Normality Test Autocorrelation Test Jointly Regression Coefficient Test F-test

H 1 = fixed effect If Chi Square statistic Chi Square table or in other word, p-value 0.005, where it means that we should reject null hypothesis H and determine that fixed effect model is the suitable model to use Winarno, 2009. Hausman test is also available through Eviews-6 command program. Table 4.1 Hausman Test Result Test Summary Chi-Sq. Statistic p -value Effect CPI 7.300554 0.0069 Fixed RER 2.667002 0.1024 Fixed GFER 4.548718 0.0329 Fixed GRVT 0.855584 0.3550 Fixed GRRVT 0.857151 0.3545 Fixed GRYPC 0.840744 0.3592 Fixed Source: Eviews-6, 2010. Note: fixed effect while p-value 0,005

4.2.3 Classic Assumption Test Analysis

a. Normality Test

Normality test is done by examining Jarque-Bera value through X 2 table. From regression through Eviews 6.0 we find that J-B statistics as shown in Table 4.2, where it is described that CPI, RER, GFER, GRVT, GRRVT, and GRYPC, has a normal distribution, where are shown from their µ residual value. Positive Auto correlation Zone of indecision Do not reject H or H 1 or both Zone of indecision Negative Auto correlation 1.720 1.746 2.254 2.280 4 Table 4.2 Normality Test Result Test Summary Df X 2 -table Jarque-Bera Result CPI 9.139 23.5893 4.0928 Normal Distribution RER 9.139 23.5893 1.1423 Normal Distribution GFER 9.139 23.5893 0.1963 Normal Distribution GRVT 9.139 23.5893 10.0599 Normal Distribution GRRVT 9.139 23.5893 15.2685 Normal Distribution GRYPC 9.139 23.5893 10.6329 Normal Distribution Source: Eviews-6, 2010. Note: Jarque-Bera JB test method is measuring value of skewness and kurtosis where if JB X 2 Chi-square value table, it means that residual value distribution is normal Gujarati, 2003.

b. Autocorrelation Test

One of formal test to detect autocorrelation is Durbin-Watson. This test is based on error model shown below; Figure 4.25 Durbin-Watson Test Note: H0: No positive autocorrelation H1: No negative autocorrelation Based on Durbin-Watson, this study found that in this research the equations are generally high potential to be free from autocorrelation, as it is described on Table 4.3. Table 4.3 Durbin-Watson Test Result Test Summary K dL Du Dw R 2 DwR 2 Result CPI 1 1.720 1.746 1.744521 0.824031 2.117 Negative Autocorrelation RER 1 1.720 1.746 1.820670 0.932277 1.953 Negative Autocorrelation GFER 1 1.720 1.746 2.280812 0.228518 9.981 Negative Autocorrelation GRVT 1 1.720 1.746 2.279058 0.227245 10.029 Negative Autocorrelation GRRVT 1 1.720 1.746 2.137546 0.153866 13.892 Negative Autocorrelation GRYPC 1 1.720 1.746 2.266261 0.058549 38.707 Negative Autocorrelation Source: Eviews-6, 2010.

c. Heteroscedasticity Test

Heteroscedasticity test purpose is to know whether all the disturbance term are similar variants or not Gujarati, 2003. This research study used White’s Heteroscedasticity-Consistent Variances and Standard Errors . White has shown that this estimate can be performed so that there is asymptotically valid i.e., large-sample statistically inference can be made about true parameter values. As the preceding result show, White’s heteroscedasticity-corrected standard errors are considerably larger than the OLS standard errors and therefore the estimated t values are much smaller than those obtained by OLS. On the basis of the latter, both the regressors are statistically significant at the 5 percent level, whereas on the basis of White’s estimators they are not. However, it should be pointed out that White’s heteroscedasticity-corrected standard errors can be larger or smaller than the uncorrected standard errors Gujarati, 2003. Table 4.4 Heteroscedasticity Test Result Test Summary Probability Result CPI 0.000000 Heteroscedasticity free RER 0.000000 Heteroscedasticity free GFER 0.000229 Heteroscedasticity free GRVT 0.000060 Heteroscedasticity free GRRVT 0.000146 Heteroscedasticity free GRYPC 0.000000 Heteroscedasticity free Source: Eviews-6, 2010. Through Eviews-6, this research study examined the heteroscedasticity by Eviews-6 Equation Estimation command of White heteroscedasticity- consistent standard errors and covariance, where the result is injured by heteroscedasticity if the probability is significant, in the other side, the result is free from heteroscedasticity if the probability 0.005. The heteroscedasticity test summary result that described in Table 4.4. d. Multicollinearity Test Multicollinearity is a condition that describes a linear relationship across independent variables. Multicollinearity happens when there are more than one independent variables in the research study. Whether this research study independent variable is only one, because of that reason this research study econometric is free from multicollinearity.

4.2.4 Regression Statistic Test Analysis Hypothesis Test

a. Jointly Regression Coefficient Test F-test

F-test goal is to determine the significance of independent variable groups in influencing the dependent variable. In this research we use 95 degree of freedom α = 5. The conclusion of jointly regression coefficient test is described in Table 4.5. It means that independent variable groups influence the dependent variable. It is significant H is rejected and H 1 is accepted. Table 4.5 Jointly Regression Coefficient Test F test Test Summary Prob F - statistic Result CPI 0.000000 Significant RER 0.000000 Significant GFER 0.000229 Significant GRVT 0.000060 Significant GRRVT 0.000146 Significant GRYPC 0.000000 Significant Source: Eviews-6, 2010.

b. Individuality Coefficient Regression Test t-Test