Impact of Board Gender Diversity on Corporate Social Responsibility and Irresponsibility: Empirical Evidence from France

: While prior studies have investigated the impact of corporate governance mechanisms on corporate social responsibility (CSR) commitment, researchers have scantly studied the poten-tially important relationship between board gender diversity and corporate social responsibility and irresponsibility separately. Drawing on the social role theory and feminist ethics, we hypoth-esize that board gender diversity is positively associated with CSR and negatively associated with corporate social irresponsibility (CSI). Here, we relied on a sample of French non-financial companies listed on the SBF 120 index between 2011 and 2016. Our results provide evidence on the positive impact of board gender diversity on CSR and the negative one on CSI. We show that women have a stronger impact on reducing CSI than on enhancing CSR. Our findings were robust to the different estimation methods. Our results show that our ﬁndings remain robust. Results show that board gender diversity is positively and signiﬁcantly related to CSP and CSR. However, gender diversity has a negative impact on CSI. Gender diversity seems to exert a higher impact on CSI than on CSR.


Introduction
The global financial crisis and the succession of financial scandals have shaken confidence in large companies and have revived debates on corporate social responsibility (CSR) and corporate governance [1,2]. The board of directors' role has expanded from an interest in shareholders to the whole stakeholders [3][4][5]. In recent years, a strand of literature has examined the link between the various governance mechanisms, particularly the characteristics linked to the board of directors and corporate performance in terms of CSR.
The composition of the board of directors is a crucial issue in corporate governance [6]. Theories of governance argue that the board of directors' composition determines its quality and influence its actions. Indeed, a diversified composition allows the board to integrate different cognitive and professional skills. Some authors consider that homogeneous groups tend to solve problems in the same way and, therefore, repeat assessment errors, which is less likely in heterogeneous groups [7]. Diversity is, therefore, appreciated as a source of wealth for the company.
Recently, companies are increasingly confronted with diversity management issues due to women's promotion in decision-making and governance bodies. The issue of the representation of women on boards of directors has attracted considerable interest from practitioners, policymakers, and researchers in recent years, which is due in part to their relatively low representation on boards of directors compared to their presence in the general population and the business world.
The academic literature has aimed to provide evidence on the economic and social consequences of board gender diversity. Most previous studies show a positive impact of gender diversity on financial performance [8][9][10] and corporate social performance (CSP) [2][3][4]. Although the majority of studies find that gender diversity on the board is likely to have a positive influence on corporate social performance [11], others show that its effect could be limited or even negative [12][13][14].
The corporate social performance of a firm includes both "doing good" through corporate social responsibility (CSR) practices and "doing bad" through corporate social irresponsibility (CSI) issues [15,16]. CSR has been broadly recognized as a highly desired firm action that benefits communities and helps firms achieve better business performance [17,18]. CSI is defined as "corporate actions that result in (potential) disadvantages and/or harm to other actors" [19] (p. 1932). In fact, companies can jointly engage in CSR and CSI [20]. Moreover, most rating agencies seek to assess a company's overall societal performance score based on a selection of measures covering social, environmental, and governance impacts. In addition, the assessment of environmental, social, and governance (ESG) practices by rating agencies contains positive and/or negative practices to represent, respectively, "strengths" and/or "concerns" regarding such practices.
Researchers have only recently begun to broaden their understanding of the impact of gender diversity on boards of directors on CSR activities through the inclusion of CSI. Nevertheless, CSI has received little attention from researchers and has scantly been examined in the literature. The impact of board gender diversity on CSR and CSI remains unresolved in the extant literature [2,21].
The present work aimed to explore the impact of women's participation on boards of directors on the corporate social performance of French companies. We drew on several studies to examine the role of gender diversity on CSR and expand it in several important ways. First, we distinguished CSR and CSI practices to explore which field women on boards impact the most. The impact of board gender diversity on CSR and CSI together has scarcely been examined. Second, we brought perspectives to the French context in the period following the adoption of the Quota Act.
The remainder of the paper is structured as follows: Section 2 provides the theoretical and empirical literature review concerning the links between gender diversity and CSR performance. Data collection and methodological approach are described in Section 3. Results and robustness checks are presented in Section 4.

Theoretical Framework and Hypotheses Development
From the resource dependency theory perspective [22], the board provides critical resources (knowledge, personal connections, or legitimacy) to the firm. Indeed, the board of directors does not only provide advice, guidance, and assistance to executives but represents a tool for creativity and strategic thinking on new growth opportunities. Therefore, the board's effectiveness depends on the resources, both human capital and relational capital, provided by the directors. Diversity is seen as a factor of expertise and improving the quality of decisions. Indeed, diversity means that the board of directors' heterogeneous composition makes it possible to expand the number of links in the network [23]. As a result, a board made up of directors with heterogeneous characteristics benefits from more fruitful links between the company and its environment.
Women's role in bringing specific resources to the board through widening access to the enriched and expended social and human capital is well established in the literature [1]. Increasing gender diversity can improve decision-making as a greater variety of perspectives and issues are examined, and more outcomes are assessed [24]. More women can also encourage more participatory communication among board members, assuming gender differences in leadership styles, as some existing research demonstrates that women have a more participatory, democratic, and communal leadership style than men [25]. Furthermore, women's psychological traits enable them to reduce information asymmetry for the stakeholders and the market [26]. In addition, having female directors on board can be recognized as a signal of compliance with stakeholder's expectations (market and regulators) [27,28]. Thus, it can be assumed that the board gender diversity implies an increase in CSP [14].
Wood [16] (p. 702) asserts that "managers are moral actors and are responsible for making choices about how to meet corporate social responsibilities . . . It seems likely that managers vary in their perceptions of choice and responsibility [29] and that personal and organizational characteristics might be related to these varying perceptions in ways that would help to more clearly express the conditions of corporate social responsibility".
According to the social role theory [30], both genders act following clichés and principles linked with their social role. Indeed, these principles may be considered as social standards expected by others and as an individual's character. The social role theory postulates that women are more compassionate than men, as females are known to be more empathetic in different cultures. Moreover, ethical decision-making depends on the characteristics of the practice under consideration. Jones [31] (p. 391) also notes that "the relative importance of personal and situational factors can vary considerably from one question to another". Therefore, comparisons of ethical gender perceptions should also take into account the characteristics of the assessed practices.
Here, we highlight the reasoning for viewing CSP as having both CSR and CSI dimensions. The responsiveness of women to social engagement should vary depending on the practice considered. Thus, women will not react the same way to good and bad social practices. In fact, nearly all ethical responsibilities and actions are identified or at the minimum connected with empathy levels [32]. Empathy has been defined as "an emotional reaction, including feelings of compassion, concern and tenderness-in helping people in need" ( [33], p.533). Empathy was considered the main mechanism of compassion [32]. Thus, the reactions based on empathy are stronger towards people experiencing "bad" situations than those who are in conditions that are "not wonderful" [32]. Therefore, gender-diverse boards may tend to engage more in social practices that induce greater empathy. As a result, higher board gender diversity implies higher responsiveness and a greater willingness to improve situations deemed absolutely bad.

Board Gender Diversity and CSP
Regarding the effect of gender diversity on CSP, the literature suggests a positive relationship. Several previous works [3,[34][35][36][37][38] point out that gender diversity on boards of directors is one of the determining factors of CSP. In addition, Landry et al. [39] found that the higher the board gender diversity, the more likely the company appears on the lists of the most ethical companies as well as those of the best corporate citizens. From this perspective, Shaukat et al. [40] suggest that companies with more independent and female directors are more likely to have a proactive and comprehensive CSR strategy.
Hussain et al. [41] studied the relationship between corporate governance and sustainability performance. They noted that the existence of more women on the board improves the company's sustainability performance and that the sustainable development committee plays a substantial positive role in companies' environmental and social performance. Based on a sample of Spanish firms belonging to the business services industry, Sánchez-Hernández et al. [42] focus on the internal side of the social responsibility of the firm. They suggest that the social and environmental potential and capabilities of employees are integrated into the social responsibility effort and lead companies to adopt a better social responsibility strategy and reinforce their competitive advantage. In the same vein, Droms, Hatch and Stephen [43] distinguish two branches of social responsibility, individual social responsibility, which focuses on the individual's perceptions of what he or she should do to help society, and corporate social responsibility, which focuses on perceptions the individuals have of the role that businesses should play in society. Their results show significant differences between the genders in the level of individual social responsibility they feel. Moreover, it results in considerable gender differences in their perception of the social responsibility that companies and organizations should generally assume. Essentially, they find that women have higher levels of internalized moral identity than their male counterparts, which leads to their perception of the role organizations play in society.
Hyun et al. [44] shed light on the role of independent female directors in CSR performance and showed that female independent directors might take CSR issues more seriously than their male counterparts not only because of their stronger moral orientations but also because they have reputational reasons for doing so. The authors used data on the board composition of a sample of Standard & Poor's (S&P) 1500 index companies over the period 2000-2009 and their CSR rating. They found that the number (and proportion) of female independent directors is positively associated with CSR ratings and that the strength of this relationship depends on the level of the company's consumer market direction. Using a sample of Italian non-financial listed companies, Romano et al. [45] concluded that greater board gender diversity has an overall positive impact on ESG performance and that chief executive officer (CEO) duality negatively moderates this relationship.
To emphasize the crucial role of gender diversity in the growth of the CSR approach and companies' reputation, Vacca et al. [46] conclude that board gender diversity increases companies' orientation to CSR disclosure. CEO gender has a positive influence on the relationship between corporate tax planning and CSR reporting. In the same vein, Lanis et al. [47] argue that women on the board of directors do not engage in tax planning activities. In addition, Kouaib and Almulhim [48] state that higher women's representation on boards is related to higher earnings quality. Triki Damak [49] confirms that board gender diversity exerts a negative impact on earnings management practice.
Although most previous studies find that gender diversity on the board is likely to have a positive influence on CSR-related performance [11], others show that its effect may be limited or even negative. A major adversity that has been widely identified in the literature is that women in senior positions are often confronted with gender biases or stereotypes, which restrict their ability to fully contribute to corporate strategy and control [50]. In this sense, some previous works studied the association between the ESG score and board diversity and revealed that gender diversity on the board has a negative effect on the ESG score [12,13].
Based on the above discussion, we state the following: Hypothesis 1 (H1). Board gender diversity is positively linked to CSP.

Gender Diversity, CSR, and CSI
The importance of board gender diversity for engaging in CSR activities is welldocumented. The debate on CSR is mainly about the additional contributions that companies make to the wellbeing of society. In particular, this is reflected in emphasizing the idea of "doing good". As a result, "doing well by doing good" is a frequently studied research topic showing that CSR in the sense of good practice can contribute to a fruitful interaction between companies and society. Nonetheless, companies not only have a social responsibility to do good but also a social responsibility to prevent harm and, therefore, to refrain from activities that disadvantage and/or harm others. Thus, CSR involves both adopting good practices and avoiding bad ones. However, both in theoretical and empirical studies, the issue of CSR associated with avoiding bad practices is rarely addressed explicitly. Boulouta [2] explains that when dealing with CSR issues, women exercise a gender-specific role as they demonstrate benevolent and empathetic behavior. She argues that gender diversity has a significant impact on negative social practices (CSI) and that the higher the board gender diversity is, the lower these practices are. Accordingly, gender diversity can have a positive impact on an overall CSP measure, but it depends on its construction. In particular, if the CSP measure is mainly constructed through the aggregation of negative social practices, then gender diversity will have a negative and significant impact on it. Negative social practices are seen as more "pertinent" than positive ones and provoke a stronger empathic reaction from managers. Consequently, these CSP mechanisms are much more associated with the women's stereotypical role and entail more interest from gender-diverse boards. Similarly, Zhang [26] studied the relationship between board gender diversity and CSP and found empirical evidence that a greater presence of women on boards is linked to better CSR performance. The author states that women tend to have certain psychological characteristics that may make them more responsive to stakeholder claims and thus increase their importance. Likewise, many researchers in psychology have argued that empathetic responses to help are related to the perceived "seriousness" of the situation [51,52]. Consequently, we presuppose that: Hypothesis 2a (H2a). Board gender diversity is positively associated with CSR and negatively associated with CSI.
Hypothesis 2b (H2b). Board gender diversity has a stronger impact on CSI than on CSR.

Sample
Our sample consisted of French firms listed on the SBF 120 index from 2011 to 2016. We relied on CSR scores provided by Thomson Reuters Datastream ASSET 4 database. Financial and accounting information was also obtained from the Thomson Reuters Datastream database. We hand-collected governance data from the reference documents or annual reports available on the firms' websites.
The ASSET 4 data cover only 96 of the 120 companies listed on the SBF 120, of which we excluded 15 companies operating in the financial sector due to differences in their account structures and governance systems.  • Control variables.
We included different variables that are likely to affect the relationship between gender diversity and corporate social responsibility. Board size (B-SIZE) is measured by the number of directors on the board. The board of directors is a competitive advantage source, and its size should be one of the main determinants of its effectiveness. An effective board will defend an organization's resources by reducing the problem of managerial opportunism. Previous literature suggests that larger boards are more likely to be diverse and have directors from different backgrounds, which may cause the company to pay more attention to social responsibility issues. We also considered CEO duality, i.e., cumulating the functions of the CEO and the chairman of the board of directors (DUAL), measured by a binary variable equal to 1 if the same person performs the functions of the CEO and the chairman of the board of directors and 0 otherwise [26,36,56]. We measured board member independence (INDEP) by the proportion of independent directors on the board. Moreover, we controlled for board cultural diversity (CULTURAL-DIV) measured as the proportion of international or intercultural (religion, race, etc.) representation on board and board-specific skills (SP-SKILLS) measured as the proportion of board members, who have either a strong financial background or an industry-specific background.
The existence of a CSR or sustainable development committee on the board of directors (CSRC) is measured by a binary variable taking 1 if the company has a sustainable development or social responsibility committee and 0 otherwise. The establishment of a CSR committee can have a positive impact on CSR [57]. Indeed, creating a CSR committee makes it possible to exploit the experience, skills, and knowledge of the committee members, which should play an important role in anchoring the CSR perspective in the strategic direction of the organization and translate it into tangible actions.
Leverage (LEV) is measured by the total net debt to common equity ratio. The reputation of the auditor (BIG4) is measured by a binary variable taking the value 1 if the company is audited by two of the big 4 audit firms (Deloitte, Touche, Tohmatsu, Ernst and Young (EY), KPMG, and PricewaterhouseCoopers) and 0 otherwise. In line with [44], we used research and development (R&D) intensity as an independent variable to control for the effects of a firm's R&D investment. This variable is measured by the ratio of R&D expenses to net sales. The former represents all direct and indirect costs associated with creating and developing new processes, techniques, applications, and products with commercial potential.
Finally, we also control for firm size as a determining factor in the practice of social responsibility. Indeed, large companies have more resources to setup efficient CSR systems. Firm size (SIZE) is measured by the natural logarithm of total assets.

Empirical Model
To test our hypotheses, we consider the following model: where corporate social practice scores are extracted from ASSET 4 (corporate social performance score, corporate social responsibility score, and corporate social irresponsibility score) of the firm i at the year t. We chose to lag the measurement of gender diversity because the impact of the presence of women on board is observable after at least a year.
To avoid the endogeneity problem, we estimated our models using the system generalized method of moments (GMM). Table 1 reports the descriptive statistics for all variables. The data show that the average CSP is 56.09%, with a minimum and a maximum of 22 and 92%, respectively. The average CSR score is 64.68%, while the CSI score is 52.41%. Regarding gender diversity, we notice that the proportion of women is 27.76% of the total number of directors, with the average number of directors is 13; then, the French companies in our sample are characterized by extended boards. This large number can be the source of richer discussions and, therefore, of higher quality decisions given the skills and experiences of the different members. However, it can cause increased agency conflicts. Independent directors represent 51.47% of the total number of directors. The proportion of companies having the same person occupying the positions of chairman of the board of directors and general manager is 58.44%. As shown in Table 1, the boards of directors of the companies in our sample are made up of directors from different cultures. Cultural diversity enriches the discussion on the boards. Furthermore, around 37% of directors have specific skills. In addition, around 81% of companies have a CSR committee on the board of directors, which shows the importance of CSR issues and responsible investment strategies in French companies. R&D expenses represent, on average, about 6% of net sales. Table 2 reports the difference between means of dependent variables (CSP, CSR, and CSI). To this end, we divided the sample into two groups based on the median of the independent variable (board gender diversity): The first group consists of firms with a proportion of women on board higher than the median for the whole sample, and the second is composed of firms having a board gender diversity lower than the median. As shown in Table 2, the comparison of means test of CSP and CSR exhibits a negative sign, as predicted, and statistically significant. However, it is positive and significant for CSI. Thus, these findings suggest that firms with higher CSP and CSR scores are more likely to have women on their boards than those with lower CSP and CSR scores. Furthermore, firms with low board gender diversity (below median) have higher CSI scores than firms with high gender diversity. These results are following our expectations. Table 3 presents the correlation matrix and variance inflation factor (VIF) testing for the existence of a multicollinearity problem between the explanatory variables. According to the correlation matrix, the correlation between the explanatory variables is not high. We note that gender diversity is positively correlated with measures of CSP and CSR and negatively correlated with the measure of CSI. Furthermore, all VIF values are less than 10, and the average of the VIFs is lower than 2. Thus, according to these findings, multicollinearity is not a concern.

Multivariate Analysis
The results of the estimations of the impact of board gender diversity on CSP, CSR, and CSI are illustrated in Table 4. Before estimating our model, we performed the specification tests and found that for the estimation with a one-period lag of the CSP measure, there is a second-order serial correlation(AR (2)). We redid the estimation by introducing two lags of the dependent variable. We then observed the absence of an AR (2) effect for the residuals. Otherwise, Sargan's tests confirm the validity of the instruments.
The results show that CSP depends on its past values. In addition, the percentage of women on the board of directors has a positive and significant impact on CSP. Our results confirm hypothesis H1. Specifically, they support the idea that women bring specific resources to boards, such as empathy and stakeholder orientation. These results corroborate those of previous studies [3,37,41,45].
Regarding our control variables, the results show that the number of directors, independence of board members, cultural diversity, and director's specific skills have no impact on the CSP. However, the CEO duality has a negative impact on it. These results are in line with those of Gul and Leung [58], suggesting that CEOs, who also serve as board chairmen, may reduce the board's ability to exercise effective control over management and, thus, negatively affect strategic choices. Moreover, the existence of a CSR committee has a positive impact on CSP. This last finding agrees with those of previous works according to which a committee is a governance tool that allows exploiting the experience, skills, and knowledge of committee members. The existence of a CSR committee can develop the CSR perspective in the strategic orientation of the company.
The assessment of the impact of gender diversity on CSR and CSI indicates that board gender diversity exerts a positive and significant impact on CSR and a negative and significant impact on CSI. These results confirm our H2a hypothesis. It can be explained by the fact that women attach importance to improving social performance through good practices. Simultaneously, they are careful to deal with negative events that affect the social performance of the company. In addition, we notice that the coefficient, which expresses the impact of gender diversity on CSI, has a greater absolute value than the coefficient on CSR. Consequently, our H2b hypothesis, which states that the impact of women is greater in terms of CSI than in terms of CSR, is validated. Our results are in line with gender role theory and feminist ethics and confirm the results found in the American context by Boulouta [2]. Indeed, women tend to react more strongly to problematic situations. They are more interested in reducing corporate irresponsible practices that can hamper the improvement of the image of the company even in the presence of a social commitment. Regarding the control variables, the cultural diversity of board members and R&D intensity exert a positive impact on CSR and a negative one on CSI. Firm size has a positive and significant impact on CSR and CSI, which indicates that a large company engages in good social practices but faces irresponsible practices jointly. In addition, we note that CEO-duality exerts a negative effect on CSI.

Robustness Checks
To test the robustness of our results, we re-estimated our models using Tobit's for panel data. This model belongs to the family of models with the censored dependent variable, i.e., continuous, however only observable over a well-determined interval. This is why these are models that are located halfway between models with qualitative variables and the linear regression model, whose endogenous variable is continuous and observable. The use of this method is motivated by the fact that our various explained variables (CSP/CSR/CSI) are censored variables (continuous and their values are between 0 and 1). The Tobit regression model is a frequently used tool to model censored variables in econometric research. Indeed, this model allows the regression of a censored variable while keeping the assumptions necessary for linear regression. Table 5 reports the estimates. Our results show that our findings remain robust. Results show that board gender diversity is positively and significantly related to CSP and CSR. However, gender diversity has a negative impact on CSI. Gender diversity seems to exert a higher impact on CSI than on CSR. For the Tobit estimate, the existence of a CSR committee attached to the board of directors has a positive and significant effect on CSP and CSR. However, the existence of a CSR committee has no significant effect on CSI. These results show that the committee's role is limited to pushing CSR engagement, not to preventing negative events (CSI). The intensity of R&D helps companies to limit CSI but has no effect on CSR. Moreover, companies audited by BIG 4 show higher CSP and CSR scores.

Conclusions
The present study extends the literature on the role of board gender diversity and particularly on its impact on CSP. It does not only investigate whether board gender diversity has an impact on CSP but also whether it brings different perspectives to boards and, therefore, demonstrate that its effect may depend on the different corporate social practices while distinguishing corporate social responsibility and irresponsibility. To this end, we drew on the social role theory [30] and feminist ethics [31]. Using a sample of French companies listed on the SBF 120 index for the period spanning 2011-2016, our results reveal that board gender diversity exerts a positive and significant impact on CSP. Furthermore, gender diversity has a positive effect on CSR and a negative one on CSI. Therefore, it is important to note here that women exert a higher impact on reducing CSI than improving the CSR score. The findings of this research are in line with the gender role theory and prior literature, according to which women tend to put in more effort to prevent bad situations [2,31]. The need to prevent CSI is more important and high-priority than the need to develop CSR practices. Therefore, this area of CSP is more strongly correlated with the stereotype of women's role and requires more attention in future research.
Although the literature on the impact of board gender diversity on corporate social performance is internationally large, it is tight in the French context. For instance, Yassen et al. [59] investigate the impact of board gender diversity on five proxies of CSR (community score, employees score, environment score, governance score, and ESG score) during the period 2012-2015. They support the positive relationship between board gender diversity and CSR. Others, such as Ben-Amar et al. [60] and Nekhili et al. [61], focus on the impact of board gender diversity on the relevance of voluntary CSR reporting. Moreover, Beji et al. [62] highlight that diversity on boards of directors of French firms is generally positively associated with CSP.
In addition to elucidating the literature examining the impact of board gender diversity on CSP, our paper elucidates the literature examining the determinants of CSP in which the internal side of social responsibility or internal firm characteristics that could have an impact on different corporate social practices have been mostly ignored.
This study has theoretical and managerial motivations. Theoretically, our study shed light on the effect of women's presence on the board of directors on the CSP by revealing areas of stronger influence through distinguishing CSR and CSI practices. Looking at the impact of board gender diversity through a gender lens may better elucidate the mixed results in the existing literature by explaining the prioritization of women's social issues on the board.
On a practical level, this study helps to understand the dynamics affecting a company's social performance and can be useful to managers in guiding the social performance of their firm strategically. Our study also contributes to the current "push" for greater gender diversity of boards by regulators and policymakers by demonstrating some of the benefits of board gender diversity.
Author Contributions: S.B. designed the research, collected and analysed data and wrote the manuscript, and A.O. supervised, revised the research and paper. All authors have read and agreed to the published version of the manuscript.