Lampiran 1 Hasil Estimasi Pengaruh Infrastruktur terhadap Pertumbuhan untuk Model di Jawa dengan Program STATA SE 10 # Mendifinisikan data dalam format panel . xtset prop tahun, yearly panel variable: prop (strongly balanced) time variable: tahun, 1993 to 2009 delta: 1 year # Penghitungan dalam model FEM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, fe Fixed-effects (within) regression Number of obs = Group variable: prop Number of groups = R-sq:
85 5
within = 0.8488 between = 0.0088 overall = 0.0028
Obs per group: min = 17 avg = 17.0 max = 17 F(6,74) = 69.23 corr(u_i, Xb) = -0.5441 Prob > F = 0.0000 ------------------------------------------------------------------------ln_pdrb | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik| .2973954 .0689863 4.31 0.000 .1599373 .4348536 ln_ab | .0181748 .0269611 0.67 0.502 -.0355463 .071896 ln_jln | -.1277735 .0246865 -5.18 0.000 -.1769625 -.0785845 ln_pskesms| .0035885 .0006412 5.60 0.000 .0023109 .004866 ln_tk | -.3602829 .2084909 -1.73 0.088 -.7757102 .0551443 dd | -.0000348 .0263029 -0.00 0.999 -.0524445 .052375 _cons | 24.53426 3.588197 6.84 0.000 17.38462 31.6839 -------------+----------------------------------------------------------sigma_u | 1.3455315 sigma_e | .06810587 rho | .99744453 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(4, 74) = 36.47 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, re Random-effects GLS regression Group variable: prop
Number of obs Number of groups
= =
85 5
R-sq:
Obs per group: min = avg = max =
17 17.0 17
within = 0.7112 between = 0.9992 overall = 0.9883
Random effects u_i ~ Gaussian corr(u_i, X) = 0 (assumed)
Wald chi2(6) Prob > chi2
= =
6594.51 0.0000
ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik| .4550678 .0502045 9.06 0.000 .3566688 .5534669 ln_ab | .1679195 .0338524 4.96 0.000 .1015699 .234269 ln_jln |-.0195267 .028054 -0.70 0.486 -.0745115 .0354582 ln_pskesms| .0064039 .0008093 7.91 0.000 .0048176 .0079901 ln_tk | .8616967 .0205902 41.85 0.000 .8213406 .9020528 dd |-.1509627 .033594 -4.49 0.000 -.2168057 -.0851197 _cons | 6.377468 .4828495 13.21 0.000 5.431101 7.323836 -------------+----------------------------------------------------------sigma_u | 0 sigma_e | .06810587 rho | 0 (fraction of variance due to u_i) -------------------------------------------------------------------------
. est sto random # Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) | fixed random Difference S.E. -------------+----------------------------------------------------------ln_listrik | .2973954 .4550678 -.1576724 .047314 ln_ab | .0181748 .1679195 -.1497446 . ln_jln | -.1277735 -.0195267 -.1082468 . ln_puskesmas | .0035885 .0064039 -.0028154 . ln_tk | -.3602829 .8616967 -1.22198 .2074717 dd | -.0000348 -.1509627 .1509279 . ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from B = inconsistent under Ha, efficient under Ho; obtained from Test:
Ho:
difference in coefficients not systematic chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 104.97 Prob>chi2 = 0.0000
Uji Hausman menunjukkan tolak H0, berarti model FEM yang paling sesuai # Penghitungan dalam Uji Woolridge untuk Asumsi Tidak Ada Autokorelasi Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation F( 1, 4) = 60.803 Prob > F = 0.0015
Uji Woolridge menunjukkan tolak H0, berarti terdapat autokorelasi dalam model terpilih # Penghitungan Uji Wald untuk Asumsi Homoskedastisitas . xttest3 Modified Wald test for groupwise heteroskedasticity in fixed effect regression model H0: sigma(i)^2 = sigma^2 for all i chi2 (5) = 51.84 Prob>chi2 = 0.0000
Modified Wald Test menunjukkan tolak H0, berarti terdapat masalah heteroskedastisitas dalam model terpilih. # Model terpilih dengan koreksi terhadap permasalahan heteroskedastisitas contemporaneously correlated across panel, and first order autokorelasi (ar1) . xtpcse ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, corr (ar1) Prais-Winsten regression, correlated panels corrected standard errors (PCSEs) Group variable: Time variable: Panels: Autocorrelation:
prop tahun correlated (balanced) common AR(1)
Estimated covariances = Estimated autocorrelations = Estimated coefficients =
15 1 7
Number of obs Number of groups Obs per group: min avg max R-squared Wald chi2(6) Prob > chi2
= = = = = = = =
85 5 17 17 17 0.9971 7560.99 0.0000
------------------------------------------------------------------------| Panel-corrected ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik | .5530455 .0662619 8.35 0.000 .4231746 .6829163 ln_ab | .1351584 .0402404 3.36 0.001 .0562888 .2140281 ln_jln | .0214779 .0354445 0.61 0.545 -.047992 .0909479
ln_pskesmas| .0045787 .0011317 4.05 0.000 .0023607 .0067967 ln_tk | .8807358 .0246133 35.78 0.000 .8324946 .928977 dd |-.1208316 .0530424 -2.28 0.023 -.2247927 -.0168705 cons | 7.235929 .5941238 12.18 0.000 6.071468 8.40039 -------------+----------------------------------------------------------rho | .5400293 -------------------------------------------------------------------------
Lampiran 2 Hasil Estimasi Pengaruh Pertumbuhan terhadap Kemiskinan untuk Model di Jawa dengan Program STATA SE 10 # Penghitungan dalam model FEM xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh, fe Fixed-effects (within) regression Number of obs Group variable: prop Number of groups R-sq:
= =
85 5
within = 0.3199 between = 0.0154 overall = 0.0238
Obs per group: min = 17 avg = 17.0 max = 17 F(3,77) = 12.07 corr(u_i, Xb) = -0.7973 Prob > F = 0.0000 ------------------------------------------------------------------------ln_miskin | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb | .5214539 .2396445 2.18 0.033 .0442608 .998647 ln_pengang~n| .2513296 .1116214 2.25 0.027 .0290629 .4735963 ln_ratarat~h|-.8307567 .6754577 -1.23 0.222 -2.175765 .5142515 _cons | 2.63936 3.189329 0.83 0.410 -3.711406 8.990126 -------------+----------------------------------------------------------sigma_u | .98905095 sigma_e | .21901915 rho | .95325492 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(4, 77) = 82.21 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh, re Random-effects GLS regression Number of obs = 85 Group variable: prop Number of groups = 5 R-sq: within = 0.2710 Obs per group: min = 17 between = 0.0841 avg = 17.0 overall = 0.1157 max = 17 Random effects u_i ~ Gaussian Wald chi2(3) = 26.48 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 ------------------------------------------------------------------------ln_miskin | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+--------------------------------------------------------ln_pdrb |-.0167787 .1641919 -0.10 0.919 -.338589 .3050315 ln_pengang| .2261664 .1029963 2.20 0.028 .0242975 .4280354 ln_ratarat| .4017806 .4166726 0.96 0.335 -.4148827 1.218444 _cons | 9.889398 2.02754 4.88 0.000 5.915493 13.8633 -------------+----------------------------------------------------------sigma_u | .29101195 sigma_e | .21901915 rho | .63839628 (fraction of variance due to u_i) ------------------------------------------------------------------------. est sto random # Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) | fixed random Difference S.E. -------------+----------------------------------------------------------ln_pdrb | .5214539 -.0167787 .5382326 .174558 ln_pengang~n | .2513296 .2261664 .0251632 .0430246 ln_ratarat~h | -.8307567 .4017806 -1.232537 .5316268 ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from B = inconsistent under Ha, efficient under Ho; obtained from Test: Ho: difference in coefficients not systematic chi2(3) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= Prob>chi2 =
9.45 0.0239
Uji Hausman menunjukkan tolak H0, berarti model FEM yang paling sesuai # Penghitungan dalam Uji Woolridge untuk Asumsi Tidak Ada Autokorelasi . xtserial ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh, output Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation F( 1, 4) = 44.045 Prob > F = 0.0027
Uji Woolridge menunjukkan tolak H0, berarti terdapat autokorelasi dalam model terpilih # Penghitungan Uji Wald untuk Asumsi Homoskedastisitas . xttest3 Modified Wald test for groupwise heteroskedasticity in fixed effect regression model H0: sigma(i)^2 = sigma^2 for all i
chi2 (5) = Prob>chi2 =
1.98 0.8524
Modified Wald Test menunjukkan terima H0, berarti tidak terdapat masalah heteroskedastisitas dalam model terpilih. # Model terpilih dengan koreksi terhadap permasalahan autokorelasi . xtgls ln_miskin ln_pdrbpre ln_pengangguran ln_rataratasklh, corr (ar1) Cross-sectional time-series FGLS regression Coefficients: generalized least squares Panels: homoskedastic Correlation: common AR(1) coefficient for all panels (0.6545) Estimated covariances = Estimated autocorrelations = Estimated coefficients =
1 1 4
Number of obs Number of groups Time periods Wald chi2(3) Prob > chi2
= = = = =
85 5 17 52.12 0.0000
------------------------------------------------------------------------ln_miskin | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb |-.7070452 .1417731 -4.99 0.000 -.9849153 -.4291751 ln_pengang| .7549185 .115229 6.55 0.000 .5290739 .9807631 ln_ratat~h| 1.421027 .3754204 3.79 0.000 .6852164 2.156837 _cons | 12.86544 1.425478 9.03 0.000 10.07156 15.65933 -------------------------------------------------------------------------
Lampiran 3 Hasil Estimasi Pengaruh Infrastruktur terhadap Pertumbuhan untuk Model di Luar Jawa dengan Program STATA SE 10 # Penghitungan dalam model FEM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, fe Fixed-effects (within) regression Number of obs = Group variable: prov Number of groups = R-sq:
within = 0.7008 between = 0.6493 overall = 0.6428
corr(u_i, Xb)
= 0.3843
Obs per group: min avg max F(6,330) Prob > F
= = = = =
357 21 17 17.0 17 128.84 0.0000
------------------------------------------------------------------------ln_pdrb | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik | .235594 .0405048 5.82 0.000 .1559139 .3152742 ln_ab |.0129663 .0218751 0.59 0.554 -.030066 .0559986 ln_jln | .020831 .0147626 1.41 0.159 -.0082097 .0498717
ln_pusksms |.0048843 .0016848 2.90 0.004 .0015701 .0081986 ln_tk |.6618725 .1038033 6.38 0.000 .4576729 .8660722 dd |.0555075 .0208609 2.66 0.008 .0144703 .0965447 _cons |8.297414 1.692421 4.90 0.000 4.96812 11.62671 -------------+----------------------------------------------------------sigma_u | .52056867 sigma_e | .11644974 rho | .95234428 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(20, 330) = 212.03 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, re Random-effects GLS regression Number of obs = 357 Group variable: prov Number of groups = 21 R-sq: within = 0.7005 Obs per group: min = 17 between = 0.6522 avg = 17.0 overall = 0.6466 max = 17 Random effects u_i ~ Gaussian Wald chi2(6) = 791.17 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 ------------------------------------------------------------------------ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik|.2500969 .0394618 6.34 0.000 .1727532 .3274405 ln_ab |.0142935 .0221207 0.65 0.518 -.0290623 .0576494 ln_jln |.0249174 .0148838 1.67 0.094 -.0042543 .0540891 ln_psksmas| .005243 .0016986 3.09 0.002 .0019139 .0085721 ln_tk |.6834638 .092101 7.42 0.000 .5029491 .8639785 dd |.0472341 .0210585 2.24 0.025 .0059601 .0885081 _cons |8.131539 1.5173 5.36 0.000 5.157686 11.10539 -------------+----------------------------------------------------------sigma_u | .41165812 sigma_e | .11644974 rho | .92590797 (fraction of variance due to u_i) ------------------------------------------------------------------------. est sto random # Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) | fixed random Difference S.E. -------------+--------------------------------------------------------------ln_listrik | .235594 .2500969 -.0145028 .0091328 ln_ab | .0129663 .0142935 -.0013273 . ln_jln | .020831 .0249174 -.0040863 . ln_puskesmas | .0048843 .005243 -.0003587 . ln_tk | .6618725 .6834638 -.0215913 .0478803 dd | .0555075 .0472341 .0082734 . ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from B = inconsistent under Ha, efficient under Ho; obtained from Test:
Ho:
difference in coefficients not systematic chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 75.95 Prob>chi2 = 0.0000
Uji Hausman menunjukkan tolak H0, berarti model FEM yang paling sesuai # Penghitungan dalam Uji Woolridge untuk Asumsi Tidak Ada Autokorelasi . xtserial ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, output Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation F( 1, 20) = 129.062 Prob > F = 0.0000
Uji Woolridge menunjukkan tolak H0, berarti terdapat autokorelasi dalam model terpilih
# Penghitungan Uji Wald untuk Asumsi Homoskedastisitas . xttest3 Modified Wald test for groupwise heteroskedasticity in fixed effect regression model H0: sigma(i)^2 = sigma^2 for all i chi2 (21) = Prob>chi2 =
1698.34 0.0000
Modified Wald Test menunjukkan tolak H0, berarti terdapat masalah heteroskedastisitas dalam model terpilih. # Model terpilih dengan koreksi terhadap permasalahan heteroskedastisitas contemporaneously correlated across panel, and first order autokorelasi (ar1) . xtpcse ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd, corr (ar1) (note: estimates of rho outside [-1,1] bounded to be in the range [-1,1]) Prais-Winsten regression, correlated panels corrected standard errors (PCSEs) Group variable: Time variable: Panels: Autocorrelation:
prov tahun correlated (balanced) common AR(1)
Estimated covariances = Estimated autocorrelations = Estimated coefficients =
231 1 7
Number of obs Number of groups Obs per group: min avg max R-squared Wald chi2(6) Prob > chi2
= = = = = = = =
357 21 17 17 17 0.9922 524.81 0.0000
------------------------------------------------------------------------| Panel-corrected ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik| .3374637 .0555478 6.08 0.000 .228592 .4463354 ln_ab | .0415669 .0271268 1.53 0.125 -.0116007 .0947346 ln_jln | .0043931 .0232843 0.19 0.850 -.0412434 .0500295 ln_pusksms| .00 5663 .0019049 2.97 0.003 .0019294 .0093966 ln_tk | .7499097 .0446191 16.81 0.000 .6624579 .8373614 dd |-.0051537 .028895 -0.18 0.858 -.0617868 .0514795 _cons | 7.80029 .8294816 9.40 0.000 6.174536 9.426044 -------------+----------------------------------------------------------rho | .8238999 -------------------------------------------------------------------------
Lampiran 4 Hasil Estimasi Pengaruh Pertumbuhan terhadap Kemiskinan untuk Model di Luar Jawa dengan Program STATA SE 10 # Penghitungan dalam model FEM . xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh, fe Fixed-effects (within) regression Group variable: prov
Number of obs Number of groups
= =
356 21
R-sq:
Obs per group: min = avg = max =
16 17.0 17
within = 0.2365 between = 0.0209 overall = 0.0316
corr(u_i, Xb)
= -0.1349
F(3,332) Prob > F
= =
34.27 0.0000
------------------------------------------------------------------------ln_miskin | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb | .3397415 .1184836 2.87 0.004 .1066683 .5728147 ln_pengangn| .1635478 .0388328 4.21 0.000 .0871584 .2399373 ln_ratarata|-.4222983 .3155182 -1.34 0.182 -1.042965 .1983686 _cons | 6.328538 1.364 4.64 0.000 3.645365 9.011711
-------------+----------------------------------------------------------sigma_u | .93059962 sigma_e | .21877232 rho | .94762826 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(20, 332) = 295.52 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh, re Random-effects GLS regression Number of obs = Group variable: prov Number of groups = R-sq:
within = 0.2365 between = 0.0208 overall = 0.0315
Random effects u_i ~ Gaussian corr(u_i, X) = 0 (assumed)
Obs per group: min = avg = max = Wald chi2(3) Prob > chi2
= =
356 21 16 17.0 17 103.26 0.0000
------------------------------------------------------------------------ln_miskin | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb |.3349692 .1144573 2.93 0.003 .110637 .5593015 ln_pengang~n|.1629681 .0386328 4.22 0.000 .0872491 .2386871 ln_ratarat~h|-.4179115 .3054819 -1.37 0.171 -1.016645 .180822 _cons |6.401366 1.336088 4.79 0.000 3.78268 9.020051 -------------+----------------------------------------------------------sigma_u | .96635108 sigma_e | .21877232 rho | .95124627 (fraction of variance due to u_i) ------------------------------------------------------------------------. est sto random
# Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) | fixe rando Difference S.E. -------------+----------------------------------------------------------pdrbpre | .3397415 .3349692 .0047723 .0306248 ln_pengang~n | .1635478 .1629681 .0005797 .0039361 ln_ratarat~h | -.4222983 -.4179115 -.0043868 .0789465 ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from B = inconsistent under Ha, efficient under Ho; obtained from Test:
Ho:
difference in coefficients not systematic chi2(3) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 0.60 Prob>chi2 = 0.8968
Uji Hausman menunjukkan tidak cukup bukti untuk menolak H0, berarti model REM yang paling sesuai
Lampiran 5 Hasil Estimasi Pengaruh Infrastruktur terhadap Pertumbuhan untuk Model Gabungan dengan Program STATA SE 10 # Penghitungan dalam model FEM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd dw_ab, fe Fixed-effects (within) regression Number of obs = 442 Group variable: prop Number of groups = 26 R-sq:
within = 0.6846 between = 0.7982 overall = 0.7943
Obs per group: min = avg = max =
17 17.0 17
F(7,409) = 126.83 corr(u_i, Xb) = 0.4407 Prob > F = 0.0000 ------------------------------------------------------------------------ln_pdrb | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik | .2367901 .035486 6.67 0.000 .1670324 .3065478 ln_ab |-.0174076 .0302395 -0.58 0.565 -.0768518 .0420366 ln_jln | .0055385 .0120538 0.46 0.646 -.0181567 .0292338 ln_pskesmas | .0016628 .0006531 2.55 0.011 .0003789 .0029468 ln_tk | .7014113 .0954085 7.35 0.000 .513859 .8889636 dd | .0305829 .0186421 1.64 0.102 -.0060635 .0672293 dw_ab | .0006064 .032123 0.02 0.985 -.0625404 .0637531 _cons | 7.916411 1.602358 4.94 0.000 4.766527 11.06629 -------------+----------------------------------------------------------sigma_u | .68413574 sigma_e | .11456928 rho | .97272029 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(25, 409) = 202.18 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd dw_ab, re Random-effects GLS regression Number of obs = 442 Group variable: prop Number of groups = 26 R-sq: within = 0.6805 Obs per group: min = 17 between = 0.8137 avg = 17.0 overall = 0.8106 max = 17 Random effects u_i ~ Gaussian Wald chi2(7) = 1047.82 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 ------------------------------------------------------------------------ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik | .190894 .032126 5.94 0.000 .1279282 .2538598 ln_ab | .02188 .028114 0.78 0.436 -.0332225 .0769824 ln_jln | .0048612 .0123382 0.39 0.694 -.0193212 .0290437 ln_pskesmas| .0021692 .000675 3.21 0.001 .0008462 .0034921 ln_tk | .8779248 .0579636 15.15 0.000 .7643183 .9915313 dd | .0313222 .0192732 1.63 0.104 -.0064525 .0690969 dw_ab | -.020199 .0286626 -0.70 0.481 -.0763766 .0359787 _cons | 4.833236 1.002299 4.82 0.000 2.868766 6.797706 -------------+----------------------------------------------------------sigma_u | .36242941 sigma_e | .11456928 rho | .90914998 (fraction of variance due to u_i) ------------------------------------------------------------------------. est sto random # Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) sqrt(diag(V_b-V_B)) | random fixed Difference S.E. -------------+---------------------------------------------------------------
ln_listrik | .190894 .2367901 -.0458961 . ln_ab | .02188 -.0174076 .0392876 . ln_jln |.0048612 .0055385 -.0006773 .0026337 ln_puskesmas |.0021692 .0016628 .0005063 .0001702 ln_tk |.8779248 .7014113 .1765135 . dd |.0313222 .0305829 .0007393 .0048913 dw_ab |-.020199 .0006064 -.0208053 . ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg Test:
Ho:
difference in coefficients not systematic chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 19.90 Prob>chi2 = 0.0058
Uji Hausman menunjukkan tolak H0, berarti model FEM yang paling sesuai # Penghitungan dalam Uji Woolridge untuk Asumsi Tidak Ada Autokorelasi Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation F( 1, 25) = 89.195 Prob > F = 0.0000
Uji Woolridge menunjukkan tolak H0, berarti terdapat autokorelasi dalam model terpilih # Penghitungan Uji Wald untuk Asumsi Homoskedastisitas . xttest3 Modified Wald test for groupwise heteroskedasticity in fixed effect regression model H0: sigma(i)^2 = sigma^2 for all i chi2 (26) = Prob>chi2 =
2074.99 0.0000
# Model terpilih dengan koreksi terhadap permasalahan heteroskedastisitas contemporaneously correlated across panel, and first order autokorelasi (ar1) . xtpcse ln_pdrb ln_listrik ln_ab ln_jln ln_puskesmas ln_tk dd dw_ab, corr (ar1) (note: estimates of rho outside [-1,1] bounded to be in the range [-1,1]) Prais-Winsten regression, correlated panels corrected standard errors (PCSEs) Group variable: Time variable: Panels: Autocorrelation:
prop tahun correlated (balanced) common AR(1)
Estimated covariances = Estimated autocorrelations = Estimated coefficients =
351 1 8
Number of obs Number of groups Obs per group: min avg max R-squared Wald chi2(7) Prob > chi2
= = = = = = = =
442 26 17 17 17 0.9885 2840.60 0.0000
------------------------------------------------------------------------| Panel-corrected ln_pdrb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_listrik | .0576344 .0274393 2.10 0.036 .0038544 .1114144 ln_ab | .1861146 .0464779 4.00 0.000 .0950196 .2772095 ln_jln | -.09368 .0215221 -4.35 0.000 -.1358625 -.0514975 ln_puskesmas| .0077231 .0013853 5.58 0.000 .005008 .0104383 ln_tk | .8150744 .031613 25.78 0.000 .7531141 .8770347 dd | .0048195 .0405262 0.12 0.905 -.0746105 .0842494 dw_ab |-.0183591 .0106673 -1.72 0.085 -.0392665 .0025484 _cons | 3.536337 .530296 6.67 0.000 2.496976 4.575698 -------------+----------------------------------------------------------rho | .7259617 -------------------------------------------------------------------------
Lampiran 6 Hasil Estimasi Pengaruh Pertumbuhan terhadap Kemiskinan untuk Model Gabungan dengan Program STATA SE 10 # Penghitungan dalam model FEM . xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh dw_pdrb dw_pengangguran dw_rataratasklh, fe Fixed-effects (within) regression Group variable: prop
Number of obs Number of groups
= =
441 26
R-sq:
Obs per group: min = avg = max =
16 17.0 17
within = 0.1894 between = 0.3403 overall = 0.3269
F(6,409) = 15.93 corr(u_i, Xb) = 0.0597 Prob > F = 0.0000 ------------------------------------------------------------------------ln_miskin | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb | .2097846 .2799074 0.75 0.454 -.340452 .7600211 ln_pengang| .4251183 .1305423 3.26 0.001 .1685007 .681736 ln_ratarat|-1.473853 .9203353 -1.60 0.110 -3.283031 .3353244 dw_pdrb |-.0011338 .3038911 -0.00 0.997 -.5985171 .5962496 dw_pengang|-.2459539 .1379898 -1.78 0.075 -.5172116 .0253039 dw_ratarat| 1.456911 .9818972 1.48 0.139 -.4732839 3.387106 _cons | 7.020798 1.425687 4.92 0.000 4.21821 9.823386 -------------+----------------------------------------------------------sigma_u | .74485943 sigma_e | .25977326 rho | .89155971 (fraction of variance due to u_i) ------------------------------------------------------------------------F test that all u_i=0: F(25, 409) = 84.17 Prob > F = 0.0000 . est sto fixed # Penghitungan dalam model REM . xtreg ln_miskin ln_pdrb ln_pengangguran ln_rataratasklh dw_pdrb dw_pengangguran dw_rataratasklh, re Random-effects GLS regression Number of obs = 441 Group variable: prop Number of groups = 26 R-sq: within = 0.1840 Obs per group: min = 16 between = 0.5081 avg = 17.0 overall = 0.4788 max = 17 Random effects u_i ~ Gaussian Wald chi2(6) = 116.35 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 ------------------------------------------------------------------------ln_miskin | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------ln_pdrb | .3982093 .0944052 4.22 0.000 .2131785 .5832401 ln_pengang~| .3043973 .1066645 2.85 0.004 .0953386 .5134559 ln_ratarat~|-1.371696 .5980721 -2.29 0.022 -2.543896 -.1994962 dw_pdrb |-.0559864 .069092 -0.81 0.418 -.1914042 .0794315 dw_pengang |-.1318088 .1154211 -1.14 0.253 -.3580301 .0944125 dw_ratarat~| 1.18289 .6518725 1.81 0.070 -.0947566 2.460537 _cons | 5.289496 1.035304 5.11 0.000 3.260337 7.318655 -------------+----------------------------------------------------------sigma_u | .65819551 sigma_e | .25977326 rho | .86522533 (fraction of variance due to u_i) -----------------------------------------------------------------------. est sto random # Penghitungan dalam Uji Hausman . hausman fixed random ---- Coefficients ---| (b) (B) (b-B) sqrt(diag(V_b-V_B)) | fixed random Difference S.E. -------------+----------------------------------------------------------ln_pdrbpre | .2097846 .3982093 -.1884247 .2635067
ln_pengang~n | .4251183 .3043973 .1207211 .0752594 ln_ratarat~h | -1.473853 -1.371696 -.1021572 .6995189 dw_pdrb | -.0011338 -.0559864 .0548526 .2959326 dw_pengang~n | -.2459539 -.1318088 -.114145 .0756251 dw_ratarat~h | 1.456911 1.18289 .2740209 .7342917 ------------------------------------------------------------------------b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg Test:
Ho:
difference in coefficients not systematic chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 7.79 Prob>chi2 = 0.2542
Uji Hausman menunjukkan tidak cukup bukti untuk menolak H0, berarti model REM yang paling sesuai .
