4.1. Descriptive Statistics and Correlation Analysis
The descriptive statistics for the variables used in this study are presented in
Table 2. The dependent variable, FESG, representing the predicted ESG score for firm i in year t + 1, has a mean of 4.281 and a standard deviation of 0.413. The minimum and maximum values are 2.235 and 5.991, respectively, indicating considerable variation in expected ESG investment performance across firms. The main independent variable, FEMALE, which captures the natural logarithm of one plus the number of female executives in firm i at time t, shows a mean of 0.320 and a standard deviation of 0.554. The median is 0.000, and the maximum value reaches 4.078, reflecting that while many firms lack female representation in top management, a minority of firms have relatively high female executive counts. The variable ABESG, reflecting the degree of abnormal ESG investment, has a mean of 0.019 and a standard deviation of 0.262. The values range from −0.691 to 0.506, suggesting a wide dispersion in ESG investment tendencies, from underinvestment to overinvestment.
Regarding control variables, firm size (SIZE), measured as the natural logarithm of total assets, has a mean of 26.930, with values ranging from 24.279 to 31.098. The market-to-book ratio (MTB) averages 1.594, but with a high standard deviation of 1.843 and a maximum of 13.092, indicating that some firms exhibit substantial growth potential. The ZSCORE, used to capture financial stability, has a mean of 2.841 and ranges from −1.446 to 30.173, highlighting the inclusion of both financially distressed and highly stable firms. The average operating cycle (OC) is 2.953, with values ranging between 1.365 and 7.098, pointing to significant differences in operational efficiency across firms. TAN, representing the ratio of tangible assets to total assets, has a mean of 0.174. STDCFO and STDSALES, which capture the volatility of cash flows and sales, average 0.047 and 0.123, respectively, with STDSALES reaching a maximum of 0.647, implying high revenue instability in some firms. The dummy variable DIV, which indicates whether a firm paid a dividend, has a mean of 0.605, suggesting that around 60.5% of firms distributed dividends. LEV, the leverage ratio, averages 0.973, with notable dispersion (standard deviation of 1.117). MEAN_K, reflecting average capital structure in the industry, averages 0.178. Finally, the LOSS dummy variable shows a mean of 0.262, indicating that approximately 26.2% of firms reported net losses. These descriptive statistics offer a comprehensive overview of the distribution and characteristics of the key variables in the sample and provide the basis for the empirical analysis that follows.
Table 3 presents the Pearson correlation coefficients among the key variables in the study: FESG (predicted ESG score), FEMALE (the number of female executives in senior management), and ABESG (abnormal ESG investment tendency).
As shown in the table, FESG is positively and significantly correlated with FEMALE, with a coefficient of 0.291 (p < 0.0001). This indicates that firms with greater female executive representation tend to exhibit higher predicted ESG performance, providing initial descriptive support for the study’s first hypothesis. The direction and significance of this correlation are consistent with the view that gender diversity in upper management contributes to more effective ESG investment. In contrast, the correlation between FESG and ABESG is extremely weak (r = 0.001) and statistically insignificant (p = 0.942), suggesting no linear association between the predicted ESG investment and the firm’s deviation from optimal ESG spending levels, as measured by ABESG. This finding reinforces the need for interaction terms and conditional analysis in the regression framework, where the role of female executives may vary depending on whether firms tend to under- or over-invest in ESG. The correlation between FEMALE and ABESG is slightly negative and statistically significant (r = −0.052, p = 0.003), implying that firms with more female executives are marginally less likely to exhibit tendencies toward abnormal ESG investment. While the magnitude of the correlation is small, the statistical significance suggests that gender diversity may play a role in moderating inefficiencies in ESG-related decision making. Overall, the correlation matrix shows no evidence of multicollinearity among the key independent variables, and the observed relationships are in line with the theoretical expectations that female leadership is positively associated with ESG performance and may contribute to curbing ESG inefficiencies.
4.2. Test Results of Hypotheses
Table 4 presents the results of the baseline regression examining the relationship between the presence of female executives and ESG investment efficiency, as specified in Equation (1). The dependent variable is FESG, which captures the predicted optimal level of ESG investment in year t + 1 based on current firm characteristics. Industry and year fixed effects are included to account for unobserved heterogeneity across sectors and time. The model explains a substantial proportion of the variation in ESG investment efficiency, with an adjusted R-squared of 0.77, and the F-statistic (374.28) confirms the overall significance of the regression at the 1% level.
The coefficient on FEMALE, which represents the natural logarithm of one plus the number of female executives in firm i at time t, is positive and statistically significant at the 5% level (β = 0.014, t = 1.97). This finding supports Hypothesis 1, indicating that greater female representation in upper management is associated with more efficient ESG investment, particularly for firms that previously exhibited underinvestment tendencies. This aligns with the argument that female executives contribute to governance quality and more balanced resource allocation.
This efficiency-enhancing role of female leadership is supported by recent studies. For instance, Alodat and Hao (2025) found that board gender diversity strengthens the link between ESG disclosure and firm performance by enhancing transparency and decision making [
16]. Similarly, Ouni et al. (2022) showed that gender-diverse boards improve intellectual capital efficiency, which is vital for sustainable investments [
17]. Moreover, Muhammad and Farooq (2025) demonstrated that female board members help mitigate ESG controversies through better oversight [
18].
The coefficient on ABESG, a ranked measure capturing the degree of abnormal ESG investment, is also positive and highly significant (β = 0.067, t = 4.25), suggesting that firms with higher tendencies toward excessive ESG investment tend to allocate more resources to ESG in the subsequent period, independent of the gender composition of their executive team.
Crucially, the interaction term FEMALE × ABESG is negative and statistically significant at the 1% level (
β = −0.087, t = −3.50). This finding supports the efficiency-based interpretation of ESG investment and indicates a moderating role of female executives. In firms predisposed to ESG overinvestment, the presence of more female executives leads to reduced ESG spending, suggesting a corrective governance effect. These findings corroborate earlier governance theories emphasizing that female leaders reduce managerial opportunism and enhance strategic alignment, particularly in sustainability contexts [
23,
24].
Recent empirical studies further reinforce this efficiency-oriented role of female leadership in ESG contexts. For example, Mensah and Onumah (2023) provide evidence that gender-diverse boards mitigate earnings management, indirectly contributing to more disciplined investment behavior [
25]. Similarly, Nadeem et al. (2019) show that female board participation improves intellectual capital allocation efficiency, which is closely linked to responsible ESG investment [
26]. In rapidly growing firms, Tran et al. (2022) report that gender-diverse top management teams enhance financial decision making and stakeholder responsiveness, reflecting a similar corrective governance effect as observed in this study [
27].
Regarding the control variables, SIZE is positively and highly significantly associated with FESG (β = 0.228, t = 77.90), consistent with prior studies showing that larger firms are more capable and often more pressured to engage in ESG activities. OC, TAN, STDSALES, DIV, and LEV all show significant positive relationships with ESG investment efficiency, suggesting that firms with stronger operations, stable sales, and shareholder payouts are more efficient in ESG allocation.
Table 5 presents the regression results that explore the impact of female executives on ESG investment efficiency, segmented by corporate lifecycle stages—specifically comparing firms in a steady state versus those in a growth state. The adjusted R-squared for both models is 0.77, suggesting a robust explanatory power for the variables considered. The model fits are further corroborated by significant F-values, 169.8 in steady state and 213.63 in growth state, both significant at the 0.1% level.
In firms classified within a steady state, the regression results show no significant effect of FEMALE or ABESG on ESG efficiency. However, in firms in a growth state, the presence of female executives has a marginally positive effect (t = 1.81), and the interaction with ABESG is significantly negative (β = −0.173, t = −4.78). This result implies that female leadership helps curb excessive ESG investment, particularly in dynamic growth contexts where resource misallocation risks are higher.
These findings align with Tran et al. (2022), who showed that female leadership in high-growth firms strengthens financial outcomes by enforcing budgetary discipline [
27]. Furthermore, Ab Aziz et al. (2025) found that board gender diversity reduces ESG-related controversies, especially during rapid expansion phases [
41]. These results underscore that female executives serve as internal governance mechanisms that adapt ESG decisions to the firm’s strategic stage, enhancing overall investment efficiency.
These conclusions are further supported by recent empirical evidence. For instance, Puntaier et al. (2022) argue that board diversity is not merely symbolic but a strategic necessity in entrepreneurial and high-growth firms [
42]. Their study shows that diverse boards make more disciplined investment decisions under uncertainty. In the ASEAN context, Abdelzaher and Abdelzaher (2019) found that post-crisis periods saw increased effectiveness of female directors, particularly in capital-intensive sectors requiring forward-looking governance [
50]. Additionally, Mastella et al. (2021) demonstrate that board gender diversity in Brazilian firms is significantly associated with improved ESG and risk management outcomes in volatile industries—offering further support for lifecycle-contingent governance effects [
51].
Control variables such as the operating cycle (OC) and sales volatility (STDSALES) show stronger effects in growth state firms, with significant coefficients of 0.025 (t-statistic = 4.21, p < 0.01) and 0.304 (t-statistic = 6.53, p < 0.01), respectively, indicating their critical roles in these dynamic settings. Dividends (DIV) and leverage (LEV) are also significant, with smaller yet impactful coefficients.
Collectively, these findings support Hypothesis 2, confirming that the positive effect of female executives on ESG investment efficiency is more pronounced in growth firms compared to steady-state firms. This reinforces the interpretation that female leadership not only enhances governance quality in general but also plays a particularly important role in dynamically evolving corporate contexts.
4.3. Additional Analyses
4.3.1. Alternative Proxy for Female Executives
In
Table 6, the study explores the impact of an alternative measurement of female executive representation, denoted as FEMALE2, on ESG investment efficiency. This alternative proxy measures the ratio of female executives to the total number of board members, providing a refined perspective on the influence of gender diversity within corporate governance structures on ESG investment efficiency.
The regression results from Panel A reveal a low baseline level of ESG investment efficiency across the firms, as indicated by a significantly negative intercept of −1.513 (t-statistic = −15.12, p < 0.01). The coefficient for FEMALE2 is −0.095 (t-statistic = −1.51), suggesting that this alternative proxy does not significantly enhance ESG investment efficiency, while ABESG shows a significant positive effect (coefficient = 0.185, t-statistic = 15.36, p < 0.05), affirming that deviations from normal ESG investment behavior correlate with higher subsequent ESG scores. Notably, the interaction term FEMALE2 × ABESG is significantly negative (coefficient = −0.108, t-statistic = 64.93), indicating that an increased proportion of female executives can effectively moderate excessive ESG investments.
Further examination in Panel B assesses the role of the corporate lifecycle, comparing steady and growth states. In steady state firms, the negative influence of FEMALE2 is minimal (coefficient = −0.036, t-statistic = −0.41), while in growth state firms, a slightly positive but non-significant effect is observed (coefficient = 0.052, t-statistic = 0.72). The interaction term FEMALE2 × ABESG in growth state firms is significantly negative (coefficient = −0.628, t-statistic = −2.58, p < 0.05), suggesting that a proportional representation of female executives is particularly effective in reducing overinvestment in ESG initiatives in dynamically growing firms. This nuanced understanding supports the broader narrative that the impact of female executives on corporate behavior varies significantly depending on how female presence is quantified and the stage of the corporate lifecycle.
4.3.2. Impact of Female Executives on the Efficiency of Segmented ESG Investments (E, S, G)
Table 7 presents the empirical results exploring how female executive representation influences the efficiency of investments across the specific components of ESG: Environmental (E), Social (S), and Governance (G). The adjusted R-squared values and the significant F-values for each model underline the robust explanatory power and the robustness of these results across the ESG spectrum. The regression analysis highlights different dynamics in each area, reflecting the nuanced roles female executives play within these domains.
For the environmental component, a significantly negative intercept (−9.577, t-statistic = −77.14, p < 0.01) indicates a generally low baseline level of investment efficiency. Female executives, however, are found to positively influence environmental outcomes with a coefficient of 0.022 (t-statistic = 2.01, p < 0.05), suggesting that their presence can enhance environmental practices. The positive coefficient of 0.247 for abnormal ESG investment (ABESG) (t-statistic = 2.24, p < 0.05) indicates that deviations from typical investment behaviors correlate with better environmental efficiency. Nevertheless, the interaction of female presence and ABESG shows a mitigating effect on excessive investments (coefficient = −0.104, t-statistic = −1.74, p < 0.1).
In the social domain, the analysis reveals a negative intercept (−4.719, t-statistic = −40.48, p < 0.01), with female executives having a slightly positive but less pronounced effect on social investment efficiency (coefficient = 0.018, t-statistic = 1.73, p < 0.1). Similarly to the environmental component, ABESG positively impacts social efficiency (coefficient = 0.131, t-statistic = 1.75, p < 0.1), with female executives again playing a role in moderating excessive social investments (coefficient = −0.085, t-statistic = −1.94, p < 0.1).
Regarding the governance component, the analysis starts with a positive intercept (1.015, t-statistic = 8.73, p < 0.01), indicating higher baseline efficiency in governance practices. However, the presence of female executives does not significantly affect governance efficiency (coefficient = 0.008, t-statistic = 0.80), nor does the interaction with ABESG (coefficient = −0.007, t-statistic = −0.80), suggesting that governance practices may be less susceptible to influence from gender diversity on the board compared to environmental or social investments.
Overall, the findings demonstrate the variable impacts of female executives on different ESG components, with more significant effects observed in environmental and social areas than in governance. This comprehensive analysis underscores the importance of considering specific ESG components when assessing the impact of female executive representation on corporate sustainability practices.
4.3.3. Addressing Endogeneity Issues
Although the research design incorporates a lagged structure—where the key independent variables are measured at time t and ESG investment performance (FESG) is observed at t + 1—to mitigate reverse causality concerns, further steps were taken to address potential endogeneity issues, particularly selection bias.
To supplement the primary model, a Propensity Score Matching (PSM) analysis was conducted. Matching was performed based on firm size using nearest-neighbor matching with a caliper width of 3%, in accordance with the matching strictness recommended by Lawrence et al. (2017) for governance-related variables [
52]. This approach was chosen to ensure that treated and control firms are similar in terms of observable characteristics that could jointly influence both female executive presence and ESG decisions.
After generating a matched dataset, the main regression model (Equation (1)) was re-estimated.
The estimation results in
Table 8 show that although the coefficient on the FEMALE variable becomes statistically insignificant, the interaction term between FEMALE and ABESG remains in the expected direction and is statistically weak but consistent. These results suggest that the core findings are not an artifact of biased sampling but rather reflect structural relationships.
Taken together, the combination of (1) a lag-based model design, and (2) a tightly calibrated PSM approach collectively reinforce the causal interpretation of the main findings. The concern of selection bias, though legitimate, does not appear to undermine the validity of the observed effects. This additional analysis helps shift the focus of the paper back to its main contribution—understanding how female executive representation contributes to ESG investment efficiency.
4.3.4. Different Classification of Life Cycle
To address concerns about the unidimensional classification of the firm life cycle based solely on R&D intensity, this study incorporates an alternative classification method that uses a broader set of financial indicators, following the methodology outlined in Kim et al. (2021) and Hwang and Choi (2022) [
53,
54].
Specifically, five firm-level indicators were used: sales growth rate, asset growth rate, fixed asset ratio, operating cash flow ratio, and profitability (net income to total assets). For each firm-year, five-year median values were calculated to smooth year-to-year fluctuations. Each median value was then divided into quintiles by year and scored from 1 to 5. The sum of these scores—ranging from 5 to 25—was used to assign firms into one of five life cycle stages. To align with the design of the main analysis, this section focuses on three key stages only: growth (score ≤ 5), maturity (score 11–15), and decline (score ≥ 20), excluding intermediate groups.
The purpose of this analysis is to provide a robustness check for the main findings, rather than to retest or reinterpret Hypothesis 2, which is formally stated and tested in
Section 4.2.
While the results largely corroborate those from the R&D-based classification, they also emphasize that the governance effect of female executives on ESG investment efficiency remains most pronounced in growth-stage firms.
The results of the alternative classification analysis are broadly consistent with the main findings. In growth-stage firms (n = 946), the interaction term between female executive representation and abnormal ESG investment is negative and marginally significant (β = −0.1064, p = 0.0528), indicating that female executives play a corrective role in limiting ESG overinvestment. By contrast, in mature-stage (n = 1313) and declining-stage (n = 594) firms, the coefficients for this interaction term are statistically insignificant and show mixed directions, suggesting that the governance effect of female leadership is more pronounced during earlier stages of development.