Soft gender diversity regulations for boards of directors: the moderating role of firm ownership and control structure

ABSTRACT

Women are significantly underrepresented worldwide on corporate boards of directors, leading to board gender diversity regulations. Drawing from agency and socioemotional wealth theories, we investigate the moderating effect of ownership structure, shareholder identity and shareholder control on how gender diversity codes and soft quotas influence women’s representation on boards. We analyse our sample of Spanish Stock Exchange firms using panel data Tobit models. The results show that a single large shareholder enhances the positive impact of regulations, while multiple blockholders diminish it. Blockholder identity also matters: family control of board seats and families as the sole large shareholders reduce the positive impact of board gender diversity regulations on women’s board representation. In family firms, the balance of power between family and non-family shareholders also affects the impact of regulations. Our findings highlight the need to consider firm ownership and control structure when developing regulations to promote board gender diversity.

KEYWORDS:

• Corporate governance
• board of directors
• female directors
• women on boards
• gender diversity regulation
• ownership and control structure

JEL CLASSIFICATION:

• G32
• G34
• G38
• M14

Acknowledgments

The authors acknowledge comments received at the Workshop of the Special Issue on Recent Developments in European Corporate Governance. Particularly, we would like to thank the suggestions and comments of the guest editors, Dr. Goergen and Dr. Pascual-Fuster, and the two reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Recommendation 15: ‘When women directors are few or non-existent, the board should state the reasons for this situation and the measures taken to correct it … .’

2. Recommendation 14: ‘The board of directors should approve a director selection policy that: […] c) […] should pursue the goal of having at least 30% of total board places occupied by women directors before the year 2020 … .’

3. Article 75: ‘[…] Companies obliged to present unabridged financial statements of income will endeavour to include a sufficient number of women on their boards of directors to reach a balanced presence of women and men within eight years of the entry into effect of this Act … .’

Additional provision one: ‘Balanced presence or membership […] will be understood to mean the presence of women and men in the context in question in a manner such that neither sex accounts for more than sixty nor less than forty per cent of the total.’

4. We apply the ultimate owner methodology and define the largest ultimate shareholder as the first largest shareholder who holds more than 10% of the company’s voting rights (La Porta et al., 1999). The 10% boundary is one of the most common approaches used in the literature to define large shareholdings (Rodríguez-Ariza et al., 2017). Moreover, as the sample firms are all listed firms, 10% ownership is sufficiently large to attain control. To identify the ultimate shareholders of each sample firm and the percentage of common shares they hold, we follow the chains of control.

5. Firms with a family or individual ultimate shareholder who holds more than 10% of the voting rights are classified as family firms (i.e. if family ultimate shareholder = 1).

6. F10 (dummy variable that takes a value of one if the firm has a first largest shareholder holding more than 10% of voting rights and zero otherwise), Controlling shareholder alone, First largest controls, Other multiple controls, and Family ultimate shareholder.

7. Most of the independent variables are highly significant and the pseudo-$R^2$ for the Probit regressions are high: F10 (0.256), Controlling shareholder alone (0.249), First largest controls (0.176), Other multiple controls (0.186), and Family ultimate shareholder (0.209). The pseudo-$R^2$ for the Probit regressions estimated on the matched sample are close to zero and all of the independent variables are non-significant but for few exceptions. The matching for the models that show a significant impact of firm ownership and control structure on gender diversity on boards (Models 1, 2 and 4 in Table 6 and Models 1, 2, 3 and 4 in Table 7) is reliable. However, it is worth nothing that leverage is significant (at the 5% level; z = 1.98) in the Probit regression estimated on the matched sample for variable Controlling shareholder alone.

8. F10 (250 firm-year observations matched), Controlling shareholder alone (902 firm-year observations matched), First largest controls (887 firm-year observations matched), Other multiple controls (875 firm-year observations matched), and Family ultimate shareholder (831 firm-year observations matched).

9. The Blau index is defined as: $1-\sum _{i=1}^n{p_i}^2$ where $p_i$ is the percentage of directors in each category (women and men), and $n$ is the number of categories. Range: 0 (no gender diversity) – 0.5 (same number of women and men). The Shannon index is defined as: $-\sum _{i=1}^n{p}_{i}ln\left(p_i\right)$ where $p_i$ and $n$ are equivalent to the Blau index definition. Range: 0 (no gender diversity) – 0.69 (same number of women and men).