Analisis Regresi Data Panel Menggunakan Gretl

Masukkan data: File → Open Data → impor Excel →

Pilih nama file → open → OK → Yes

Pilih Panel → Forwad → Stacked time series → Forward → Forward → OK

Klik Menu model → pilih ordinary least square

Pada dependen variable choose NP
Pada independent variable: add UK, KIND_, K_AUD
Klik OK

Maka akan muncul hasil:

Model 1: Pooled OLS estimates using 25 observations

Included 5 cross-sectional units

Time-series length = 5

Dependent variable: NP

     VARIABLE       COEFFICIENT       STDERROR     T STAT   P-VALUE

const               -6.13465         3.37871     -1.816   0.08373 *

UK                   0.0946867       0.184124     0.514   0.61244

K_IND               12.9318           3.88385       3.330   0.00318 ***

K_AUD                 2.80815          2.21417       1.268   0.21858

Mean of dependent variable = 1.5736

Standard deviation of dep. var. = 1.48048

Sum of squared residuals = 34.3047

Standard error of residuals = 1.27811

Unadjusted R-squared = 0.347868

Adjusted R-squared = 0.254706

F-statistic (3, 21) = 3.73402 (p-value = 0.027)

Log-likelihood = -39.4286

Akaike information criterion (AIC) = 86.8571

Schwarz Bayesian criterion (BIC) = 91.7326

Hannan-Quinn criterion (HQC) = 88.2094

Selanjutkan akan dilakukan running model sekaligus uji terhadap model fixed dan random effect sekaligus (ini adalah kelebihan dari software gretl)

Klik panel tets-->diagnotis test:

model fix effect

Diagnostics: assuming a balanced panel with 5 cross-sectional units

                        observed over 5 periods

Fixed effects estimator

allows for differing intercepts by cross-sectional unit

slope standard errors in parentheses, p-values in brackets

             UK:       0.74465       (1.5177)       [0.62996]

           K_IND:        2.0963       (16.377)       [0.89965]

           K_AUD:         1.8778       (4.1499)       [0.65663]

Means of pooled OLS residuals for cross-sectional units:

unit 1:       0.61947

unit 2:       0.16288

unit 3:     -0.16007

unit 4:     -0.47706

unit 5:     -0.14523

5 group means were subtracted from the data

Berikut hasil Uji Chow Test (Pooled vs Fixed)

Residual variance: 24.4465/(25 - 8) = 1.43803

Joint significance of differing group means:

F(4, 17) = 1.71386 with p-value 0.193295

(A low p-value counts against the null hypothesis that the pooled OLS model

is adequate, in favor of the fixed effects alternative.)

Terlihat terima H₀ berti model terpilih adalah common/pooled effect model

Model Random Effect

                        Random effects estimator

           allows for a unit-specific component to the error term

           (standard errors in parentheses, p-values in brackets)

           const:       -5.9405       (3.4921)       [0.10368]

             UK:       0.09607       (0.19122)       [0.62061]

           K_IND:         12.548       (4.0224)       [0.00518]

           K_AUD:         2.7418       (2.2893)       [0.24439]

Berikut hasil LM Test (Pooled vs Random)

Breusch-Pagan test statistic:

LM = 0.784705 with p-value = prob(chi-square(1) > 0.784705) = 0.375706

(A low p-value counts against the null hypothesis that the pooled OLS model

is adequate, in favor of the random effects alternative.)

Variance estimators:

between = 0.320402

within = 1.43803

theta used for quasi-demeaning = 0.0525617

Sehingga disimpulkan model pooled/common yang terpilih

Berikut hasil Hausman Test (Random vs Fix)

Hausman test statistic:

H = 5.83566 with p-value = prob(chi-square(3) > 5.83566) = 0.119885

(A low p-value counts against the null hypothesis that the random effects

model is consistent, in favor of the fixed effects model.)

Sehingga disimpulkan model fix yang terpilih

Selanjutnya yang akan diuji asumsi klasiknya adalah model common effect:

1.Normality

Test→ Normality of Residual

Test for null hypothesis of normal distribution:

Chi-square(2) = 9.586 with p-value 0.00829

Heterosedastis

Test→ Hetrosedastycity (White Test)

White's test for heteroskedasticity

OLS estimates using 25 observations

Dependent variable: uhat^2

    VARIABLE       COEFFICIENT       STDERROR     T STAT   P-VALUE

const             -242.867         300.287       -0.809   0.43050

UK                   24.9716         35.0780       0.712   0.48679

K_IND               524.670          680.175         0.771   0.45173

K_AUD               177.286         194.344         0.912   0.37519

sq_UK               -0.626410         1.31245     -0.477   0.63962

UK_K_IND           -28.9264         24.4306       -1.184   0.25370

UK_K_AUD             -9.26173         16.3592       -0.566   0.57915

sq_K_IND           -114.349         405.956       -0.282   0.78180

K_IND_K_AUD       -326.227         300.383       -1.086   0.29355

Unadjusted R-squared = 0.38001

Test statistic: TR^2 = 9.500262,

with p-value = P(Chi-square(8) > 9.500262) = 0.301865

Uji Autokolerasi

Test→ Autocorelation

Breusch-Godfrey test for first-order autocorrelation

OLS estimates using 20 observations

Dependent variable: uhat

     VARIABLE       COEFFICIENT       STDERROR     T STAT   P-VALUE

const               -2.91791         2.61600     -1.115   0.28223

UK                   0.0893927       0.139526     0.641   0.53139

K_IND                 5.29939         2.63697       2.010   0.06281 *

K_AUD                 1.25887         1.73846       0.724   0.48013

uhat_1               1.05234         0.203363     5.175   0.00011 ***

Unadjusted R-squared = 0.651358

Test statistic: LMF = 28.024023,

with p-value = P(F(1,15) > 28.024) = 9.01e-005

Alternative statistic: TR^2 = 13.027151,

with p-value = P(Chi-square(1) > 13.0272) = 0.000307

Multikolinearitas

Test→ collinearity

Variance Inflation Factors

Minimum possible value = 1.0

Values > 10.0 may indicate a collinearity problem

   3)              UK   4.068

   4)           K_IND   1.481

   5)           K_AUD   4.877

VIF(j) = 1/(1 - R(j)^2), where R(j) is the multiple correlation coefficient

between variable j and the other independent variables

Properties of matrix X'X:

1-norm = 1137.33

Determinant = 217.82109

Reciprocal condition number = 5.0760414e-005

 

download data model regresi panel dalam m.excel