4.1. Descriptive Statistics
The descriptive statistics concerning the financial and CSR performance indicators utilised in this research are presented in
Table 1. Focusing on the accounting-based performance metrics, the mean values for ROA and ROE are recorded at 7.118% and 7.884%, respectively. It is noteworthy that the firms within the sample exhibit significant performance variability, with the top-performing firm achieving an ROA of 43.040% and the lowest-performing firm showing an ROA of −30.41%. Similarly, the ROE figures range from 77.270% for the highest performer to −158.65% for the lowest. The observed disparity in profitability metrics is further supported by standard deviation values of 11.265 for ROA and 30.650 for ROE. The average EPS across the sample is ZAR 760.343, with the highest and lowest EPS recorded at ZAR 8843.49 and ZAR–923.300, respectively, during the sample period. The substantial standard deviation in EPS indicates considerable variation among the selected JSE firms regarding their earnings distribution decisions for shareholders.
Regarding market-based performance indicators, the average Tobin’s Q for the sample of JSE companies stands at 1.119, accompanied by a standard deviation of 0.885. The Tobin’s Q values among the selected firms range from a minimum of 0.140 to a maximum of 4.530 throughout the sample period. Additionally, EVA and MVA exhibit average values of −2.894 and 1.637, respectively, during the study period. EVA values fluctuate between −106.219 and 29.820, with a standard deviation of 15.172. Conversely, MVA shows minimum and maximum values of 0.470 and 9.520, respectively, with a standard deviation of 1.433. A key conclusion from this analysis is that the selected JSE firms demonstrated varied performance across the financial outcome metrics throughout the study period.
The performance indicators for CSR reveal that the average values for the CSR strategic score (csr_score), composite ESG score, and the social dimension of ESG (socia) are approximately 51%, 50%, and 54%, respectively. Notably, the CSR score for the firms listed on the JSE ranges from a minimum of 0.39% to a maximum of 99.58%, with a standard deviation of 29.4. Regarding other CSR performance metrics, the composite ESG score varies between 14.14% and 88.88%, while the social component of ESG fluctuates between 1.62% and 96.34%. This variation indicates that the firms in the sample exhibit differing levels of CSR performance, as evidenced by the observed minimum and maximum values. The disparities are further substantiated by the standard deviations, which are calculated to be 29.4 for the CSR score, 17.611 for the overall ESG score, and 21.053 for the social pillar of ESG.
Regarding the control variables related to board characteristics, the firms in the sample report an average board size of approximately 11 members, with independent directors, female directors, and black directors constituting about 61%, 28%, and 34% of the board, respectively. It is noteworthy that the highest proportions of independent, female, and ethnic directors within the board are approximately 91%, 69%, and 85%, respectively. On average, the JSE firms convened for board meetings six times per year during the study period, with the number of meetings ranging from a minimum of one to a maximum of 24. In terms of firm-specific control variables, the average leverage ratio for the sample firms is recorded at 0.560, accompanied by a standard deviation of 0.224. The average total assets of the firms under study amount to ZAR 141.858 billion, with a substantial standard deviation of 350,062,000. Furthermore, there exists a significant disparity in firm size among the JSE sample firms, with the smallest and largest asset sizes recorded at ZAR 1.038 billion and ZAR 2004.402 billion, respectively.
4.2. Correlational Analysis
Correlational analysis offers preliminary insights into the relationships between the dependent variable and the independent variables. Additionally, this analysis can indicate the potential presence of multicollinearity among the explanatory variables within the model. To assess the strength of the associations among the variables of interest, the study utilises the Pearson moment correlation coefficient, as detailed in
Table 2. In examining column 1, it is evident that the three indicators of CSR performance—csr_s, esg, and soc—exhibit positive and statistically significant correlations with the accounting-based performance measure ROA, with correlation coefficients of 0.084, 0.210, and 0.206, respectively. Similar patterns are observed between the CSR performance indicators and other accounting-based metrics, such as ROE and EPS. Conversely, regarding market-based performance indicators, the composite ESG and the social pillar (soc) show positive and significant correlations with TOBQ, with correlation coefficients of 0.098 and 0.125, respectively. Furthermore, MVA is positively and significantly correlated with ESG (0.146) and soc (0.124) at the 5% and 1% significance levels, respectively. In contrast, the market-based performance measure, EVA, demonstrates negative and significant correlations with csr_s (−0.198), esg (−0.093), and soc (−0.069) at the 1%, 5%, and 10% significance levels, respectively.
The investigation into the relationship between board characteristics and firm performance indicators reveals that board size (bsize) exhibits a negative and statistically significant correlation with performance metrics such as ROA at −0.109 and EVA at −0.201. Conversely, it positively correlates with accounting-based performance indicators, including ROE at 0.092 and EPS at 0.308. This indicates that the relationship between board size and firm performance depends on the specific performance proxy utilised. In contrast, board independence (bind) demonstrates a negative and significant relationship with performance indicators such as EPS (−0.068), TOBQ (−0.110), and EVA (−0.096), significant at the 10%, 1%, and 5% levels, respectively. Additionally, gender diversity (gend) is positively and significantly correlated with firm performance measures, including roa (0.118), eps (0.103), and MVA (0.131) at the 1%, 5%, and 1% significance levels, respectively. The relationship between ethnic diversity (ethn) and firm performance appears to be mixed, varying according to the performance measure employed. Furthermore, board meetings (bmeet) are negatively and significantly associated with all the performance indicators at the 1% significance level. Regarding firm-specific control variables, a negative and significant correlation is observed between the leverage ratio (lev) and all the performance metrics, except for MVA, which shows a positive association with lev at the 5% significance level. Similarly, firm size (fsize) is negatively and significantly correlated with ROA (−0.125) and market-based performance measures such as TOBQ (−0.108) and EVA (−0.343) at the 1% level, while a positive correlation is noted between fsize and accounting-based performance metrics like ROE (0.180) and EPS (0.358) at the 1% level.
The analysis of the correlation coefficients presented in
Table 2 indicates that the relationships among the explanatory variables in this study are characterised by weak to moderate correlations. According to
Gujarati and Porter (
2008), a correlation coefficient of 0.8 is considered the minimum threshold for identifying multicollinearity. The data in
Table 2 reveals that the highest correlation coefficient observed is between fsize and bsize, which stands at 0.629, falling below the established threshold. This finding implies that multicollinearity does not pose a significant concern across the estimated models.
4.6. Regression Results
The findings derived from the GLS methodology regarding the influence of CSR score (csr_s) on FP are presented in
Table 7, with the overall sample and nonfinancial firms categorised as panel A and panel B, respectively. Initially focusing on the primary variable, the analysis indicates a significant positive correlation between CSR score and FP across the complete sample, except for the EVA in model 6. Statistically, an increase in one unit in the CSR score results in enhancements of 0.097, 0.095, 8.374, 0.005, and 0.010 units in ROA, ROE, EPS, TOBQ, and MVA, respectively, at varying levels of significance.
This finding confirmed the previous findings of
Aftab et al. (
2024),
Kaimal and Uzma (
2024), and
Li et al. (
2023), grounded in the stakeholder theory. It also demonstrates that CSR performance enhances a company’s financial performance, thereby corroborating
Freeman’s (
1984) assertion that a CSR-focused strategy can lead to improved financial results over the long term. The beneficial effect of CSR involvement on FP is evident across both accounting-based financial metrics (models 1, 2, and 3) and market-oriented performance indicators such as TOBQ and MVA in models 4 and 5. Conversely, within the nonfinancial subsample (panel B), the CSR score positively influences financial performance as indicated by EPS, TOBQ, and MVA at a significance level of 10%. This finding implies that a company’s commitment to CSR initiatives fosters profitability and enhances firm value, thereby contributing to improved FP for the selected companies listed on the JSE. Furthermore, when employing the composite ESG score (esg) and the social dimension (socia) as indicators of CSR performance, the results illustrated in
Table 8 and
Table 9 reveal that CSR performance consistently bolsters corporate performance both in Panel A and Panel B. This consistency is observed across both accounting-based and market-based performance measures, applicable to both the full sample and the nonfinancial subsample. In
Table 8, statistically, an increase of one unit in the ESG score (esg) results in enhancements of 0.253, 0.178, 16.37, 0.013, 0.077, and 0.017 units in ROA, ROE, EPS, TOBQ, EVA, and MVA, respectively, at varying levels of significance. In
Table 9, statistically, an increase of one unit in the social score (socia) results in enhancements of 0.211, 0.204, 15.56, 0.012, 0.105, and 0.012 units in ROA, ROE, EPS, TOBQ, EVA, and MVA, respectively, at varying levels of significance. Thus, it can be concluded that irrespective of the CSR performance metric utilised, CSR practices significantly enhance the performance of firms listed on the JSE. This is consistent with the rational choice and stakeholder–agency theories, and the findings of prior research (
Chininga et al., 2024;
Fu & Li, 2023;
Shan et al., 2024), which found that ESG performance enhances the FP of firms.
Masongweni and Simo-Kengne (
2024) have also found a positive association between the social pillar score and FP. Thus, a growing body of evidence, supported by established theories, indicates that strong ESG performance is positively linked to improved FP for businesses.
The analysis of board characteristic variables, as presented in
Table 7,
Table 8 and
Table 9, indicates that both board independence (bind) exerts a negative and statistically significant effect on various measures of FP within the aggregate sample (panel A) and among nonfinancial firms (panel B). The results are consistent with those of
Khan et al. (
2024) and
Sahoo et al. (
2023), who suggest that independent directors play a passive role and maintain strong connections with company management. Therefore, while promoting independence in board composition is essential for good governance, firms should carefully consider the optimal balance between independence and direct managerial involvement to foster FP. The frequency of board meetings (bmeet) also exerts negative and statistically significant effects on various measures of FP within the aggregate sample (panel A) and among nonfinancial firms (panel B) in
Table 7,
Table 8 and
Table 9. These findings are in contrast with those of
Sahoo et al. (
2023). This finding suggests that an increase in the proportion of external directors and the frequency of board meetings may detract from FP. While regular meetings are necessary for oversight, excessive frequency may indicate governance issues or a reactive approach to problem-solving, which can hinder a firm’s agility and overall performance. Companies should therefore aim to balance adequate board engagement with maintaining a strategic, outcomes-focused meeting structure.
Conversely, gender diversity (gend) appears to positively influence financial performance metrics across both samples. This effect is robust and consistent across all models detailed in
Table 7,
Table 8 and
Table 9, highlighting that female representation on boards is a significant contributor to FP in South Africa. These results are consistent with studies conducted by
Sahoo et al. (
2023). In contrast, other board characteristics, such as board size (bsize) and ethnic diversity (ethn), exhibit mixed effects depending on the specific performance measure analysed. For example, findings from Model 1 in
Table 7,
Table 8 and
Table 9 indicate that an increase in board size correlates with a significant decrease in ROA (
p < 0.05). This finding supports the study conducted by
Sarpong-Danquah et al. (
2023), who found that board size negatively impacts FP. At the same time, it shows a positive and statistically significant relationship with the EPS (
p < 0.1) for the aggregate sample. In the context of nonfinancial firms (panel B), board size does not emerge as a critical determinant of FP, particularly when CSR scores are utilised as a performance metric, as shown in
Table 5. However,
Table 8 and
Table 9 reveal that board size is negatively and significantly related to ROA at a 10% significance level (model 7 in
Table 8 and
Table 9), while its influence on other performance indicators remains negligible. These results suggest that while larger boards may dilute decision-making efficiency, they also potentially bring diverse expertise that positively affects FP. Similarly, the effect of ethnic diversity (ethn) on FP is predominantly negative across all model specifications. The influence of the proportion of black directors on FP, as indicated by market-based indices, reveals a correlation where an increase in such representation is linked to a decline in both profitability and market value among the sampled firms in South Africa. This may reflect challenges associated with diverse perspectives, potentially leading to slower consensus in decision making.
In alignment with prior expectations, the leverage ratio exhibits a negative and statistically significant effect on FP across various significance levels, except for the MVA specification (models 6 and 12 in
Table 7,
Table 8 and
Table 9). This adverse effect of the leverage ratio remains consistent across different performance metrics and is evident in both the full sample and the regressions focused on nonfinancial firms. The findings strongly emphasise the detrimental impact of elevated leverage ratios on financial and market performance indicators. Specifically, as illustrated in panel A of
Table 7, a one-unit increase in the leverage ratio results in declines of 15.84 units in ROA, 30.71 units in ROE, 2.166 units in EPS, 0.659 units in TOBQ, and 11.22 units in EVA at a 1% significance level. In contrast, a similar increase in leverage ratio adversely affects the ROA, ROE, EPS, TOBQ, and EVA of nonfinancial firms (Panel B,
Table 7) by 17.82 units, 51.30 units, 2.948 units, 0.562 units, and 4.226 units, respectively. Interestingly, the analysis indicates that an increase in leverage ratio is associated with a higher MVA of 0.518 units for the entire sample, as shown in
Table 8. When examining the nonfinancial companies, the results regarding the leverage ratio’s impact on financial outcomes are consistent with those of the aggregate sample, indicating that firms with high leverage tend to exhibit lower performance, except for the MVA regression.
The examination of the second firm-specific control variable reveals that FP indicators exhibit varying responses to alterations in firm size (fsize). As illustrated in Panel A of
Table 7, (fsize) demonstrates a positive and statistically significant influence on ROE in model 2 and EPS in model 3, indicating that larger firms tend to achieve superior performance across the overall sample. Comparable results are observed in
Table 8 and
Table 9 when utilising the composite ESG and its social component as measures of CSR performance. Conversely, market-based performance metrics, such as TOBQ and EVA, negatively correlate with firm size when analysing the aggregate sample (Panel A). This finding is consistent with the outcomes presented in models 4 and 5 of
Table 8 and
Table 9. This suggests that the impact of firm size on performance indicators is heterogeneous, contingent upon whether accounting-based or market-based metrics are employed. Furthermore, the relationship between firm performance indicators and firm size within nonfinancial sectors lacks consistency across various specifications. In summary, accounting-based financial performance measures exhibit a positive and significant response to changes in firm size (as seen in models 7, 8, and 9 of
Table 7,
Table 8 and
Table 9). In contrast, the influence of the firm-specific variable on market-based performance appears negligible.