Regression and hypothesis testing

Regression is the test which analyzes the relationship between variables is negative or positive and the strength of the relationship. To define which model is suitable to analyze this data, Hausman test is conducted and after that regression testing is performed.

4.4.1 Hausman Test Result

Table 4.7 Output of Hausman Test Correlated Random Effects - Hausman Test

Test cross-section random effects

Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.

Cross-section random 4.385578 7 0.7344

Depend on the consequences of the Hausman tests, probability of the Chi Square statistics is worth 0.7344. The hypothesis is

H0: Random Effect Model H1: Fixed Effect Model

Known the number of k or the degree of freedom is 7 at the level 5%, then the value of the table chi squares is 4.385. The result also achieved p_value is 0.7344 that is bigger than alpha 0.05 so the researcher do not reject H0 which mean the best model to used is Random Effect Model (REM) or we can conclude that and independent variable is not correlated.

4.4.2 Multiple Linear Regression Estimation Results

The productivity of Random Effect Model is shown below:

ROE = -0.24 + 0.245AT + 0.016LEV + 0.012NPM + 3.91E-05TIE + 0.041LOGA - 0.05BETA

The results of model testing with Eview 8.1 showed that LOG (ASSETS), AT, NPM, RISK, ATE, TIE variables were statistically significant and these variables accounted for 22.27% of the variation of ROE. Specifically, LOGA (ASSETS), NPM, ATE, TIE and AT have positive associated on the ROE dependent variable, and RISK (BETA) variables have negative associated on ROE changes.

Table 4.8 Output of Random Effect Model Dependent Variable: ROE

Method: Panel EGLS (Cross-section random effects) Date: 04/21/17 Time: 22:34

Sample: 2007-2016 Periods included: 10 Cross-sections included: 35

Total panel (unbalanced) observations: 305

Variable Coefficient Std. Error t-Statistic Prob.

C -0.24441 0.12937 -1.88916 0.0598

LOGA 0.04123 0.02043 2.01799 0.0445

BETA -0.05331 0.01869 -2.85221 0.0046

Weighted Statistics

The coefficient of LEV is 0.016 indicated that if ratio asset to equity increase 1%, the ROE will surge 1.6%. The coefficient of NPM is 0.012 indicated that if ratio net profit margin increase 1%, the ROE will increase 1.2%. The coefficient of AT is 0.245 showed that if ratio asset turnover increase 1%, the ROE will increase 24.5%. The coefficient of TIE is 3.91E-05 indicated that if ratio times interest earned increase 1%, the ROE will increase 0.0039%. The coefficient of LOGA is 0.041 indicated that if asset ratio increase 1%, the ROE will increase 4.1%. In contrast, the coefficient of BETA is -0.05 indicated that if ratio Risk increase 1%, the ROE will decrease 5%.

4.4.3 The Results of Coefficient Determination Test (R

)

The result of random effects model (REM) showed that the value of the coefficients determination (R²) is 0.222738. Therefore, 22.27% variation or the changing to ROE explained by the independent variables, which are AT, TIE, LOG(ASSET), NPM, BETA and LEV except

CR towards real estate companies listed on Ho Chi Minh Stock Exchange. Prob (F-statistic) = 0.0000 < α (0.05) indicates the suitability of the model or the model exists. The p_value of coefficients less than α (0.05) shows that only these variables in the model are statistically significant, explaining well the variability of the ROE at the significance level and the 95%

confidence level.

4.4.4 The Results of t-statistics value

The results of current ratio (CR) on the ROE are 0.356297 for t-statistics value while the p-statistics value is 0.7219. The result of the probability is larger compared with the level of α

= 5 %. It means partially current ratio have no significant impact on ROE.

The results of leverage (LEV) on the ROE are 2.189749 for t-statistics value while the p-statistics value is 0.0293. The result of the probability is smaller compared with the level of α

= 5 %. It means partially leverage have positive significant impact on ROE.

The results of net profit margin (NPM) on the ROE are 2.154763 for t-statistics value while the p-statistics value is 0.0320. The result of the probability is smaller compared with the level of α = 5 %. It means partially net profit margin have positive significant impact on ROE.

The results of asset turnover (AT) on the ROE are 5.602498 for t-statistics value while the p-statistics value is 0.0000. The result of the probability is smaller compared with the level of α = 5 %. It means partially asset turnover have positive significant impact on ROE.

The results of time interest earned (TIE) on the ROE are 3.523776 for t-statistics value while the p-statistics value is 0.0005. The result of the probability is smaller compared with the level of α = 5 %. It means partially time interest earned have positive significant impact on ROE.

The results of Risk (BETA) on the ROE are -2.852211 for t-statistics value while the p-statistics value is 0.0046. The result of the probability is smaller compared with the level of α

= 5 %. It means partially Risk have negative significant impact on ROE.

The results of assets (LOGA) on the ROE are 2.017998 for t-statistics value while the p-statistics value is 0.0445. The result of the probability is smaller compared with the level of α

= 5 %. It means partially assets have positive significant impact on ROE.

In a summary, the results in table 4.9 show that H1 is rejected and H2, H3, H4, H5, H6 and H7 are all supported.

Table 4.9 Hypothesis Summary

Statement Result H1 There is a negative relationship between Current Ratio and ROE. Rejected

H2 There is a positive relationship between Asset Turnover and ROE Supported H3 There is a positive relationship between Leverage and ROE Supported H4 There is a positive relationship between Net Profit Margin and ROE Supported H5 There is a positive relationship between Times Interest Earned and ROE. Supported H6 There is a positive relationship between Assets and ROE Supported H7 There is a negative relationship between Risk and ROE Supported

CHAPTER V CONCLUSIONS AND