Board/Executive Gender Diversity and Firm Financial Performance in Canada: The Mediating Role of Environmental, Social, and Governance (ESG) Orientation

The objective of this study is to verify the effect of gender diversity on the board of directors (BD) and the executive committee (EC) of participating Canadian firms with regard to the financial performance and the mediating role of environmental, social, and governance (ESG) orientation in this relationship. The study sample was composed of 133 Canadian firms, and the data cover an 18 year timeline (2002–2019), with 925 observations. This paper provides empirical support for the effect that gender diversity in turnover has on the financial performance of firms and explains 53% of its variance. In addition to supporting the beneficial effect of gender diversity on performance, the study reveals the mediating mechanism through the ESG orientation of companies explaining almost 4% of the total effect of gender diversity on performance. By analyzing two levels of diversity, the study revealed the superiority of the effect of gender diversity in BDs as compared to ECs. We discuss the theoretical and empirical implications of the results found, as well as the limitations and future prospects of research on the subject.


Introduction
Governance has been the subject of growing interest within both public administrations and private companies, particularly those listed on the stock exchange. With the rise of market globalization, trade liberalization, the arrival of the 4.0 industry and the 5th generation mobile network (G5), citizens' expectations, and the scandals that have occurred in Europe (e.g., Vivandi, Suissair) and North America (Enron, WorldCom), this interest has increased in the last two decades. Corporate governance refers to the art of governing, i.e., the ability of a company's executive committee to assume responsibilities such as the management, direction, and control of a company. However, good governance is first and foremost the responsibility of the board of directors (BD) in terms of decision-making, which affects the life of the company, major strategic orientations, identification of potential risks facing the company, benchmarking of the company against others in its sector, determination of efficiency indicators, evaluation of CEOs, sharing of information, and transparency.
In fact, a BD should not only be the outcome of a regulatory system (e.g., Canada, France, Spain, Norway) or the manifestation of corporate conformism, but a true value creator in terms of principal-agent problem solving (agency problems) [1] and strengthening the social legitimacy of the company and its performance, particularly in the private sector. Given the variety of skills and the scope of knowledge that board members should have in order to better govern in an increasingly complex business context, a number of studies have focused on the effectiveness of a board through the effect of some of its endogenous variables, such as diversity [1], i.e., the variety of the profiles of the board members in terms of age, gender, seniority (as a board member), nationality, education, experience or personality.
The current debate on the issue of diversity on the board of directors is concentrated around two positions. The first, which is more ethical, argues for diversity on the board because it is the right thing to do. Countries such as Canada, France, Spain, and Norway have legislated to ensure diversity by, for example, imposing a quota (or parity) for women on the board of directors. The second, which is more utilitarian, analyzes diversity through the prism of financial performance and sees diversity on the board of directors as a lever for creating value for the company and its stakeholders. It is probably the introduction of a legal constraint to appoint women to the board of directors and the lack of conclusive evidence that has sparked research on the impact of gender diversity.
In principle, one can subscribe to the idea that gender diversity allows board members to grasp the framework of a company's business context from a variety of perspectives and cognitive frameworks. Gender diversity on the board also has the potential to improve the quality of information provided by board members to the executive committee [2]. However, this idea has much to gain when supported by empirical studies linking diversity and performance. Despite the number of studies on the relationship between gender diversity and performance, the current state of the literature does not allow the resolution of this issue or the creation of a consensus around a position affirming or denying this relationship. In fact, certain studies show that BDs that include more women tend to generate a higher return on assets [3,4] and lead to a positive reaction on the stock market [5]. Conversely, in other studies, gender diversity in the boardroom reduces accounting performance [6,7] and leads to a loss in the value of the share [8]. Several other studies have found no relationship between women's representation on the board of directors and corporate performance [9][10][11]. As an example, using seven cases, in D'Souza, Simkins, and Simpson's study [9] regarding the relationship between gender diversity and corporate financial performance, it was shown that there was a positive relationship in two cases, no relationship in three cases, and a negative relationship in two cases. Based on the study by Hermalin and Weisbach [1], other more recent literature related to the affirmation of the relationship between gender diversity on the board and performance has tried to understand actions related to the presence of women on the board and to explain how diversity influences performance. In other words, beyond the question of the existence of a diversity-performance relationship, this literature attempts to understand the underlying mechanism explaining this relationship.
The present paper aims to reconcile these inconsistent results by addressing several gaps in the existing literature and capitalizing on recent developments on the subject. More specifically, this paper offers several contributions to the knowledge on the relationship between gender diversity in the boardroom and corporate financial performance. First, capitalizing on the fact that the existing literature on this subject covers companies in several countries on five continents, and that regulatory conditions and the socio-cultural context could explain the mixed results, this study examines only Canadian companies. Looking at a single country, Canada, isolates the confounding effects of factors related to the legislative framework and national, cultural, and social contexts that could increase or decrease the strength of the relationship between gender diversity in the boardroom and financial performance. Second, instead of considering Canadian companies at a specific time in their growth cycle, this paper extends the timeline to 18 years (2002-2019), thus avoiding situations that are contingent on a particular period in the company's performance or the composition of its board. Third, this paper makes an empirical contribution to the debate on issues related to the integration of women in corporate decision-making bodies. Fourth, in contrast to several previous, more descriptive studies, this study explores the intermediate mechanisms (environmental, social, and corporate governance (ESG) orientation) or the "black box", linking gender diversity to the corporate board and financial performance. In fact, we are not only looking at the direct relationship between female representation on the board and financial performance, but also at the board's orientations as mediators in this relationship. More specifically, the mediating role of the ESG (environmental, social, and corporate governance) orientation is examined through the relationship between gender diversity and corporate financial performance. ESG orientation thus represents a proxy for the influence of women on the board of directors. Fifth, the gender diversity analysis is broadened to include gender diversity on both the board and the executive committee (EC). This double analysis makes it possible to compare the effects of diversity at two levels of women's action, the board and the executive committee.
Thus, the objective of this study is twofold. The first concerns the combined effect of gender diversity on both the board and executive committee with regard to the financial performance of participating Canadian firms over the period 2002 to 2019. The second is to verify the mediating effect of ESG orientation on the relationship between gender diversity (on the board of directors and the executive committee) and financial performance.

Theoretical Background
Studies on corporate governance mobilize diverse and varied theoretical frameworks, but the effectiveness of a board of directors (BD) remains the central question for most of these frameworks. Thus, for so-called contractual theories, such as the agency theory [12] and the stakeholder theory [13], the effectiveness of a board lies in its ability to fulfill its contract. Therefore, the studies that are part of these approaches have tried to understand the determinants of the effectiveness of a BD through its characteristics (e.g., size, structure, composition). More specifically, for the agency theory, the BD is able to fulfill its contract (manage the agency relationship between the agents and the principal, i.e., to protect the interests of the shareholders when it has members (internal or external) with sufficient competence and independence. Stakeholder theory, however, broadens the scope of interest to include, in addition to the shareholders' interest (financial perspective), the interest of other contractors (partnership perspective), such as employees and the community (stakeholders). Thus, according to this theory, board members are expected to protect the interests of the various stakeholders; these stakeholders influence and are influenced by the board's decisions. The stakeholder theory provides a broader framework for explaining the composition of a board and the diversity of its members.
Another current view of so-called strategic theories takes an approach that goes beyond the control role of the BD through the reward-punishment mechanism of managers. In fact, strategic theories, such as resource dependency theory and cognitive theories, consider board members as active actors in the creation of value in several ways. For the resource dependency theory, firms operate in an open system and are forced to exchange with their environment for the acquisition of certain resources. Thus, board members contribute in terms of reducing the risks of access to resources (e.g., skills, relationships) that are critical and indispensable for the survival of the firm. According to this theory, the effectiveness of a BD lies in its ability to monitor but also to facilitate access to resources. For cognitive theories, board members make a strategic contribution through the variety of their knowledge, cognitive cues, and perspectives. Thus, the diversity of the board members guarantees access to a diversity of cognitive frameworks, and thus to multiple readings of reality and a wealth of strategic orientations. It is from this perspective that studies on the effectiveness of the board of directors find their theoretical basis, justifying the importance of diversity, including gender diversity, on the board of directors. However, cognitive theories face operationalization challenges, i.e., the difficulty of grasping cognitive frameworks and measuring the cognitive, interpretative and informational ("black box") processes mobilized by board members in the exercise of their role. Regarding this issue, the "Upper Echelons Theory" (UET) [14,15] provides an important theoretical foundation for studying the effect of the diversity of cognitive frameworks of board members on certain organizational outcomes, such as financial and extra-financial performance, through a proxy: gender diversity on the board [16]. Although the UET theory focuses primarily on senior management, several researchers have applied it to board members by equating the board with a high-level "supra-management team" [16,17].
"With the increasing call for board activism . . . , the influence of Boards over organizational outcomes will only grow. Perhaps best thought of as "supra-TMTs," Boards are an important target for strategic leadership research." ( [17] p. 11) "Although UET originally focused on top management teams, research has applied the theory to Boards of directors by likening Boards to "supra top management teams"." ( [16], p. 1548) In fact, the UET theory suggests two major principles: that managers act according to their interpretations of situations, and that these interpretations are a function of their experiences (seniority, function, education, international), their values, and their personality. In addition to these two principles, there are two ideas: (1) focusing on the characteristics of managers as a team provides a better understanding of organizational outcomes as compared to studies conducted only with a single manager (e.g., a CEO); (2) the demographic characteristics of managers can be used as valid, albeit incomplete, proxies for managers' cognitive frameworks. Empirically, the study has produced substantial evidence on the relationship between manager profile and organizational outcomes [18][19][20]. Others have examined the psychological and social processes that intervene between the demographic profile of leaders on the one hand and their behaviors on the other [21,22]. More recently, several studies have applied the UET theory to board members by studying their characteristics (e.g., gender, ethnicity) as proxies for their cognitive frameworks [23,24]. Based on the various elements presented above, this study mobilized the cognitive theory, stakeholder theory, and UET framework to study the effect of gender diversity in both the board and the executive levels on financial performance of firms in the Canadian context.

Empirical Studies on Gender Diversity and Firm Performance
Diversity research with regard to the board of directors (BD) has used the term "diversity" using two approaches, apparent or demographic diversity (gender, age, ethnicity), and non-apparent or cognitive diversity (e.g., knowledge, education, experience, values, personality). The measure takes three forms: a dummy variable, a percentage, or a Blau-index. This paper looks at the apparent gender diversity at the board and executive committee levels, i.e., the representation of women as indicated by the percentage of women on the board or executive committee. According to the UET theory, gender diversity is used as a proxy for cognitive diversity. With respect to the board of directors, Finkelstein and Hambrick [25] sketched the functioning of the board according to two functions strongly linked to the performance of the firm, namely decision-making with regard to strategic orientations and monitoring with respect to the interest of shareholders, the use of resources, response to threats, reward, and control of senior management.
Since the functioning of the board of directors is strongly linked to the performance of the company [26], it is quite legitimate to ask the question about the effect of the composition of the board of directors in terms of gender diversity on its functioning and, by extension, on the performance of the company. Studies regarding the relationship between gender diversity and organizational performance (accounting and stock market) are numerous, mixed, and do not lead to conclusive results. This observation remains the same for so-called meta-analytic studies [16,27].
In fact, Gulamhussen and Santa's [28] study, based on 461 observations, revealed a positive relationship between gender diversity on the board and its committees and both accounting (return on assets ROA and return on equity ROE) and stock market performance (Tobin's Q). This is the case as well with studies by Campell and Minguez-Vara [29] and Carter, Simkins and Simpson [30] on board gender diversity and Tobin's Q on a sample of 68 and 638 companies, respectively. Studies by Martin-Ugedo and Minguez-Vera [31], Ngyuen, Locke, and Reddy [32], and Brahma, Nwafor and Boateng [33] went in the same direction, using any of the performance measures ROA, ROE, or Tobin's Q. Other studies have also revealed the link between gender diversity and other outcomes, such as efficiency [34], increased surveillance [32,35,36], reduced level of risk [35,36], and increased strategic control [37]. However, according to Wellalage and Stuart [38], gender diversity leads to reduced business performance and increases agency costs. Other researchers have found no relationship between gender diversity and business performance [9][10][11]39,40].
Disparity of results is also present in meta-analytic studies. Indeed, Post and Byron's first meta-analysis [16] of 140 studies from 35 countries revealed that gender diversity is associated only with accounting performance (ROA, ROE) (r = +0.047, Confidence Interval CI [+0.033, +0.061] at 95%). Their results were not conclusive for stock market performance, with a weak and non-significant correlation (r = +0.014, 95% CI [−0.002, +0.031]). In other words, gender diversity explained less than 1% of the variance in accounting performance (profit, sales, return on investment). The second meta-analysis by Pletzer, Nikolova, Kedzior, and Voelpel [27], using 20 studies (published only in peer-reviewed academic journals) and covering more than 3097 companies from 1997 to 2012, showed the correlation between the percentage of women on the board of directors and financial performance (measured retrospectively by ROA and ROE and prospectively by Tobin's Q) was low and insignificant (r = +0.01, 95% CI [−0.04, +0.07]). In sum, the two meta-analyses revealed either no relationship or a very weak relationship between gender diversity on the board and financial performance (accounting or stock market). With an effect, if it exists, that is weak, it is very debatable to conclude that there is a relationship between the feminization of the board of directors and corporate performance. Table 1 captures the points of convergence and divergence on the relationship between gender diversity on the board of directors and financial performance according to different measures. Table 1. Summary of the literature review.

Reference
Results [28] Positive relationship between gender diversity on the board and its committees and both accounting (ROA and ROE) and stock market performance (Tobin's Q). [29,30] Board gender diversity and Tobin's Q on a sample of 68 and 638 companies, respectively.
[ [31][32][33] Similar to the above studies, using any of the performance measures ROA, ROS, ROE, or Tobin's Q. [34] Link between gender diversity and efficiency. [32,35,36] Link between gender diversity and increased surveillance. [35,36] Link between gender diversity and reduced level of risk. [37] Link between gender diversity and increased strategic control. [38] Gender diversity leads to reduced business performance and increases agency costs. [9][10][11]39,40] No relationship between gender diversity and business performance. [16] Gender diversity is associated only with accounting performance (ROA, ROE). [27] Correlation between the percentage of women on the board of directors and financial performance (measured retrospectively by ROA and ROE and prospectively by Tobin's Q) is low and insignificant.
Different elements may explain the divergence of these results. The first element lies in the very management of diversity. In fact, in some countries, gender diversity, or even gender parity, is a matter of compliance with the law and not the will or orientation of board members. Thus, some countries are more advanced than others in the management of gender diversity, and, therefore, in the feminization of the boards of directors of their companies. Women were under-represented in the majority of the samples studied. For example, in Pletzer, Nikolova, Kedzior, and Voelpel's [27] meta-analysis, the percentage of women on the board of directors did not exceed 37%, with an average of only 14%. Furthermore, international studies that include several countries do not differentiate between companies operating in an emerging market and those operating in an established market. Additionally, the lack of control over the economic (e.g., crisis period), political (e.g., law and jurisdiction), and social (e.g., women's participation rate, women's education levels) context of each country as well as the under-representation of companies from certain countries could mask the combined effect of gender diversity and certain contingent variables (moderating effect). The weak or absent direct relationship between gender diversity and performance is also indicative of a complex relationship whose contours have yet to be determined. Thus, placing our study in the Canadian context helps to reduce the adverse effects of heterogeneity in the study sample.

The Link between Gender Diversity and Firm Financial Performance in Canada
In Canada, the participation rate of women had reached 82% in 2014, compared to 91% for men, but their presence on the board of directors of companies remains modest. Indeed, according to the Catalyst report [41], the representation of women on boards of directors in companies on the Canadian stock exchange is 20.8%. In order to increase the feminization of BDs, the Canada Business Corporations Act (CBCA) requires all publicly traded companies to disclose diversity information.
As a result, the Canadian Securities Administrators adopted a "comply or explain" diversity disclosure regime in 2014. Under this regime, companies listed on the TSX (Toronto Stock Exchange) are required to disclose information on their gender diversity policies. Published data from the Canadian Securities Administrators on the representation of women on the board of directors shows a marked improvement over previous years. In 2019, a total of 17% of the board seats of TSX-listed companies were held by women (11% in 2015 and 15% in 2018).
In terms of education, 2011 was a special year for women in Canada. For the first time ever, more women than men graduated from post-secondary education. More specifically, 64.8% of women between the ages of 25 and 64 had completed post-secondary education, compared to 63.4% of men. In 1991, 14% of women had a college diploma compared to 26% in 2015; for men, this percentage was 9% in 1991, and it increased to 19% in 2015. The proportion of women aged 25 to 64 with a university certificate or degree increased more rapidly than that of men. Indeed, this proportion increased from 15% in 1991 (19% for men) to 35% in 2015 (29% for men). In addition, the proportion of Canadian women with no formal educational credentials (no certificate, diploma or degree) was greatly reduced, from 31% in 1991 to 9% in 2015. Finally, between 2001 and 2012, women received 60.7% of all university degrees awarded to Canadians. Thus, with such an increase in women's education, the knowledge and experience of women board members is likely to broaden the available human capital and provide new cognitive frameworks for integrating diverse perspectives into the reading of contemporary challenges facing Canadian businesses.
In terms of values, the report "Women in Canada: A Gender-based Statistical Report" revealed that more women than men did unpaid work (e.g., health care and support, volunteering) as a form of concurrent activity, with 60.8% of women and 40.2% of men doing so at the same time. More specifically, for volunteer activities (e.g., serving on a committee or board, giving advice, protecting the environment, teaching or mentoring), according to a 2013 Statistics Canada survey, women in the 15-74 age group volunteered as much or more than men in the last 12 months for an organization. For example, for the age group of 25 to 54 years old, this proportion was 47% for women as compared to 43% for men (a significantly different value at the 5% cut-off than that observed for women in the same age group). Consequently, because of the differences in male-female values, it is legitimate to think that women's cognitive frameworks also influence the decision-making processes within a board not only through the information used in decision-making, but also through the way decisions are made. Indeed, several studies support the idea that heterogeneous groups are superior to homogeneous groups because of the tendency of members to engage in substantive discussions and to integrate the variety of knowledge and information available [42,43].
Adams and Funk's [44] study of 502 board members and 126 CEOs regarding their value system according to the Schwartz model [45][46][47], revealed significant male-female differences that were similar to the general population for certain aspects and different from the population for others. More specifically, female board members were more involved in volunteer work (activities that preserve and improve the well-being of those with whom they are in frequent personal "group" contact), and were more universalist (understanding, appreciative, tolerant, and protective of the well-being of all and of nature), but less power-oriented (social status and prestige, control or domination over people and resources) than their male counterparts. However, in contrast to the general population, women on BDs had fewer traditional values (from culture and religion), less security orientation (preserving existing social arrangements), and comparable risk aversion to men. Hypothesis 1. The percentage of women on the board (H1a) and on the executive committee (H1b) has a positive impact on the financial performance of Canadian companies.

The Mediating Role of ESG
As previously mentioned, the disparity of results with regard to the relationship between gender diversity in BDs and performance (positive, negative or neutral) and the low correlation between the two variables reveals the complexity of this relationship, and the plausibility of the existence of indirect and underlying mechanisms linking the two variables. According to Hermalin and Weisbach [1], board characteristics (e.g., gender diversity, size, independence of members), affect the financial performance of the company through actions that members influence, such as extra-financial performance or ESG orientation (environmental, social, and corporate governance). Hermalin and Weisbach's [1] theory is in line with the logic of the UET (Upper Echelons Theory) in the sense that UET uses the characteristics of board members as a proxy for their cognitive framework, while Hermalin and Weisbach [1] refer to the actions of firms as mediators of the effect of gender diversity, and, therefore, of the cognitive frameworks of members, on financial performance. For example, previous studies showed that the gender diversity of BDs seems to be linked to companies' actions, which are in turn linked to employee policies [48], the increase in acquisitions [49], and to the ESG [50][51][52]. In line with studies by Boulouta [50], Harjoto, Laksmana, and Lee [51], and Jain and Jamali [52], and based on the stakeholder theory, the interest of shareholders is considered to be served when the interest of stakeholders is also served. In other words, the diversity of the type of board allows board members to mobilize different cognitive frameworks, and these, due to their variety, allow the interest of several stakeholders to be considered, including those of shareholders [13,51,[53][54][55]. According to several studies, ESG orientation is one of the corporate actions that influence financial performance [54,[56][57][58]. The effect of ESG on financial performance is made possible in two ways: (1) the quality of exchanges with stakeholders facilitates exchanges with actors other than those involved in business transactions and increases the company's share capital; (2) cost reduction by increasing employee and customer loyalty and avoiding the consequences of non-compliance with regulatory frameworks.

Hypothesis 2.
The effect of the percentage of women on the board (H2a) and on the executive committee (H2b) on the financial performance of Canadian companies is mediated by the ESG focus.

Sample
Based on data from the DataStream and Asset4 ESG databases, this study focused on medium and large companies operating in Canada between 2002 and 2019. The choice of this period was based on two criteria. The first was the scope of the 18-year timeline, which allows firms to be considered in both good times and bad. The second was the availability of all the data relating to the variables of interest (independent variables, dependent variable, mediator variables, and control variables). Thus, a total of 925 observations were retained, representing 133 Canadian companies.

Measurement of Variables
Dependent variable "financial performance": The majority of empirical studies have used Tobin's Q or ROA for measuring the financial performance of firms. For the present study, Tobin's Q was chosen because, unlike performance measurement with ROA, the disparity of results between gender diversity and performance is more present with the use of the Q index, and has the advantage of combining financial and stock market data. Additionally, Tobin's Q is a market-based measure that reflects unstructured and volatile bank performance better than ROA [2]. Tobin's Q is computed based on the most usual proxy: the book value of total assets plus the market value of equity minus the book value of equity divided by total assets [10,29,35]. A Tobin's Q value greater than 1 reflects investor confidence in the company and its growth opportunities, which is not the case for companies with a Tobin's Q of less than 1.
Independent variables "gender diversity on the board and the executive committee (EC)": Gender diversity on the board/EC is calculated as the ratio of the number of women on the board/EC to the total number of members on the board/EC. This measure is consistent with several studies on the characteristics of board members [2,55,59].
Mediating variable "ESG": The ESG Score is an overall company score based on its self-reported information in the environmental, social, and corporate governance pillars.
Control variables: In order to avoid bias for some confounding variables, and to increase the strength of the results, control variables were introduced. At the firm level, a control was conducted for their characteristics (i.e., performance, leverage and Altman Z-score) and the quality of their governance (i.e., CEO-chairman separation dummy variable and shares held by insiders). At the country level, a control was conducted principally with regard to the country inflation rate, and in all models dummies were introduced to control for industry effects.
For the robustness check, we considered "PERFORM" (the ratio of Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) to total assets) as another proxy for financial performance.

Data Analysis Plan
The data analysis was based on structural equation modeling using the bootstrap approach for mediation tests and the maximum-likelihood estimator. It was structured in five steps: (1) descriptive analysis of the data in order to draw a portrait of the sample by presenting the mean, the standard-deviation, and the correlations between the variables of interest; (2) verification of a direct relation model between the two independent variables "gender diversity on the board of directors and executive committee" and the dependent variable "financial performance"; (3) verification of the strength of the direct relation model by introducing the control variables; (4) verification of the mediating effect of the ESG orientation in the relationship between gender diversity on the board of directors and the executive committee, and financial performance; and (5) verification of the robustness of the mediation model by introducing the four control variables.

Result of the Analyses
In this section, empirical results are presented with regard to the relationship between gender diversity in the boardroom and financial performance, as well as the mediating role of the ESG orientation. Table 2 presents the bivariate correlations between the variables of interest (independent, dependent, mediating, and control variables), and Table 3 presents the descriptive statistics (mean and standard deviation). On average, the representation of women on the board of directors of the Canadian companies participating in the study was just under 16%, a percentage close to that found in the Pletzer, Nikolova, Kedzior, and Voelpel study [27]. This percentage is 14% in the EC with almost the same standard deviation as the percentage of women on the board, i.e., 12 points. Participating companies operate in six sectors of activity (68.4% industrial, 9.3% utilities, 2.8% transportation, 11.6% bank/savings and loans, 3.1% insurance, 4.8% other financial). It is clear that the industrial sector is over-represented as compared to other sectors, which is why this "sector" variable was introduced in the subsequent analyses as a control variable. The mean performance on Tobin's Q was 0.53 and the standard deviation was 0.28. Some firms had a Tobin's Q greater than 1 (max = 1.36), which indicates strong performance, while for others the Tobin's Q was 0 (min = 0). The Pearson correlation coefficient varied between −0.20 and +0.73. The correlation between the two independent variables (gender diversity in the BD and EC) was moderate (+0.37), indicating a lack of multicollinearity in the model. The strongest correlation was between gender diversity on the board and financial performance (Q index). Table 4 presents the estimated results of the study's basic hypothetical model of explaining financial performance through diversity on the BD and the EC while controlling for certain variables. Linear regression revealed a significant direct effect between gender diversity on the board and financial performance; the standardized regression beta is estimated to be 0.59 (β = 0.59; p = 0.000) (it is 0.74 without the control variables (β = 0.74; p = 0.000)). Thus, for a one-unit increase in the percentage of women on the board of directors, financial performance increases by 0.59 units. At the 5% significance level, the lower and upper confidence interval (CI) limits of the standardized regression beta coefficient are 0.53 and 0.64, respectively. This result is consistent with Hypothesis H1a, which is, therefore, accepted. However, regression analyses, including the control variables, did not reveal a significant relationship between EC gender diversity and financial performance (β = 0.05; p = 0.052). Therefore, Hypothesis H1b is rejected.
In accordance with the objective of the study, subsequent analyses were carried out to verify the mediating role of the variable "ESG orientation". The results presented in Table 5 show a significant mediating effect between gender diversity in the boardroom and financial performance. More specifically, the total effect of gender diversity in BDs on financial performance while introducing the control variables was estimated to be 0.59 (0.71 without the control variables), and can be decomposed into two effects, a significant direct effect of 0.577 (β = 0.57; p = 0.000, 5% CI = [0.51, 0.64]) and a significant indirect effect of 0.022 (β = 0.02; p = 0.000, 5% CI = [0.005, 0.039]). Thus, Hypothesis H2a is accepted. On the other hand, because Hypothesis H1b is rejected, mediation Hypothesis H2b is also rejected.
In summary, as shown in Figure 1, when women were more represented on the board, financial performance increased (1 for 0.59). Almost 4% (0.022/0.59) of this effect was achieved through ESG orientation. However, the representation of women on the EC did not affect performance. Furthermore, as shown in Figure 2, the robustness test also confirmed the mediating effect (indirect effect) of ESG in the relation between gender diversity on the board and financial performance measured by the "PERFORM" proxy. Thus, hypotheses H1a and H2a are supported while H1b and H2b are not. In addition, the model explains 52.20% of the variance in financial performance measured with Tobin's Q.

Discussion and Conclusions
The objective of this study was to test the effect of gender diversity on boards of directors (BDs) and on executive committees (ECs) (defined as the percentage of women) on corporate financial performance (as measured by Tobin's Q index) in the Canadian context from 2002 to 2019. In addition, we wanted to identify the mechanism by which this effect occurs while controlling for confounding variables. Based on the cognitive theory, the UET, and the stakeholder theory, the study began with the idea that characteristics of the board/EC, such as gender diversity, are a proxy for the cognitive frameworks of the board/EC members, and from the idea that the actions of board members, such as ESG orientation, are manifestations of these cognitive frameworks explaining the financial performance of firms.
The results of the analyses of 925 observations from 133 Canadian companies reveal that this line of reasoning is well tested for gender diversity on the board, but is not the case for gender diversity in the EC. Thus, in response to the debate arising from the disparity of results from previous studies, particularly those that measured financial performance using Tobin's Q index, this study supports the effect of gender diversity on financial performance in the Canadian context, which is consistent with several studies conducted in countries other than Canada [28][29][30][31][32][33]. Moreover, the influence of women is most evident when their rank is high (BD vs. EC), and the effect of gender diversity on the BD is not matched by the effect of gender diversity in the EC. Therefore, the results of our study do not support the idea that gender diversity should first be proven in the EC before being considered on the board, or the idea that gender diversity in the EC is a first step towards gender diversity on the board. On the contrary, gender diversity seems to have an effect when considered at the highest decision-making level.
As previously mentioned, Post and Byron's [16] study found that the influence of women on the board of directors appears to be greater in the paid workforce promoting gender parity, which is consistent with this study. In fact, Canada, like other countries (France, Spain, Norway), has a regulatory framework (Canada Business Corporations Act, CBCA, etc.) which is based on the principle of gender parity; and the Diversity Disclosure Regime of the Canadian Securities Administrators, which supports the representation of women on the board. Thus, in the Canadian context, gender diversity can be supported both from an ethical (the right thing to do) and utilitarian (corporate profitability) point of view.
This study demonstrates the viability of a mediation model based on the ESG variable. As such, the results reveal that the total effect (direct and indirect) of gender diversity on the financial performance of companies (up to 9% of the total effect), is partly due to the company's ESG orientation. Consistent with Galbreath's study [58], this study reveals the mechanism by which the diversity of cognitive frameworks of board members influences performance. For example, it appears that the representation of women on the board of directors influences the sensitivity of board members regarding issues affecting stakeholders other than shareholders, and thus act on different fronts, i.e., economic, social, and environmental (ESG).
As we have already pointed out, studies conducted in a single country are rare or use outcomes other than the financial performance of firms. However, overall, the results of our study are consistent with the few studies conducted in a single country with respect to the beneficial effect of gender diversity on the board of directors, but less so with respect to gender diversity on the executive committee. Indeed, contrary to our study (Hypothesis H1b), a study conducted in the USA by Li and Zhang [60] revealed that the role of gender diversity on the executive committee reduces potential problems and improves the effectiveness of the board of directors. It should be noted, however, that our model considers the two-tiered effect of diversity at the board and executive levels. Cabeza-Garcia, Fernandez-Gago, and Nieto's [61] study conducted in Spain is consistent with our study regarding the beneficial effect of gender diversity, including disclosure of ESG information. Also in Europe, the study by Jiraporn, Bouattour, Hamrouni, and Uyar [62] conducted in France showed how gender diversity helps increase the chances of paying dividends-a positive effect, but limited to one stakeholder, the shareholders. The only study, to our knowledge, using a model close to ours is the one conducted by Galbreath [58] in Australia. That study revealed a mediating effect of ESG orientation in the relationship between gender diversity and financial performance. However, the timeline of that study was limited to two years (2004)(2005), gender diversity did not consider members of the executive committee, and the method chosen for the mediation test is not unanimously accepted by the scientific community.

The Study's Contributions
This study makes several contributions to the debate on the effect of the feminization of boards of directors (BDs) and executive committees (ECs). First, the study focused on a single country, Canada. It should be of interest to firms operating only in the Canadian context, and is, therefore, subject to sample heterogeneity biases, such as the legislative framework and national, cultural, and social contexts. Our study also contributes to understanding the role of gender diversity in BDs and ECs over a broader timeline (18 years, from 2002 to 2019), thus avoiding situations contingent on a particular period. The results of this study provide empirical support for the beneficial effect of gender diversity for Canadian companies in financial, social, and environmental terms. In addition, beyond the direct effect of gender diversity at the board level regarding financial performance, the results reveal the mechanism of mediation by the ESG and the superiority of diversity at the board level over diversity at the EC level.

Limits and Future Perspectives
As in all previous studies, the argument for introducing gender diversity on the board is based on what this diversity brings in terms of a cognitive framework that allows board members to understand issues from different perspectives. In all studies on the subject, the measurement of the effect of cognitive frameworks involves measuring certain proxies (antecedents or consequences). It would be appropriate in future studies to introduce, in addition to measures at the organizational level, measures at the individual level that focus on the perception of women themselves, as board members, of their role on the board and their contribution to financial performance. In fact, crossing the two levels of analysis (individual and organization) allows a better understanding of the effect of the diversity of cognitive frameworks resulting from gender diversity without going through proxies such as the characteristics of the board of directors. To do this, qualitative research strategies (participant observation, ethnographic design) are more appropriate to capture the behavior of members as a result of gender diversity and the dynamics within the board.
While this study encourages gender diversity on the board for its effects on ESG orientation and financial performance, it does not provide information on the existence of a critical threshold below which the effect may not be manifest and a threshold above which the effect will be reversed. It would be interesting to explore this avenue on the basis of critical mass theory and group heterogeneity theory.
This study found that the relationship between gender diversity on the board and corporate financial performance is complex, with only 9% of the total effect passing through the ESG. It is important to think further in order to explore other mechanisms for mediation, but also for moderation by certain variables specific to women board members (e.g., education, external-internal status, independence, experience). In addition, the hypothetical model of this study expresses the idea of the effect of gender diversity on performance. However, it is also plausible to argue that a firm's performance could influence the composition of its board of directors or that the two influence each other (endogenous relationship). It is important that future studies test cross-lag analysis models to explore these issues.
Author Contributions: Z.O. and J.B.M. contributed to conception, analysis and interpretation of data as well as writing and revising of the manuscript; S.A. contributed to understanding practical gender diversity and ESG in Canadian context, specially banking Industry. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.