
75
Karena nilai probabilitas Chi-Square berdasarkan hasil estimasi diperoleh probabilitas sebesar 0,0012 yang berarti tolak Ho. Kesimpulan : model Fixed Effect lebih baik daripada Random Effect.
3.
Uji Asumsi Homoskedastisitas Untuk medeteksi adanya heteroskedastisitas dapat menggunakan metode
General Least Square (Cross section Weight) yaitu dengan membandingkan sum square Residual pada Weighted Statistics dengan sum square Residual Unweighted Statistics. Jika sum square Residual pada Weighted Statistics lebih kecil
dari
sum
heteroskedastisitas. heteroskedastisitas.
square Hasil
Residual output
Unweighted
Statistics,
memperlihatkan
maka
adanya
terjadi indikasi
76
Dependent Variable: LOG(MISKIN) Method: Panel EGLS (Cross-section weights) Date: 12/06/11 Time: 18:24 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Linear estimation after one-step weighting matrix Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
11.38855 -0.517855 -1.091039 -0.002860 0.027069
1.250168 0.126073 0.430742 0.003038 0.020699
9.109618 -4.107568 -2.532929 -0.941462 1.307705
0.0000 0.0003 0.0172 0.3545 0.2016
Effects Specification Cross-section fixed (dummy variables) Weighted Statistics R-squared Adjusted R-squared S.E. of regression F-statistic Prob(F-statistic)
0.992760 0.989916 0.052219 349.0327 0.000000
Mean dependent var S.D. dependent var Sum squared resid Durbin-Watson stat
3.058640 1.521740 0.076350 1.785111
Unweighted Statistics R-squared Sum squared resid
0.990284 0.080003
Mean dependent var Durbin-Watson stat
2.530722 1.332777
77
Treatment pelanggaran ini dapat dilakukan dengan mengestimasi GLS dengan white-heteroscedasticity. Berdasarkan prosedur di atas, maka hasil estimasi Fixed Effect sebagai berikut: Dependent Variable: LOG(MISKIN) Method: Panel EGLS (Cross-section weights) Date: 12/06/11 Time: 18:25 Sample: 2005 2009 Periods included: 5 Cross-sections included: 8 Total panel (balanced) observations: 40 Linear estimation after one-step weighting matrix White cross-section standard errors & covariance (d.f. corrected) Variable
Coefficient
Std. Error
t-Statistic
Prob.
C LOG(PDRB) LOG(MYS) SHARE_PERTANIAN LOG(PENGANGGURAN)
11.38855 -0.517855 -1.091039 -0.002860 0.027069
1.545020 0.089410 0.274328 0.001622 0.007465
7.371135 -5.791907 -3.977127 -1.763247 3.626274
0.0000 0.0000 0.0004 0.0888 0.0011
Effects Specification Cross-section fixed (dummy variables) Weighted Statistics R-squared Adjusted R-squared S.E. of regression F-statistic Prob(F-statistic)
0.992760 0.989916 0.052219 349.0327 0.000000
Mean dependent var S.D. dependent var Sum squared resid Durbin-Watson stat
3.058640 1.521740 0.076350 1.785111
Unweighted Statistics R-squared Sum squared resid
0.990284 0.080003
Mean dependent var Durbin-Watson stat
2.530722 1.332777
4. Uji Asumsi Autokolerasi Hasil estimasi menunjukkan nilai statistic Durbin Watson sebesar 1,785111. Nilai Durbin Watson tersebut berada pada interval dU < DW < 4-dU (1.721 < 1,785111 < 2.279). Hal ini menunjukkan tidak adanya autokolerasi.
78
5. Uji Multikolinieritas Berdasarkan matriks korelasi pearson antar variabel independen terlihat bahwa korelasi antarvariabel kurang dari 0,8, sehingga dapat disimpulkan model telah memenuhi asumsi terbebas dari multikolinieritas. Variabel PDRB MYS Share_Pertanian Pengangguran
PDRB
MYS
1 0,347285 -0,580666 0,612862
1 -0,749964 0,590386
Share_Pertanian
Pengangguran
1 -0,698824
1
6. Uji Normalitas. Hasil uji normalitas pada Eviews 6.0 sebagai berikut: 6
Series: Standardized Residuals Sample 2005 2009 Observations 40
5
4
3
2
1
Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis
9.02e-18 0.002639 0.090148 -0.085441 0.044246 0.090564 2.305704
Jarque-Bera Probability
0.858091 0.651130
0 -0.05
0.00
0.05
0.10
Berdasarkan nilai probabilitas Jarque-Bera yang lebih besar dari taraf nyata lima persen, dapat disimpulkan bahwa error term terdistribusi dengan normal.
