Chapter 4. Research Results
4.2. Empirical Findings
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
4.2. Empirical Findings
Table 12 to Table 14 shows the results of the effects of corporate governance on performance. In this study, the impact of the independent variables, rating of corporate governance (CGLV) and firm types, whether it is family firm or non-family firm (FAM), have been evaluated based on the ROA1, ROA2, and ROE of the selected firms. The summary output of the regression models ROA1, ROA2 and ROE run based on the information collected is as shown below each tables.
Table12 and 13 show the results of the regression models, their dependent variables are ROA1 and ROA2 which indicates firm performance. ROA1 is the net income divided by total assets, ROA2 which is the operating profit divided by total assets.
Both variables refer to Return on Assets, which represents how much a company has benefited from using their assets. In other words, it shows how efficiently the firm has managed its assets.
First, according to Table 12, model 1 column incorporates only control variables, and reports their influence on firm performance. Model 2 column incorporates only explanatory variables FAM; CGLV, and reports their influence on firm performance.
Model 3 column (the full model) combines two types of variables to explain the impact on the firm performance. The F -value for each model is significant at the 1% level and the adjusted R2 is 0.322 for the full model.
‧
Table 12 Regression Result of the Effect of Corporate Governance on Firm Performance ( measured ROA1 )
‧
In the model 1, as for the Board of Directors' variables, the coefficient of BDM is -0.446 and statically significant at 1% level. Based on the results, the coefficient of the AUDIT indicates that audit committee has a positive effect on firm performance. FSHRP represents the number of family member and large shareholders in the board of directors, and USH represents the shareholding of family and family related stakeholders. Both variables are closely related to the family dominant firm. As a result, both FSHRP and USH variables showed a positive value. The coefficient of FSHRP is 1.819 and statically significant at 10% level, the coefficient of USH is 0.079 and statically significant at 1%
level. As a result of this study, the coefficient 0.048 and 1% significant positive relationship is observed between FOR and firm performance. The higher the stake of the foreign investor, the smaller the problem of the agent of the manager. Therefore, there is a positive relationship between the foreign stake and the firm performance. The RPT on the presence or absence of transactions between related parties was expected to have a negative impact on corporate performance. As a result, the relationship between RPT and corporate performance was negative. In the case of MB, which is a substitute variable of firm value, it is expected to have a positive effect on firm performance. The coefficient of MB is 0.465 and statically significant level is 10%. GROW represents the sales growth rate of each company and it is expected to have a positive relationship with the corporate performance because it is a variable indicating the growth rate of the corporation. As a result, The coefficient of GROW is 0.027 and statically significant level is 5%. TURN indicates the efficiency of the invested capital, which is also expected to have a positive relationship with corporate performance, the coefficient of TURN is 0.047 and statically significant at 1% level. In addition, SIZE is the logarithm of the asset, which indicates the size of the company and is expected to have a positive relationship with corporate performance. As a result, the coefficient of SIZE is 1.882 positive with 1% significantly level. In a study of RND, we expected RND and firm performance to have a positive relationship, but the results shows that has a coefficient -0.363 and is statically significant at 5% level. As the result of the analysis, it is found that LEV is negatively related to firm performance and significant at 1% level.
The model 2 examines the effect of explanatory variables CGLV and FAM. In the case of CGLV, there is positive correlation with ROA1, ROA2 and ROE in the correlative analysis as mentioned above previously. In the results of regression, the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
coefficient of CGLV is 1.567 and statically significance 1% level. Here another explanatory variable is significant that is FAM. As a result, the coefficient of FAM is 7.062 and statically significance 1% level.
In the model 3, when all variables and firm performance are analyzed, the CGLV has a positive effect on firm performance. Because, the coefficient of CGLV is 0.437 and statically significance 10%. The coefficient of FAM is 4.191 and statically significance 1% as model 2. The variable BDM, representing the number of members of the board, is found to be significantly related to firm performance at the 5% level. In terms of OUT which is ratio of outside directors in the board, the results indicate a significant 5%
negative relationship with firm performance. We can know that both FSHRP and USH, corporate governance and family-related variables, have a positive effect on firm performance. The coefficient of FSHRP is 1.462 and statically significance 10%, the coefficient of USH is 0.046 and statically significance 1%. This result also shows that the hypothesis 2 is supported. As like the expectation, the coefficient of FOR is 0.048 and statically significant level 1%. It means that foreign ownership 4.8% positively influences on firm performance. The control variable MB is found to have a significant 10% negative relationship with firm performance. And, GROW has a coefficient 0.027 and statically significant level 5%. As like the expectation, Both TURN and SIZE are found to be significantly 1% associated with firm performance. The coefficient of RND is -0.354 and statically significance 5%. LEV is found to have a coefficient -0.207 and significant 1%
relationship with firm performance for the full model.
‧
Table 13 Regression Result of the Effect of Corporate Governance on Firm Performance ( measured ROA2 )
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
As mentioned above, Table 13 is the results of the relationship between family firm, corporate governance and firm’s performance adopting ROA2. Table 13 also analyzes the three models. Model 1 is a description of the control variables analysis, model 2 is a description of explanatory variables analysis, and model 3 is an explanation of all variables analysis. Compared with the results of the regression models for ROA1 in Table 12 described above, there is no big difference. In addition, the signs of the positive and negative which represent the impacts of each variable on firm performance is exactly as in Table 12. The F -value for each model is significant at the 1% level and the adjusted R2 is 31% for the full model.
First, in model 1, the coefficient of BDM is -0.343 and statically significant level is 1%. It means that number of board of director influences the on firm performance 34.3%
negatively. The coefficient of USH is 0.041 and statically significant level is 1%. It means that the shareholding of family and family related stakeholders 4.1% positively influences the firm performance. The variable FOR, represents foreign stakeholders, has a coefficient 0.059 and statically significant 1% level. As expected, in other control variables which are MB, GROW, TURN and SIZE have a positive effect on firm performance with 1% significant level. In case of RND, it has a negative effect on firm performance as not expected. Its coefficient is -0.321 with significantly 10%. And, the coefficient of LEV is -0.81 and statically significance 1%, it also has a negative effect on firm performance.
In model 2, there are CGLV and FAM which are explanatory variable in this research. In the case of CGLV, it has a coefficient 1.085 and statically significance 1%
level. And, the coefficient of FAM is 3.264 and statically significance 1% level. Both CGLV and FAM have a positive effect on firm performance, so the result of ROA2 seems as same as the result of ROA1.
In the model 3, as the results of CGLV and FAM, the coefficient of CGLV is 0.256, and the coefficient of FAM is 1.762 and statically significant at 5% level. So, both explanatory variables have a positive effect on the firm performance. In case of variables which are related corporate governance, BDM has a coefficient -0.329 and statically
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
foreign stakeholder’s ratio, has a coefficient 0.059 with significantly 1%. As for expected, control variables which are MB, GROW, TURN, SIZE have a positive effect on firm performance with significantly 1%. But, RND has a negative effect on performance as not expected. In case of LEV, its coefficient is -0.080 and statically significant at 1%
level.
‧
Table 14 Regression Results of the Effect of Corporate Governance on Firm Performance ( measured ROE )
‧
Table 14 shows the result of the regression models, its dependent variable is ROE which also indicates the firm performance. The F -value for each model is significant at the 1% level and the adjusted R2 for each of model is 0.221 for the full model.
ROE shows the ability of a corporation to generate profits, which indicates how much profit is raised by the manager using the capital invested in the corporation. A company with a high ROE is a company that uses capital efficiently and has high profitability. It tends to have a high share price, which is used as an investment indicator and closely related to corporate performance. Tables 12 and 13 also illustrate the dependent variables ROA1 and ROA2, which are related to firm performance. ROE is explained separately because there is a noticeable difference from the previous analysis.
This is as follows.
The relationship between the ROE and the explanatory variables CGLV and FAM is as positive as in the previous tables. And, among control variables, BDM, OUT, AUDIT and FSHRP, which are related to corporate governance, are also the same as those of the previous tables. They have a positive effect on firm performance.
However, other variables which are related corporate governance, USH, FOR and RPT, came out differently. In the case of USH, the relationship between control variables and firm performance in model 1 is positive, but in model 3 it shows negative relationship. In the previous tables, FOR showed a positive relationship with firm performance, but when viewed in relation to ROE, it shows a negative relationship of 10%
significantly. The coefficient of FOR is -0.396 in model 1 and -0.397 in model 3. In the case of RPT, it was judged to have a negative impact on firm performance in the previous analysis, but it has positive in relation to ROE. Since then, GROW has also been positively related to firm performance in the preceding tables, but negative in relation to ROE.
The results of the regression analysis of the effects of corporate governance and family dominant on corporate performance are summarized as follows:
First, CGLV is a statistically significant positive relationship among dependent variables ROA1, ROA2 and ROE which are indicated firm performance. This is consistent with previous studies that positive corporate governance has a positive effect on firm performance and firm value (Yoon and Oh, 2005; Jeon, 2007; Lee and Son, 2009;
Sohn, 2010). Therefore, this supports Hypothesis H1 of this paper.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Second, the relationship between firm performance variables and FAM is also statistically significant positive. This indicates that if different types of companies have the same corporate governance structure, effect on the firm performance shows differently. So, the hypothesis H2 of this paper is verified. In addition, it is found that the family governance structure has a positive relationship with corporate performance through the variables related with corporate governance and family among the control variables. In the analysis of each ROA1 and ROA2, FSHRP and USH, which are variables related to family and corporate governance, FSHRP have a positive effect on firm performance and USH also have a statistically significant 1% positive relationship with firm performance. Therefore, this also supports hypothesis H2 of this paper.
Third, the fit of the three regression equations above shows that the F-values are all significant at the 1% level. And, the adjusted R2 of each model in each regression table shows that when analyzed together with the explanatory variables and the control variables, it is significantly increased better than when analyzed the explanatory variables and the control variables separately.