BGD and environmental performance

Gender diversity is defined as the presence of female directors on corporate boards (Carter, Simkins, & Simpson, Reference Carter, Simkins and Simpson2003). Resource dependence theory postulates that gender diversity on corporate boards increases board effectiveness by bringing new insights, expertise, innovation, and strategic decision-making (Pfeffer & Salancik, Reference Pfeffer and Salancik1978), leading to improved overall performance (Apesteguia, Azmat, & Iriberri, Reference Apesteguia, Azmat and Iriberri2012). Hillman, Cannella, and Paetzold (Reference Hillman, Cannella and Paetzold2000) argued that women's presence on corporate boards provided many benefits and helped firms understand diverse perspectives, thus accessing better resources. Ben Barka and Dardour (Reference Ben Barka and Dardour2015) postulated that female directors were more likely to serve on several boards, implying that they had a larger social network through board interlocks which increased awareness and distribution of actions. Prior literature showed that women were more rigorous and involved in board effectiveness and monitoring activities (Adams & Ferreira, Reference Adams and Ferreira2009), more cautious in decision-making, and more sensitive to social and environmental issues (Liao, Luo, & Tang, Reference Liao, Luo and Tang2015). Research on gender diversity also indicated that boards with female members had higher levels of environmental concerns (Kassinis et al., Reference Kassinis, Panayiotou, Dimou and Katsifaraki2016), were actively involved in pro-environment behaviours (Li, Zhao, Chen, Jiang, Liu, & Shi, Reference Li, Zhao, Chen, Jiang, Liu and Shi2017), and had an excellent perception of environmental risks (Davidson & Freudenburg, Reference Davidson and Freudenburg1996).

However, evidence on the relationship between BGD and environmental performance is mixed. For instance, research suggested that BGD was positively associated with environmental performance (Kassinis et al., Reference Kassinis, Panayiotou, Dimou and Katsifaraki2016), and that the presence of women on corporate boards brought in different opinions and enhanced discussion quality to assess environmental-related decisions (Post, Rahman, & McQuillen, Reference Post, Rahman and McQuillen2015). Sampling 94 Fortune 500 firms, Ciocirlan and Pettersson (Reference Ciocirlan and Pettersson2012) showed that firms with more female boardroom representations were more concerned about climate change and environmental performance. Walls, Berrone, and Phan (Reference Walls, Berrone and Phan2012) supported their findings, discovering that one of the main reasons for less environmental efficiency was boards becoming more independent, bigger, and less diverse. McGuinness, Vieito, and Wang (Reference McGuinness, Vieito and Wang2017) sampled Chinese joint-stock firms, revealing that firms with more female directors were associated with higher corporate environmental performance. Glass, Cook, and Ingersoll (Reference Glass, Cook and Ingersoll2016) examined a sample of US firms, showing that firms with more female directors were more effective in implementing sustainability strategies. Shoham, Almor, Lee, and Ahammad (Reference Shoham, Almor, Lee and Ahammad2017) sampled cross-country firms, revealing that only one woman on the corporate board positively encouraged firms towards environmental sustainability, regardless of the local culture. Birindelli, Iannuzzi, and Savioli (Reference Birindelli, Iannuzzi and Savioli2019) examined 96 listed banks and found that gender diversity on corporate boards exerted a positive influence on firms' environmental performance, and that women leaders were essential drivers of environmental sustainability and were increasingly involved in environmental issues. Lu and Herremans (Reference Lu and Herremans2019) analysed the industry effect, showing that BGD in more environmentally impacting industries was associated with higher scores in environmental performance; however, such impact was insignificant in less environment-impacting sectors, including retail and services. Orazalin and Baydauletov (Reference Orazalin and Baydauletov2020) sampled firms from 10 European countries, revealing that BGD was positively related to environmental performance, reinforcing the notion that gender diversity fostered sustainable development. Sampling US firms, Cordeiro, Profumo, and Tutore (Reference Cordeiro, Profumo and Tutore2020) showed similar findings: the higher the proportion of women on a corporate board, the better the firm's environmental performance. Nadeem, Bahadar, Gull, and Iqbal (Reference Nadeem, Bahadar, Gull and Iqbal2020a) sampled a large data set of publicly listed firms in the US and revealed a positive relationship between BGD and environmental innovation. Furthermore, the authors found that this association was more prominent in firms that were less profitable and environmentally sensitive. Using a sample of publicly listed firms in the UK from 2007 to 2017, Nadeem, Gyapong, and Ahmed (Reference Nadeem, Gyapong and Ahmed2020b) showed that in addition to economic returns, BGD significantly boosted firms' social and environmental value creation. More recently, Nuber and Velte (Reference Nuber and Velte2021) analysed a cross-country sample of European firms from 2009 to 2018, showing a robust, linear, and positive relationship between BGD and environmental performance.

Inconsistent with these results, other studies showed that the presence of women on corporate boards negatively impacted environmental performance. Deschênes et al. (Reference Deschênes, Rojas, Boubacar, Prud'homme and Ouedraogo2015), for instance, reported that the presence of women on corporate boards was negatively related to environmental performance. Alazzani, Hassanein, and Aljanadi (Reference Alazzani, Hassanein and Aljanadi2017) found that Malaysian female board members were more concerned with social issues than environmental ones. More recently, Nadeem (Reference Nadeem2021) sampled environmental violations and supplemental data on environmental projects and found that BGD might not nurture environmental initiatives, particularly in firms that had been found guilty of environmental violations. Another group of studies, however, failed to find any meaningful impact of BGD on environmental performance. For instance, Galbreath (Reference Galbreath2011) and Boulouta (Reference Boulouta2013) revealed that the presence of women on corporate boards was not directly associated with a firm's environmental quality. These contradictory results warrant further empirical investigation into how female directors contribute to environmental efficiency. In 2011, Italy enacted BGD quotas, mandating that firms have one-third of female directors on the board by 2015. Therefore, we believe that more women in the boardroom would positively affect environmental performance. Hence, we propose the following hypothesis:

Critical mass of female directors and environmental performance

The extant literature widely uses critical mass theory to explain how BGD affects board effectiveness (Schwartz-Ziv, Reference Schwartz-Ziv2017), firm performance (Joecks, Pull, & Vetter, Reference Joecks, Pull and Vetter2013), innovation (Torchia, Calabrò, & Huse, Reference Torchia, Calabrò and Huse2011), and investment in research and development (Saggese, Sarto, & Viganò, Reference Saggese, Sarto and Viganò2021). However, little research has been conducted on the critical mass of female directors and environmental performance. Torchia, Calabrò, and Huse (Reference Torchia, Calabrò and Huse2011) argued that the critical mass theory offered possible explanations for the influence of female directors on corporate issues. For example, in the boardroom, most male directors often dismissed or devalued the input of a minority of female directors. Nevertheless, when more female directors served on corporate boards, their influence increased. Accordingly, Konrad, Kramer, and Erkut (Reference Konrad, Kramer and Erkut2008) contended that a firm's approach towards environmental sustainability and disclosure seemed to be affected by the presence of a critical mass of female directors. Based on the critical mass theory, Nadeem (Reference Nadeem2020) revealed that firms with at least two female directors communicated better with investors through voluntary intellectual capital disclosure in initial public offering prospectuses than those with a single female director. As a result, the critical mass theory recommends that having an unbalanced board of directors hinders innovative decision-making (Konrad, Kramer, & Erkut, Reference Konrad, Kramer and Erkut2008) because social pressures inspire minority board members to embrace or adapt to the majority's opinions (Torchia, Calabrò, & Huse, Reference Torchia, Calabrò and Huse2011). Nevertheless, when a minority of female directors form a critical mass or reach a recommended threshold, interpersonal interactions improved, and information overload and environmental risks were mitigated (Nuber & Velte, Reference Nuber and Velte2021).

Consistent with the critical mass theory, several studies have established a positive association between the critical mass of female directors and higher levels of environmental performance. For instance, Post, Rahman, and Rubow (Reference Post, Rahman and Rubow2011) examined Fortune 1000 firms, showing that a critical mass of three female directors enhanced the environmental performance. Fernandez-Feijoo, Romero, and Ruiz-Blanco (Reference Fernandez-Feijoo, Romero and Ruiz-Blanco2014) posited that boards with at least three women favoured and enhanced CSR performance. Cook and Glass (Reference Cook and Glass2018) sampled Fortune 500 firms from 2001 to 2010 and found that the critical mass of female directors led to a greater commitment to CSR activities. Nuber and Velte (Reference Nuber and Velte2021) examined a sample of European non-financial firms from 2009 to 2018 and found that a critical mass of at least two female directors was to be reached to improve environmental performance. De Masi, Słomka-Gołębiowska, Becagli, and Paci (Reference De Masi, Słomka-Gołębiowska, Becagli and Paci2021) sampled Italian FTSE-MIB firms from 2005 to 2017 and revealed that female directors positively impacted ESG disclosure only if the critical mass was reached. The authors showed that the contribution of female directors was insignificant when the board consisted of only one or two women. Other recent studies also indicated that firms with a higher proportion of women on board were likely to demonstrate higher ability in promoting sustainability and environment-related strategies and practices (e.g., Birindelli, Iannuzzi, & Savioli, Reference Birindelli, Iannuzzi and Savioli2019; Orazalin & Baydauletov, Reference Orazalin and Baydauletov2020). In line with the existing literature, we believe that female directors' critical mass would allow them to wield sufficient influence to alter firm decisions and improve environmental performance. Thus, we propose the following hypothesis:

Sustainability committee and environmental performance

The board of directors is primarily committed to implementing long-term value creation strategies. The board's effectiveness depends not only on the board's composition but also its governance structure. In this regard, a sustainability committee's existence is one of the board attributes that influences sustainable efficiency (Ortiz-de-Mandojana & Aragon-Correa, Reference Ortiz-de-Mandojana and Aragon-Correa2015). Its aim is to systematically plan, implement, and review sustainability policies and activities (Liao, Luo, & Tang, Reference Liao, Luo and Tang2015). Carter, D'Souza, Simkins, & Simpson (Reference Carter, D'Souza, Simkins and Simpson2010) posited that the relevant specialised committees allow the board to reflect on interests and issues more carefully. Michaels (Reference Michaels2009) showed that establishing a sustainability committee signalled a firm's stakeholders about all relevant issues. Liao, Luo, and Tang (Reference Liao, Luo and Tang2015) documented that a sustainability committee improved employees' awareness of the environmental aspects of their jobs and responsibilities to reduce negative impacts. A sustainability committee assists corporate boards with the formulation and implementation of policies, principles, and practices to manage social and environmental risks, and achieve sustainable development goals (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018). Eccles, Ioannou, and Serafeim (Reference Eccles, Ioannou and Serafeim2014) argued that highly sustainable firms were more likely to establish separate board committees to achieve sustainable development. A sustainability committee makes social and environmental recommendations to the board of directors and helps board members in their tasks related to environmental concerns (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018). Thus, having a well-structured sustainability committee responsible for sustainability-related strategy and performance would facilitate a competitive advantage for the business.

Addressing environmental problems is considered a crucial challenge for a firm's success (Post, Rahman, & McQuillen, Reference Post, Rahman and McQuillen2015). Therefore, a responsible investor or stakeholder could also pressurise firms to establish a sustainability committee or promote initiatives to improve environmental performance (Liao, Luo, & Tang, Reference Liao, Luo and Tang2015). Stakeholder theory also suggests that to better accommodate stakeholders' needs, the board of directors should establish a sustainability committee that presumably focuses on promoting and implementing firm sustainability initiatives (Rodrigue, Magnan, & Cho, Reference Rodrigue, Magnan and Cho2013), and ensure the quality of the stakeholder engagement process (Michelon & Parbonetti, Reference Michelon and Parbonetti2012). Research shows that a board-level sustainability committee helps address CSR concerns (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018), efficiently tracks CSR strategies and policies (Liao, Luo, & Tang, Reference Liao, Luo and Tang2015), improves the efficacy of CSR strategies (Orazalin, Reference Orazalin2020), manages CSR-related risks and opportunities (Peters & Romi, Reference Peters and Romi2014), and expands the scope of sustainability disclosures accessible to stakeholders (Michelon & Parbonetti, Reference Michelon and Parbonetti2012).

Regarding empirical evidence, there is an ongoing debate on the relationship between sustainability committees and environmental performance. For instance, Berrone and Gomez-Mejia (Reference Berrone and Gomez-Mejia2009) examined longitudinal data from US firms and found no significant association between the existence of sustainability committees and environmental performance. Rupley, Brown, and Marshall (Reference Rupley, Brown and Marshall2012) sampled 127 US firms from 2000 to 2005 and found an insignificant relationship between sustainability committees and sustainability performance. Similarly, Rodrigue, Magnan, and Cho (Reference Rodrigue, Magnan and Cho2013) analysed environmentally sensitive firms, revealing that a sustainability committee could be established under a symbolic approach to manage stakeholder concerns, thus having less impact on environmental performance. More recently, Nadeem (Reference Nadeem2021) used environmental violations and supplemental data on environmental projects and showed that the existence of a sustainability committee might not foster environmental initiatives, particularly in firms that have been found guilty of environmental violations.

Other studies showed that the existence of a sustainability committee had a positive impact on a firm's environmental performance. For instance, Liao, Luo, and Tang (Reference Liao, Luo and Tang2015) reported that UK firms with environmental committees were more environmentally transparent than firms with no such committees. Elmaghrabi (Reference Elmaghrabi2021) used a sample of 100 non-financial firms from the Financial Times Stock Exchange (FTSE) from 2015 to 2017, demonstrating that firms with board-level sustainability committees had better CSR performance, better CSR strategy, and fewer controversies than firms without a sustainability committee. The author further showed that sustainability committees chaired by female directors exhibited better CSR performance and had lower CSR-related controversies. Kuzey, Uyar, Nizaeva, and Karaman (Reference Kuzey, Uyar, Nizaeva and Karaman2021) analysed three sectors (tourism, healthcare, and financial sectors) from the Thomson Reuters Eikon database from 2011 to 2018 and showed that it was imperative to establish CSR committees for tourism firms to enhance environmental performance. Similarly, Martínez-Ferrero, Lozano, and Vivas (Reference Martínez-Ferrero, Lozano and Vivas2021) sampled a group of firms located in Argentina, Brazil, Chile, and Mexico between 2012 and 2018 and found that the existence of a sustainability committee positively impacted a firm's sustainability performance. Other recent studies also showed that the presence of a sustainability committee heightened corporate transparency, especially concerning environmental information and above all, regarding the firm's environmental performance (e.g., Glass, Cook, & Ingersoll, Reference Glass, Cook and Ingersoll2020; Orazalin & Baydauletov, Reference Orazalin and Baydauletov2020). Given the documented positive relationship between sustainability committees and environmental performance, we expect the sustainability committee to positively influence environmental performance. Hence, we propose the following hypothesis:

Sample

This study used a data set of large firms listed on the Italian Stock Exchange from 2014 to 2018. Starting with a sample composed of all Italian publicly listed firms, we excluded firms that were not consistently listed during the sample period, firms with missing and duplicate data, and firms with missing environmental data. We also excluded financial firms, including banks and insurance companies, because of their specific governance and regulatory rules. Our final sample comprised 82 Italian listed firms covering 13 different industries, including energy and natural resources, health care, information technology, telecommunications, utilities, textile, retail, pharmaceutical, beverage and tobacco, fashion, tourism, manufacturing and construction, and automotive industries. The environmental data were retrieved from the Thomson Reuters Eikon ASSET4 database. This comprehensive platform provided applicable, consistent, and auditable data on the environmental, social, and governance dimensions of firms' performance. This database has increasingly been validated in the literature and has been widely used in recent studies (e.g., Kassinis et al., Reference Kassinis, Panayiotou, Dimou and Katsifaraki2016; Orazalin & Baydauletov, Reference Orazalin and Baydauletov2020). We also used the Bureau van Dijk Aida database to extract data related to the control variables in this study.

Variable's measurement

Independent variables

This study used BGD and sustainability committee as independent variables. In line with the existing literature, we employed different measures of female directors to capture gender diversity within corporate boards. First, we took the percentage of female directors on board (Glass, Cook, & Ingersoll, Reference Glass, Cook and Ingersoll2016). Second, we used the BLAU Index of heterogeneity, defined as;

$$BLAU = 1-\mathop \sum \limits_{i = 1}^n P_i^2 , \;$$

where ‘P_i’ is the percentage of board members in each gender and ‘n’ is the total number of gender groups (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018).Footnote ¹ Regarding the critical mass of women directors, we created three dummy variables: Critical Mass_1 (CM_1) equals ‘1’ if a firm has only one woman on the corporate board and ‘0’ otherwise; Critical Mass_2 (CM_2) equals ‘1’ if a firm has two women on corporate board and ‘0’ otherwise; Critical Mass_3 (CM_3) equals ‘1’ if a firm has more than two women on corporate board and ‘0’ otherwise.

Following prior studies such as Liao, Luo, and Tang (Reference Liao, Luo and Tang2015) and Biswas, Mansi, and Pandey (Reference Biswas, Mansi and Pandey2018), we measured sustainability committee as a dummy variable equal to ‘1’ if a firm has a sustainability committee and ‘0’ otherwise. A sustainability committee on corporate boards indicates that firms are more engaged in environmental sustainability (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018).

Quantile regression and model specifications

Quantile regression is a technique used to evaluate the conditional quantiles of a dependent variable in a linear model (Koenker & Bassett, Reference Koenker and Bassett1978). The traditional classical regression (OLS) assumes the average (i.e., conditional mean) relationship between the dependent and independent variables. Unlike OLS, quantile regression offers a more comprehensive picture of the potential relationship between the dependent and independent variables at different levels of conditional distribution (Koenker & Hallock, Reference Koenker and Hallock2001). Certainly, it can be used to predict the 25th or 75th percentile of the dependent variable or any percentile that the researcher is interested in. Quantile regression, according to Johnston and Di Nardo (Reference Johnston and Di Nardo2007), does not require any strict assumptions, unlike traditional linear regression of normality, homoscedasticity, and outliers. In addition, quantile regression is less prone to the impact of extreme data points. Thus, an essential rationale for employing quantile regression is to quantify a complete picture of the relationship between the dependent and independent variables.

Prior studies mostly used linear regression, relying on ‘mean’ as a measure of central tendency, and could not provide a complete picture of the relationship between BGD, sustainability committees, and environmental performance. Assuming that the impact of BGD and sustainability committees is different across the conditional distribution of firms' environmental performance, the mean effect might be zero. However, this can hide differences between firms with different levels of environmental performance scores. Hence, previous studies showed the positive, negative, or insignificant findings did not imply that the impact of BGD and sustainability committees on environmental performance was positive, negative, or insignificant for all levels of environmental performance. We believe that these relationships may vary as the dependent variable may differ in scale. Thus, we proposed a quantile regression method to re-examine this relationship. This method produced several coefficients, examining the impact of BGD and sustainability committees on environmental performance at different points of the conditional distribution.

We followed Buchinsky (Reference Buchinsky1998) and summarised the linear quantile regression model. Suppose (y _i, x _i), where x _i is a vector of independent variables that explains the dependent variable y _i and i denotes a sample of observations from the population. Assuming that the θ^thquantile of the conditional effect y _i is linear in x _i, we proposed the following model for conditional quantile regression.

(1)$$y_i = x_i^{\prime} {\rm \beta }_{\rm \theta } + u_{{\rm \theta }i}, \;$$

where y _i is the dependent variable, x _i is a k × 1 vector of the explanatory variables, β_θ is an unknown k × 1 vector of regression parameters related to the θ^thquantile, and u _θi is the unknown error term.

The θ^th conditional quantile regression of y _i assumed x _i would be as follows:

(2)$$Quant_{\rm \theta }\, ( {y_i/x_i} ) \equiv {\rm inf}\{ {y_i = F_i( {y/x} ) \ge 0} \} = x_i^{\prime} {\rm \beta }_{\rm \theta }$$

and

(3)$$Quant_{\rm \theta }( {u_{{\rm \theta }i}/x_i} ) = 0.$$

The θ^th quantile regression (0 < θ < 1) of y _i is the clarification of the minimisation of the sum of the absolute deviation residuals.

(4)$${\rm mi}{\rm n}_{{\rm \beta \theta }}\left\{{\mathop \sum \limits_{i\colon y \ge x_i^{\prime} {\rm \beta }} {\rm \theta }\vert {y_i-x_i^{\prime} {\rm \beta }_{\rm \theta }} \vert + \mathop \sum \limits_{i\colon y < x_i^{\prime} {\rm \beta }} ( {1-{\rm \theta }} ) \vert {y_i-x_i^{\prime} {\rm \beta }_{\rm \theta }} \vert } \right\} = {\rm mi}{\rm n}_{{\rm \beta \theta }}\mathop \sum \nolimits^ {\rm \rho }_{\rm \theta }( {u_{{\rm \theta }i}\,} ) .$$

The quantile regression model for our proposed research models generally took the following form:

(5)$$Quant_{\rm \theta }( {y_{it}/x_{it}} ) = {\rm \alpha } + {\rm \beta }_{\rm \theta }^{\rm ^{\prime}} BGD_{it} + {\rm \beta }_{\rm \theta }^{\rm ^{\prime}} SC_{it} + {\rm \beta }_{\rm \theta }^{\rm ^{\prime}} x_{it} + {\rm \mu }_{it}, \;$$

where y _it is the outcome variable at quantileθ. BGD _it, shows BGD variables, SC _it represents the sustainability committee, x _it vector includes board and firm characteristics, and μ_it is the error term. Specifically, we regressed environmental performance for BGD, sustainability committees, and their interaction effect on environmental performance. Hence, we showed all the models in the following ways. First, we examined the impact of BGD on environmental performance, as shown in Equation (6). Second, we investigated the effect of the critical mass of female directors on environmental performance, as shown in Equation (7). Third, we examined the impact of sustainability committees on environmental performance, as indicated in Equation (8). Finally, we explored the interaction effect of BGD and sustainability committee on environmental performance, as reported in Equation (9).

(6)$$Quant_{\rm \theta }( {EP_{it}/x_{it}} ) = {\rm \alpha } + {\rm \beta }_1BGD_{it} + {\rm \beta }_2BLAU_{it} + {\rm \beta }_3\mathop \sum \nolimits ControlVariables_{it} + {\rm \varepsilon }_{it}, \;$$

(7)$$Quant_{\rm \theta }( {EP_{it}/x_{it}} ) = {\rm \alpha } + {\rm \beta }_1CM\_1\;_{it} + {\rm \beta }_2CM\_2\;_{it} + {\rm \beta }_2CM\_3\;_{it} + {\rm \beta }_3\mathop \sum \nolimits ControlVariables_{it} + {\rm \varepsilon }_{it}, \;$$

(8)$$Quant_{\rm \theta }( {EP_{it}/x_{it}} ) = {\rm \alpha } + {\rm \beta }_1SC_{it} + {\rm \beta }_2\mathop \sum \nolimits ControlVariables_{it} + {\rm \varepsilon }_{it}, \;$$

(9)$$\eqalign{Quant_{\rm \theta }( {EP_{it}/x_{it}} ) & = {\rm \alpha } + {\rm \beta }_1BGD_{it} + {\rm \beta }_2CM_{it} + {\rm \beta }_3BLAU_{it} + {\rm \beta }_4SC_{it} + {\rm \beta }_5BGD \times SC_{it} \cr & \quad + {\rm \beta }_6BGD \times SC_{it} + + {\rm \beta }_7CM \times SC_{it} + {\rm \beta }_8BLAU \times SC_{it} \cr & \quad + {\rm \beta }_9\mathop \sum \nolimits ControlVariables_{it} + {\rm \varepsilon }_{it}, \;} $$

where EP is the firm's environmental performance, BGD refers to board gender diversity, CM_1 denotes one woman on board, CM_2 denotes two women on board, CM_3 denotes more than two women on board, BLAU denotes the gender diversity BLAU Index; as for subscripts, (i) refers to the individual firms, (t) refers to the time period, and (ε) is the error term (residual). The definitions of all variables are reported in Table 1.

Table 1. Definition of variables

Correlation matrix and summary statistics

Table 2 shows the descriptive statistics for all variables. The mean (SD) value of the environmental performance score was 72.90% (26.75%), with a minimum of 5.261% and a maximum of 98.58%. This was consistent with the European Commission's recent report on Italy's environmental performance score.Footnote ² Nevertheless, Birindelli, Iannuzzi, and Savioli (Reference Birindelli, Iannuzzi and Savioli2019) reported an environmental performance score of 67.9% for firms from Europe, Middle East, and Africa (EMEA) from 2011 to 2016. Biswas, Mansi, and Pandey (Reference Biswas, Mansi and Pandey2018), Baalouch, Ayadi, and Hussainey (Reference Baalouch, Ayadi and Hussainey2019), and Cordeiro, Profumo, and Tutore (Reference Cordeiro, Profumo and Tutore2020) reported 34.60%, 83.16%, and 50.68% for Australian, French, and US firms, respectively. The mean (SD) of BGD was 32.05 (14.25), indicating 32.05% female boardroom representations. This result reflected the impact of Italian Law 120 of 2011, which stipulates that one-third of corporate board directors must comprise women. The mean (SD) of the critical mass of women directors was 19.52 (23.21), 31.16 (24.33), and 50.01 (27.60), indicating that 19.52% of firms had only one woman on board, 31.16% of firms had two women on board, and 50.01% of firms had more than two women on board. The mean (SD) of sustainability committees was .63 (.311), showing that 63% of sample firms designated a sustainability committee on board. The mean (SD) of board size (BS) was 09.33 (4.631), with a minimum of six and a maximum of 22 board members, in line with Lipton and Lorsch (Reference Lipton and Lorsch1992), who argued that a board with nine members was ideal for monitoring purposes, and followed the best practices of corporate governance, wherein the ‘ideal board size should not be less than seven’ (Yermack, Reference Yermack1996). Of the total firms, 32.25% of firms represented independent directors on board, while 43.8% of CEOs also served as the chairman of the board of directors.

Table 2. Descriptive statistics

Note: This table reports descriptive statistics of all the variables used in this study. SD denotes standard deviation. Min denotes minimum. Max denotes maximum. All variables are defined in Table 1.

Table 3 presents the correlation matrix for all variables used in this study. We employed the Pearson correlation matrix to measure the direction of the linear relationship between the variables. The correlation results revealed a significant relationship between the main variables. Specifically, the results provided a preliminary indication that environmental performance was positively and significantly associated with BGD and sustainability committees. To rule out the presence of multicollinearity, we ran the variance inflation factor (VIF) for each estimated model. As a rule of thumb, multicollinearity was considered when the VIF of the explanatory variables exceeded 10 (Field, Reference Field2013). As shown in Table 3, the problem of multicollinearity was not a significant concern as it fell below the recommended threshold.

Table 3. Correlation matrix

Note: This table reports Pearson's correlation of the variables used in the regression of this study. All variables are defined in Table 1. VIF denotes the variance inflation factor. Significance level: ***p < .01, **p < .05, *p < .1.

Quantile regression results

Table 4 examines the impact of BGD on environmental performance and presents the quantile regression results. We conducted an empirical investigation by using Equation (6) for different values of quantiles of θ (i.e., the 25th, 50th, and 75th quantiles). In doing so, we explored the influence of explanatory variables at different levels of conditional distribution. We examined the differential effect of BGD on environmental performance, depending on the portion of firm distribution. We found that BGD indicators (i.e., BGD% and BLAU Index) were positively and significantly associated with environmental performance at different percentile levels. This confirmed that BGD affected not only the conditional mean but also the dispersion of environmental performance. Specifically, we showed that the coefficients for BGD at low quantiles (i.e., 25th) were lower than the median and upper quantiles (50th and 75th). This indicated that in each case, BGD variables (BGD and BLAU Index) had a larger quantitative impact in firms performing better environmentally. Overall, the results were consistent with the extant literature (e.g., Kassinis et al., Reference Kassinis, Panayiotou, Dimou and Katsifaraki2016; McGuinness, Vieito, & Wang, Reference McGuinness, Vieito and Wang2017; Orazalin & Baydauletov, Reference Orazalin and Baydauletov2020) and supported Hypothesis 1, which states that the presence of women on boards tended to pay more attention to the firm's environmental performance. However, the results contradicted the findings of Deschênes et al. (Reference Deschênes, Rojas, Boubacar, Prud'homme and Ouedraogo2015) and Alazzani, Hassanein, and Aljanadi (Reference Alazzani, Hassanein and Aljanadi2017), who found that female directors were more concerned with social issues than environmental issues. In addition, the results were in line with resource dependence theory and stakeholder theory, suggesting that a gender-diverse board was likely to represent diverse stakeholders (Rao, Tilt, & Lester, Reference Rao, Tilt and Lester2012), allowing firms to draw from a variety of experiences and education (Shaukat, Qiu, & Trojanowski, Reference Shaukat, Qiu and Trojanowski2016), and eventually providing a more accurate environmental performance disclosure (Hillman, Cannella, & Paetzold, Reference Hillman, Cannella and Paetzold2000).

Table 4. BGD and environmental performance

Note: This table reports the quantile regression results of environmental performance (EP) on BGD and a set of control variables. All variables are defined in Table 1. Standard errors are in parentheses. Significance level: ***p < .01, **p < .05, *p < .1.

Table 4 further shows that board size positively and significantly impacted environmental performance in all quantile estimations. This indicated that boards with enhanced workload environments and broader collective backgrounds and skills were better at improving the firm's environmental performance (Jizi, Reference Jizi2017). We found that firm performance had a negative and significant impact on environmental performance in the first percentile (25th) and a positive impact in the median and upper percentiles (50th and 75th). This indicated that firms with lesser profitability were discouraged from external forces, while firms with more profitability utilised the external resources to improve environmental performance. We revealed that firm size had a positive and significant impact on environmental performance in all quantile regressions. This postulated that large firms performed better than small firms in terms of environmental performance. This might also be because large firms could have more resources than small firms to handle environmental hazards (Li et al., Reference Li, Zhao, Chen, Jiang, Liu and Shi2017). We, however, failed to find any significant impact of board independence, CEO duality, and financial leverage on environmental performance.

Table 5 examines the relationship between the critical mass of female directors and environmental performance by using Equation (7) for different quantile values (i.e., 25th, 50th, and 75th). We employed three indicators to gauge the critical mass of female directors (i.e., CM_1, CM_2, and CM_3). We found that the relationship between firms with only one woman on their board (CM_1) and environmental performance was insignificant for all quantiles. However, it improved for firms with two women on their board (CM_2) and was even more significant for firms with more than two women (CM_3). The results showed that a board with a single female director was not enough to eliminate the evidence of tokenism. Therefore, it is difficult to prevent women from being categorised, stereotyped, and ignored by the majority of men, making them unable to make any contributions to the board. On the contrary, two or more women on corporate boards enabled their interaction and potential impact on corporate decisions, ultimately improving the environmental performance. These findings were consistent with those of previous studies (e.g., Cook & Glass, Reference Cook and Glass2018; Orazalin & Baydauletov, Reference Orazalin and Baydauletov2020; Post, Rahman, & Rubow, Reference Post, Rahman and Rubow2011) and supported the critical mass prediction of Hypothesis 2. The results showed that unless boards have a critical mass of two or more female directors, the influence of BGD on the corporate decision-making process was limited. Nevertheless, when this critical mass was reached, BGD had a positive effect on corporate response to stakeholders' demands to enhance the firm's environmental performance. Similarly, the critical mass theory suggested that one woman was a ‘token’, two women were a ‘presence’, and three or more women were a ‘voice’. It further stated that a higher proportion of female directors would be more willing to support each other, thus collectively having a more significant impact on corporate decisions (Torchia, Calabrò, & Huse, Reference Torchia, Calabrò and Huse2011). In addition, minorities on boards were usually considered merely a token; however, as their numbers increased to a majority, they influenced the firm's strategic decisions, including environmental performance.

Table 5. Critical mass of female directors and environmental performance

Note: This table reports the quantile regression results of environmental performance (EP) on critical mass (CM) of women directors and a set of control variables. All variables are defined in Table 1. Standard errors are in parentheses. Significance level: ***p < .01, **p < .05, *p < .1.

Table 6 shows the influence of sustainability committees on environmental performance by using Equation (8) for different values of θ (i.e., the 25th, 50th, and 75th quantiles). We found a positive and significant effect of sustainability committees on environmental performance at different percentiles (i.e., 25th, 50th, and 75th percentiles), obtaining different coefficient values. In particular, we found that the 25th and 75th percentile coefficients were lower in magnitude than the median (i.e., 50th percentile). The median percentile (50th percentile) showed a significant and strong effect. This indicated that a sustainability committee had a substantial and quantifiable influence on firms with a median environmental performance. These results were consistent with prior studies (e.g., Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018; Liao, Luo, & Tang, Reference Liao, Luo and Tang2015), which asserted that a sustainability committee led to better environmental performance. Nevertheless, these findings contradicted those of Chams and García-Blandón (Reference Chams and García-Blandón2019), who did not find any significant relationship between the sustainability committee and environmental performance. Michelon and Parbonetti (Reference Michelon and Parbonetti2012) posited that board committees were strategic dynamics for auditing, monitoring, and assessing governance performance, agendas, and policy-making. Likewise, a sustainability committee provided specialised counsel, involving the firm with strategic environmental constituencies and monitoring decision-making about environmental issues, and thus positively influencing environmental performance (Biswas, Mansi, & Pandey, Reference Biswas, Mansi and Pandey2018). Walls, Berrone, and Phan (Reference Walls, Berrone and Phan2012) argued that a board-level sustainability committee played a significant role in the firm's overall performance by giving directors specialised responsibilities and facilitating boards more effectively towards better environmental awareness. Moreover, the results were in line with the stakeholder theory and argued that the existence of a sustainability committee showed the firm's commitment to its stakeholders in terms of corporate social responsibility (Michelon & Parbonetti, Reference Michelon and Parbonetti2012). Firms were more likely to form a dedicated board committee on sustainability issues to better manage their stakeholder relationships, adequately address stakeholders' concerns, and demonstrate their commitment to responsible corporate practices. Hence, Hypothesis 3 was supported.

Table 6. Sustainability committee and environmental performance

Note: This table reports the quantile regression results of environmental performance (EP) on sustainability committee (SC) and a set of control variables. All variables are defined in Table 1. Standard errors are in parentheses. Significance level: ***p < .01, **p < .05, *p < .1.

To inspect the interaction effect of BGD and sustainability committees on environmental performance, we used environmental performance as a dependent variable and sequentially introduced the independent variables of BGD, sustainability committees, the control variables, and the interaction terms between BGD variables and sustainability committees into the regression model, as indicated in Equation (9). The results presented in Table 7 revealed that when the interaction term between BGD variables and sustainability committees were included, the main influence of BGD variables on environmental performance remained significant, but notably, at a higher magnitude than in Tables 4 and 5. We found a significant and positive interaction effect of BGD and sustainability committees on environmental performance. In particular, the interaction term between the percentage of female directors and sustainability committees (BGD% × SC) showed a positive and significant effect in all quantile regressions. This revealed that, on average, if the presence of female directors and sustainability committees in firms increased by one point, the firm's environmental performance would increase by 1.243 points. This indicated that BDG accentuated the positive effect of sustainability committees on environmental performance. The interaction term between the critical mass and sustainability committees (CM_3 × SC) had a positive and significant effect on all quantile estimations. This showed that, on average, if the critical mass and sustainability committees in firms increased by one point, the level of environmental performance would increase by 7.647 points. This showed that the critical mass of female directors reinforced the sustainability committee's positive effect on environmental performance. Furthermore, the interaction term between the gender diversity BLAU Index and sustainability committees (BLAU Index × SC) indicated a positive and significant effect on environmental performance in all quantile regressions. We further witnessed that an increase in female directors on boards and the presence sustainability committees increased environmental performance.

Table 7. BGD, sustainability committee, and environmental performance

Note: This table reports the quantile regression results of environmental performance (EP) on BGD, SC, and a set of interaction variables. All variables are defined in Table 1. Standard errors are in parentheses. Significance level: ***p < .01, **p < .05, *p < .1.

Consistent with the resource dependence theory, we found that female directors provided unique resources to firms by holding many key positions to monitor both formal and informal environmental activities, being actively involved in many environmental campaigns, and could transfer these related competencies to the board to boost the firm's overall performance. At the same time, we supported both the stakeholder theory and critical mass theory and revealed that more women on corporate boards led to environmental sensitivity and better policy-making because of their specific characteristics, ultimately improving environmental performance. These characteristics of women directors included a superior propensity to be ‘green’, more effective monitoring agents, more rigorous enforcement of ethical conduct, and a better likelihood to participate with multiple stakeholders and respond to their necessities – all of which result in greater environmental awareness.

Overall, we found that BGD and sustainability committees led to higher environmental performance. We showed that female directors enhanced board effectiveness and environmental performance by bringing diverse perspectives, knowledge, experience, and ideas to the boardroom (Schwartz-Ziv, Reference Schwartz-Ziv2017). Firms with a higher number of women on board were more persuaded to embrace advanced sustainability agendas and select long-term rewarding social and environmental activities (Jizi, Reference Jizi2017). We also provided empirical evidence that sustainability committees with a higher proportion of female directors were more effective for firms' environmental performance. Hence, firms with sustainability committees and more women on board tended to tackle a broader spectrum of financial and non-financial issues, facilitating sustainable practices to improve firms' environmental performance.

Robustness check

We conducted several additional tests to check the robustness of the results. The problem of endogeneity is pervasive in almost all governance studies (Wintoki, Linck, & Netter, Reference Wintoki, Linck and Netter2012), as well as BGD and performance studies (Adams & Ferreira, Reference Adams and Ferreira2009). To reduce the influence of endogeneity, this study considered the issue of self-selection bias, reverse causality and simultaneity, as well as dynamic endogeneity, which are considered the main sources of endogeneity (Muhammad, Migliori, & Mohsni, Reference Muhammad, Migliori and Mohsni2021). To this end, we used a series of estimation checks that included alternative measures, the Heckman two-stage model, instrumentation techniques, lagging structures, and dynamic panel data methods. Following Lara, Osma, Mora, and Scapin (Reference Lara, Osma, Mora and Scapin2017), we first implemented the Heckman two-stage model to mitigate potential self-selection problems and other forms of endogeneity. As shown in columns (1) and (3) of Table 8, in the first-stage regressions, we employed all those variables that potentially influenced a sustainability committee and a firm's decision to hire female directors. Specifically, we used a probit regression with a BGD and a sustainability committee dummy as outcome variables and all control variables from the main model as explanatory variables. We then computed the inverse mills ratio (IMR) from probit regressions and included it in stage-two OLS estimations (i.e., columns [2] and [4] of Table 8) to determine the association between BGD, sustainability committees, and environmental performance. We found that the coefficient for IMR in both regressions was statistically insignificant, confirming that the main results were not biased towards self-selection problems.

Table 8. Robustness check

Note: This table reports the Heckman two-stage model, two-stage least-squares (2SLS), and system GMM estimations of environmental performance (EP) on BGD, SC and a set of interaction variables. BGD%^ is the percentage of women instrumented with industry average gender diversity. All variables are defined in Table 1. Standard errors are in parentheses. Significance level: ***p < .01, **p < .05, *p < .1.

Second, we employed an instrumental variable regression (i.e., two-stage least squares [2SLS]) to account for possible endogeneity concerns. Wintoki, Linck, and Netter (Reference Wintoki, Linck and Netter2012) argued that endogeneity was pervasive in all corporate governance studies, and that the selection of board members was not an exogenous process. Rather, board attributes were endogenously chosen by firms to make their tasks and information uniform. The appointment of female directors on boards was also a choice made by firms and was, therefore, an endogenous decision (Sila, Gonzalez, & Hagendorff, Reference Sila, Gonzalez and Hagendorff2016). To execute the 2SLS method, we needed an exogenous instrument, a variable that was closely correlated to the explanatory variables but did not influence the outcome variable. Hence, we followed Adams and Ferreira (Reference Adams and Ferreira2009) and used the industry average of BGD (BGD^) as an exogenous instrument to deal with possible endogeneity issues. We justified using the BGD industry average as an instrument because BGD in a firm's industry could influence the firm's BGD but might not influence the firm's environmental performance.Footnote ³ The present governance variables might then be affected by past performance. Thus, we lagged all variables by one year to mitigate any possible simultaneity issues. The 2SLS results, reported in column (5) of Table 8, confirmed the consistent results.

Finally, we used a dynamic panel data approach with a two-step system generalised method of moments (GMM) and a one-year lag to instrument the lagged dependent variable (Arellano & Bond, Reference Arellano and Bond1991; Roodman, Reference Roodman2009).Footnote ⁴ The GMM estimator dealt with dynamic endogeneity by drawing instrumental variables from the data set itself. It accounted for time-invariant firm characteristics using first differences to transform regressors and remove any fixed firm-specific effect, and thus adjusted for any unobserved shock in the firm's fixed effect, which might distort the results. Roodman (Reference Roodman2009) posited that GMM could improve OLS or fixed-effect regression in three ways: first, by containing firm-fixed effects to account for unobservable heterogeneity; second, by allowing current corporate governance variables to be influenced by past performance; and third, by allowing the firm's past characteristics to be valid instruments to account for simultaneity. Hence, the GMM method was more robust in dealing with endogeneity issues (Wintoki, Linck, & Netter, Reference Wintoki, Linck and Netter2012). We re-estimated model (9), and the results are reported in column (6) of Table 8. Overall, the robustness checks offered consistent results and reaffirmed the findings of this study.