B. CLASSICAL ASSUMPTION TESTS
This study includes several classic assumption tests: 1.
Autocorrelation Test Autocorrelation test is performed to determine any indication of
correlation in a time series observations. Autocorrelation test is done by calculating the value of Durbin-Watson in every regression model:
Table 4.2 Autocorrelation Test Results
DURBIN-WATSON SCORE Conclusion
MODEL 1 0.281
Autocorrelation
MODEL 2 0.075
Autocorrelation
MODEL 3 1.552
Autocorrelation
MODEL 4 1.125
Autocorrelation
ROBUST 1 0.271
Autocorrelation
ROBUST 2 0.074
Autocorrelation
ROBUST 3 1.551
Autocorrelation
ROBUST 4 1.124
Autocorrelation Source: Processed Data
Autocorrelation normally occurs in the time series data, including the data used in this study.
2. Multicollinearity Test
Multicollinearity test is conducted to identify whether there is a strong correlation between two or more independent variables or not. The table
below shows the results of multicollinearity test for both regression models:
Table 4.3 Multicollinearity Test Results I
ROE SOE
LOCAL FOREIGN
VENTURE LNTA
EQTA LDR
ROE 1
SOE -0.11461
1
LOCAL
0.09893 -0.12805
1
FOREIGN 0.19757
-0.06418 -0.14773
1
VENTURE 0.03181
-0.08813 -0.20284
-0.10167 1
LNTA
0.085716 0.44413
-0.03024 0.111204
0.057519 1
EQTA -0.07509
-0.07475 -0.18237
-0.20563 0.008002
-0.42899 1
LDR -0.00538
-0.05173 -0.19606
0.199326 0.16873
-0.06309 0.199686
1
Source: Processed Data
Table 4.4 Multicollinearity Test Results II
ROE SOELOC
VENFOR LNTA
EQTA LDR
ROE 1
SOELOC 0.03533
1
VENFOR 0.15373
-0.30371 1
LNTA 0.085716
0.19271 0.11954
1
EQTA -0.07509
-0.20743 -0.12575
-0.42899 1
LDR -0.00538
-0.20876 0.26506
-0.06309 0.19686
1
Source: Processed Data
When the correlation score is 0.90 or above, the variable might be suffers the multicollinearity problem. From the table above, it can be
concluded that this research do not have multicollinearity effect. 3.
Heteroscedasticity Test Heteroscedasticity test in this study was not done separately, but rather
included as a factor considered in the regression test.
C. EMPIRICAL RESULTS