*5.1. Descriptive Statistics and Multicollinearity Diagnostic*

Table 3, Panel A reports the descriptive statistics of all the variables used in our methodology. The overall descriptive statistics show the mean value of performance measures used in our analysis at 4.3 percent, 6.49 percent, and 0.39 per share for ROA, ROE, and EPS, respectively. The average value of the administrative expense ratio is 9.8 percent, while the free cash flows to total assets have a negative average value of −18.7 percent of total assets. The absolute value of discretionary accruals denoted by EM has a mean value of 5.43 percent of total assets. The mean value of the corporate governance quality index represented by CGQ is 5.79. The maximum amount of the shareholding percentage of the largest shareholder (top 1) is 89.99 percent.


**Table 3.** Summary statistics.

We divided the descriptive statistics into two more panels based on state ownership and non-state ownership. The mean values of ROA, ROE, and EPS in Non-SOE (4.9 percent, 8.6 percent, and 4.01 respectively) are higher than those of the SOE (3.5 percent, 3.27 percent, 4.02 per share respectively). State-owned enterprises have a higher level of absolute discretionary accruals (0.07 > 0.06), as well as a higher administrative expense ratio (0.102 > 0.081) compared to non-SOE. The non-state companies have a better corporate governance score of 6.07 compared to state companies, which have a score of 5.35. The state-owned companies are larger (22.46 > 21.49), as well as having a higher average leverage (0.505 > 0.353), than non-state-owned companies.

Panel B presents the descriptive statistics of the corporate governance variables used in this study divided into three groups: Overall, State, and Non-State. The A-listed companies from the period 2008 to 2016 have an average independent director of 3.29, reported CEO duality of 26.27 percent of the total sample, average board size of 8.8, while 16.7 percent on average present female representation on the board. On average, non-state firms pay a higher level of dividends (0.13 > 0.11), have a higher female ratio on the board (0.18 > 0.13), have higher managerial ownership (16.63 percent > 11.57 percent), and a higher degree of separation between control rights and cash flow rights (6.26 percent > 4.38 percent) as compared to state-owned companies. All these governance attributes lead to a better governance environment, resulting in less managerial expropriation. State-owned companies have a greater board size (9.46 > 8.38), more institutional ownership (8.8 > 5.8), more access to getting audited by Big Four auditors (0.087 > 0.02) and more independent directors on the board (3.44 > 3.09).

Table 4 shows the results for pairwise correlation analysis. Looking at all of the independent variables, we find no sign of multicollinearity, as the values of the coefficients are less than 0.8. Additionally, we performed individual VIF analysis and found all the values to be less than the critical level of 10 in every case (Shan 2015). When performing the regression analysis for the moderating effect, the interaction term usually gives a value for VIF greater than 10. Following Allison (2010), we mean-centered the interaction terms (agency cost variables and their interaction with CGQ, Top1, SOE, and NSOE). The use of mean centering does not affect the probability values, and at the same time, the multicollinearity is reduced, as seen from Table 5. All the dependent variables are positively correlated with corporate governance quality. On the other hand, the agency cost proxies are negatively related to the performance measures. Ownership concentrations also have a positive association with the firm performance.

#### *5.2. Moderating E*ff*ects of Corporate Governance Quality*

Table 6 shows the effect of corporate governance and agency cost on firm performance by using the fixed-effect model and system dynamic panel data estimation. ROA and EPS were the dependent variables, while agency cost and corporate governance were the independent variables. Four variables, namely, Size, Leverage, Growth, and Firm age, were used as control variables. The asset size used has a positive effect on firm performance while, on the contrary, a higher level of leverage hampers performance. These results are similar to the study conducted by Vithessonthi and Tongurai (2015).



Note: Table 4 reports the correlation coefficients of the variables used in Equations (1) and (2); the detailed calculation is presented in Table 1. signify *p*-values of 1



Note: Table 5 reports the VIF diagnostics with tolerance values; the detailed calculation is presented in Table 1.



Note: Table 6 reports the regression results from estimating Equations (1) and (2) respectively. Variable definitions are provided in Table 2. \*, \*\*, \*\*\* signify *p*-values at 1 percent, 5 percent,and 10 percent, respectively.

#### *JRFM* **2020** , *13*, 154

Looking at the primary results, agency cost (administrative expense ratio and free cash flow) has a significant adverse effect on firm performance measured by ROA and EPS. Corporate governance has a positive impact on both returns on ROA (FE: β = 0.003, *p* < 0.01; GMM: β = 0.002, *p* < 0.01) as well as EPS (FE: β = 0.003, *p* < 0.01; GMM: β = 0.004, *p* < 0.01). The objective of the study was to investigate the moderating effect of corporate governance quality on the relationship between agency cost and firm performance. To do so, we incorporated the interactions terms of corporate governance quality with agency cost proxies. The results are depicted in Table 6 with ROA and EPS as dependent variables. Corporate governance quality significantly positively moderates the relationship between agency cost and firm performance. When interaction terms are introduced, the coefficient of agency cost has changed the sign from negative to positive for both AC1 × CGQ (ROA—FE: β = 0.004, *p* < 0.01; GMM: β = 0.012, *p* < 0.01 and EPS—FE: β = 0.002, *p* < 0.05; GMM: β = 0.03, *p* < 0.1) and FCF-AC × CG (ROA—FE: β = 0.009, *p* < 0.01ta; GMM: β = 0.007, *p* < 0.1 and EPS—FE: β = 0.002, *p* < 0.05; GMM: β = 0.007, *p* < 0.1). These results indicate that the objectives of the principles and agents are aligned if the firms have the adopted good corporate governance practices reported by Jensen and Meckling (1976). Looking at the control variables, the size of the firm has a positive effect on the firm performance in all cases, which is in agreement with what we found in the literature. Leverage is negatively related to the performance, as can be seen by the signs of the negative coefficients in the table. Based on these results, we can accept our alternative hypothesis that corporate governance positively moderates the agency–performance relationship.

## *5.3. Moderating E*ff*ects of Ownership Concentration*

Table 7 indicates the effect of ownership concentration measured by the shareholding percentage of the top shareholder on the relationship between agency cost and firm performance. The results show a positive impact of ownership concentration on the ROA and EPS for both the fixed-effect model and GMM. These results are in line with those of Shleifer and Vishny (1986) and Li et al. (2008). We introduced two interaction terms to measure the moderating effect of ownership concentration on the relationship between firm performance and agency cost. As shown in the results, the interactions terms possess a positive coefficient value, with ROA (TOP\_1 × AC1 β = 0.002, *p* < 0.01 and TOP\_1 × FCF-AC β = 0.0014, *p* < 0.1) and EPS (TOP\_1 × AC1 β = 0.002 and TOP\_1 × AC2 β = 0.006) in the fixed effect model. Similarly, the GMM model also exhibits a positive moderating effect of ownership concentration on the relationship between agency cost and firm performance. At higher levels of ownership concentration (for the largest shareholder), the shareholders are highly vigilant, and this helps facilitate the alignment of interests among the agents and the principles. The results show that as the ownership concentration is increased, the firm performance increases on the one hand, while the agency cost is decreased on the other side. These results support the second hypothesis that the ownership concentration positively moderates the relationship between agency cost and firm performance.
