**Table 1.**

costs. On the other hand, the idiosyncratic risk is inherent in the company's operations and its management expertise. The overall risk of an investment is commonly measured by the standard deviation of its return [40–42]. In accordance with previous research [43–45], we use the standard deviation of monthly stock returns over a one-year period as global risk proxy (RISK) formulated as follows.

$$
\sigma R\_{\text{it}} = \sqrt{\frac{\sum\_{i=1}^{n} \left(R\_{\text{it}} - \overline{R}\right)^2}{n-1}} \tag{1}
$$

σ *Rit* = standard deviation of company's monthly stock return, *Rit* = company's monthly stock return, *R* = company's average monthly stock return, *n* = Number of observation.

Thus, a higher standard deviation value reflects a higher risk.

The gender variable (GEND) was measured by the percentage of women on the board of directors (number of women on the board of directors/total number of directors). This measure, more and more used by similar researchers [5, 46, 47], makes it possible to go beyond the limits of binary measurements of women's presence and their representativeness in the boards and to take into account board size.

In addition, we have deemed it useful to introduce control variables that are likely to have an effect on the relationship between women's presence on boards of directors and firm risk.

First, the size of the board of directors (B-SIZE) was included as a control variable as a large board has more chance to contain female members. In addition, previous literature suggests that the size of boards of directors is a major determinant of its effectiveness and governance mechanism to protect company assets, to secure better allocation of resources, to limit managerial opportunism, and to prevent the risks of insolvency risk or dangers of risky investments. B-SIZE is measured by the number of directors on the board. Larger boards tend to be also associated with lower return volatility [48].

Second, R&D spending and innovation effort are considered as an indicator of risk-taking insofar as their impact often remains uncertain [7]. R&D intensity (RID), measured by the "R&D expenditures divided by total annual sales" ratio, was introduced into the model as a control variable. Third, we control the model by the


**Table 2.** *Variables description.* leverage (LEV). Leverage is considered to be an indicator of solvency often associated directly with firm financial risk. According to Bodie et al. [49], leverage level allows welling assess the risk of financial distress and bankruptcy and provides information on the ability of the company to honor its financial commitments. Among others, Abobakr and Elgiziry [50] found that percentage of females to the total board to be significantly negatively correlated to ratio leverage. We define leverage as the "total debt to total assets" ratio. Forth, firm size (F-SIZ), measured by the natural logarithm of total assets (in millions of Euros), was introduced into our model as a control variable. In this regard, Bruna et al. [9], for example, have shown that risk aversion was negatively associated with firm size. Finally, since profitability level could imply greater risk-taking [51], we opted for controlling for profitability (FFP) calculated as the Return on Assets. For a detailed description of variables see **Table 2**.

### **3.3 Model**

To test our hypotheses, we use panel data regression running the following model.

$$\text{RISK}\_{\text{at-}} \, f \left( \text{GEND}\_{i}^{t} \, \text{B} - \text{SIZ}\_{i}^{t}, \text{FFP}\_{i}^{t}, \text{F} - \text{SIZ}\_{i}^{t}, \text{LEV}\_{i}^{t}, \text{RID}\_{i}^{t} \right) \tag{2}$$

Model 1

$$\begin{aligned} RISK\_{\rm \tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\tiny\atop\langle\,^2D \choose\langle\,^2D \, ^2F \, ^2F \, ^3F \, ^3F \, ^3F \, ^3F \, ^3F \, ^3F \, ^3F \end{aligned} \tag{3}$$

In equation 3, the index t presents the year of the observation, while the index i refers to the company. The variable RISK is our dependent variable. GEND is the proportion of women on the board of directors. B-SIZ, FFP, F-SIZ, LEV, and RID are our controls as defined above. The βj with j {0…6} are the model parameters. μ is the time fixed effect and *e* is our idiosyncratic error term.

#### **4. Results and discussion**

#### **4.1 Descriptive statistics**

**Table 3** reports the descriptive statistics for all variables. The data show that the average 10-year stock return volatility is 34.6%, with a minimum and a maximum of 1.3 and 90%, respectively. Nevertheless, RISK presents a significant dispersion with a high standard deviation, which is understandable given the diversity of the firms being studied. Regarding gender diversity, the proportion of women is 23.8% of the total number of directors. It should be noted that this average has increased considerably over time after the Copé-Zimmermann enacted the law in 2011. The average size of boards of directors is around 14, and the average return on assets of the companies in our sample is around 4%.

#### **4.2 Multivariate analyzes**

The correlations between all test variables are reported in **Table 4**. At this point, and without claiming to draw a definitive conclusion, we notice that GEND has a high negative correlation with the risk metric. This is consistent with our hypothesis

#### *Board Gender Diversity and Firm Risk DOI: http://dx.doi.org/10.5772/intechopen.100189*


#### **Table 3.**

*Descriptive statistics.*


*This table presents the correlations between all variables and the variance inflation factor (VIF). It shows that correlation between the explanatory variables as each of the variables used range from 1.05 to 1.90 along with mean VIF value below 10.\*Significance at the 5% levels.*

*\*\*Significance at the 1% levels.*

*\*\*\*Significance at the 0.01% levels.*

#### **Table 4.**

*Correlation matrix***.**

and is in line with the result of the current literature. GEND also has a strong correlation (p-value <0.0001) with both board size (0.1654) and financial performance (0.2211). This is reasonable as large boards are more likely to include the female gender. There is a significant correlation between risk and board size in our sample, which is consistent with previous researches [52]. Besides, we note a positive correlation at the 1% level of confidence between risk and ROA and a significantly positive association between risk and leverage ratio. Finally, there is a significant relationship between risk and R&D expenditures.

To test our hypothesis, we ran least squares regressions and then we had selected the appropriate estimation based on the different usage tests. Our estimation satisfies the assumptions of normality. Furthermore, we had ensured the independence between the error terms by the Durbin-Watson autocorrelation test and the absence of heteroscedasticity problems by the Breusch-Pagan test. Finally, the F-test allowed us to opt for the fixed-effect estimation, which allows overcoming unobserved heterogeneity issue over time.

According to **Table 5** results, the model's explanatory power is acceptable (R<sup>2</sup> within equal to 24.12%). The statistic of the Wald test (Wald Chi), presents a statistically significant p-value (Prob > Chi) at 1% level. The coefficient of the dependant variable (gender) is negative and significant (Coeff. = −0.903; t-stat = − 2.82)


*This table presents results from fixed effects regression of the risk measure on board gender diversity, over the period 2011–2018. A firm's total risk (RISK) is the annualized standard deviation from the monthly stock returns over the past year. GEND is the percentage of women on the board of directors.*

*\* Significance at the 10% level.*

*\*\*Significance at the 5% level. \*\*\*Significance at the 1% level.*

**Table 5.**

*Multifactor regression results RISK as a dependant variable.*

indicating that women board presence mitigates firm risk. This result supports our assumption; that is, to say that an increase in women representation on board decreases firm risk. It is also in line with the main arguments presented in our theoretical background. Women are less "adventurous" than men and are likely to have an innate aversion to risk. Then, this result joins those obtained in previous empirical research [5, 53]. However, these findings do not support those found by [9] who have found no evidence to support the assumption of a significant relationship between women on corporate boards and firm risk-taking on a sample of SBF 120 index listed companies. This can be explained by the fact that the authors carried out their investigation over the period 2006–2010 before the adoption of the Copé-Zimmermann law in 2011 or by the measurement of the risk retained.

Other than the GEND effect, we find a negative association between board size and firm global risk (under the 5% risk error). This is in line with the risk aversion hypothesis, which suggests that large boards tend to control leverage, as this will increase the volatility of equity returns [54]. Likewise, this can be explained by the fact that large boards can lead to slower decision-making processes and search for compromises inducing less risky behaviors. This result is in accordance with Nakano and Nguyen [52] who observed lower performance volatility and lower risk bankruptcy in companies with larger boards of directors. Another explanation could be found in the arguments put forward by Cheng [53] and Wang [54] suggesting that a high number of administrators would prevent boards from functioning properly, thus limiting performance variability. Bureaucracy and vulnerability to agency issues that characterize large boards may cause less performance volatility and higher risk.

The positive coefficient of the variable FFP confirms that risk has a positive effect on return on assets. However, this influence is not statistically significant. The proportion of women on the board of directors, therefore, has no impact on economic performance. This observation is inconsistent with the risk-based rationale

#### *Board Gender Diversity and Firm Risk DOI: http://dx.doi.org/10.5772/intechopen.100189*

presented by Akbas et al. [55] according to which a greater risk is not associated with better expected profitability, nor with the results of Nartea et al. [56] who determined that risk has a positive effect on return.

In addition, we found a negative link between firm size and risk. This may be because large companies, which enjoy better governance and less information asymmetry, tend to control the volatility of stock returns [57]. This negative relationship was also found by Damanpour [58] who argues that large companies better control fluctuations in stock prices.

Also, **Table 5** indicates a coefficient on leverage variable of 0.091, significant at 5% level of confidence (t-statistics of 2.46). These results confirm the close link between stock return volatility the leverage suggesting that the volatility of returns represents a risk hindering the company' ability to go for debt financing [59].

Finally, R&D intensity has a coefficient of 0.0009 (t-statistic 5.49) significant at 1% level. This relationship most likely exists. One of the plausible explanations has been given by Mazzucato and Tancioni [60] suggesting that the volatility of stock returns is indeed linked to innovation since financial markets react to signals provided by companies about their future growth prospects through their R&D spending and innovation behavior.

## **5. Conclusion**

The aim of this research was to examine the impact of gender diversity on firm risk. More specifically, we studied the relationship between board women presence and firm global risk measured by stock return volatility. To this end, we used a sample of French companies listed SBF 120 over the period 2011–2018. Using a panel data regression method, we were able to highlight a negative and statistically significant link between the percentage of women on boards of directors and firm risk. Our results suggest that better women representativeness on boards could lead to better financial performance through risk mitigation. This is in line with gender socialization theory suggesting that women would be more risk-averse as well as with the agency theory, which states that women would exercise better control and participate actively in conflict resolution. The potential explanations for this negative effect would lie not only in the crucial role played by women in boards in ensuring better risk oversight [7, 13], in the reduction of agency costs [61], but also in the risk aversion generally observed among women [62]. Our results help to enrich the debate on this issue. They are consistent with a current of the literature [26, 31, 63] but remain in contradiction with the results of other empirical research [9]. Thus, this should lead us to consider the role of contingent variables that may moderate/mediate this relationship, such as cultural differences, organizational visibility, or intersectorial differences. On the other hand, we should also recognize that reverse causation is likely to exist and deserves to be examined, which means that companies with lower risk might intentionally choose more female directors [36]. Thus, it is important to address the issue of endogeneity when studying the relationship between gender diversity on the board of directors and business risk.

Moreover, although our results suggest that women inclusion on boards is likely to strengthen corporate governance mechanisms, mitigate risks, and maximize value, it is important to note that such appointments should not be a reaction to normative pressures just to legitimize governance modes or to comply with rules deemed socially acceptable. Our results should therefore not be analyzed from an instrumentalist perspective, that is to say, from purely economic and financial angles, but rather based on notions of gender equity and justice as well as on fundamental legal principles. Hence, it is also strongly recommended that nomination

and governance committees in the boards of directors take steps to achieve menwomen parity within boards. The laws of the various European countries appear to be evolving in this direction.

Our study has also several limitations. First, by focusing our investigation only on large companies, the results cannot be generalized. The extension of the sample to cover small- and medium-sized enterprises would make it possible to give more robustness to observed results. The second limit is related to the risk measurement: the variability of stock return may not capture all risks incurred by the company. Finally, an impact of the female presence may not be immediate: Considering a delay effect (by introducing lagged variables in the econometric model) would probably be interesting since the appointment of women may take time to observe its risk impact.

### **Author details**

Zyed Achour National Institute of Labor and Social Studies, University of Carthage, Tunisia

\*Address all correspondence to: zyed.achour@intes.rnu.tn

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*Board Gender Diversity and Firm Risk DOI: http://dx.doi.org/10.5772/intechopen.100189*
