Does the Gender of Directors Matter?

In the paper, Does the Gender of Directors Matter?, which was recently made publicly available on SSRN, I investigate how having gender-balanced boards affects the working of boards. There has always been interest in the makeup of boards of directors, both in the academic literature and in the popular press. Lately, board diversity, and particularly the gender of directors has been a topic of much attention, since there is a recent movement to impose diversity requirements on boards. In the United States, the Securities Exchange Commission requires that companies disclose whether they have a diversity policy, and how it applies to board recruitment practices (Regulation S-K, Item 407(c)). In Europe, several countries (including Norway, France, Spain, and Italy) have already legislated laws enforcing gender quotas. Furthermore, Dr. Viviane Reding, the vice president of the European Commission, is promoting legislation enforcing gender quotas in all European countries (New York Times, March 4th, 2012).

Despite this attention, it nonetheless remains unclear whether diversity has a meaningful impact upon boards of directors, and how such impact might be measured. As emphasized by Hermalin and Weisbach (2003), board composition is jointly determined with firm performance, so it is problematic to draw inferences from the associations between firm performance and board composition. These scholars argue that instead of looking at the impact of boards on a firm’s overall financial performance, it is better to understand the impact of board composition by considering how it affects the actions the board or the firm take, given the board composition at the time the action is taken.

Most “action-based” studies rely on publicly available information, and consequently they usually study only a narrow set of actions that are observable, such as the decision to replace a CEO, to adopt a poison pill, or to acquire another company. However, boards take many actions in each meeting that are unobservable to outsiders, but are recorded in the board minutes. I am able to obtain detailed minutes, which are quasi-transcripts, of the board and board-committee meetings of eleven business companies in which the Israeli government holds a substantial equity interest (Government Business Companies, or GBCs) for one year each in the period 2007-2009—altogether 155 board meetings and 247 meetings of board committees. These minutes document the details of the meetings, including all the statements made by every participant in each meeting. Altogether, 2,459 issues were discussed in these meetings.

I use this database to evaluate the extent to which the gender of a board affects its actions. These data are ideal for understanding the effect of gender on board dynamics for at least three reasons. First, unlike studies based on publicly available information, I can observe virtually all the actions of directors, most of which are unobservable to outsiders. Second, because I know the attendance at every board meeting and which actions were taken at each one, I can control for firm-level characteristics by using within-firm variation across meetings. Third, these companies have boards that are relatively gender-balanced, containing roughly 36% women, and have included a large proportion of women for a decade and a half. This diversity is different from most boards of directors, since with a few exceptions such as the Scandinavian countries, most countries’ boards contain too few women to evaluate the effect of diversity.

As a point of departure, I take the hypothesis offered by Shrader et al. (1997), Rosener (1995), and Kramer et al. (2007): In board meetings, a critical mass of three women directors will catalyze board activeness/performance. I examine empirically whether indeed having boards that include a critical mass of three women directors, and also one of three men directors, catalyzes the effort a board exerts. Based on the minutes-data I document whether the board (1) requested to receive further information or an update and (2) whether it took an initiative, such as proposing which action should be taken. These two actions document the extent boards are active and involved in monitoring on the one hand, and supervising on the other, respectively.

I examine how the gender composition of the directors in attendance impacts upon the likelihood that a board will take each of these actions. The empirical results indicate that boards are most active when they are relatively gender-balanced—when at least three men and three women directors are in attendance, a situation I term a “dual-critical mass”. Boards with such a dual critical mass were found, in comparison to boards without one, to be at least twice as likely both to request further information or an update and also to take an initiative. These results are particularly driven by the presence of a critical mass of women directors.

To address the possible concern that non-random attendance is driving these results (i.e., that one of the genders is likely to attend meetings especially when high/low activeness is expected to be required), I use instrumental variables (IVs) that capture the financial incentives of men and women directors to attend meetings on a particular day. Because the compensation GBC directors receive depends only on the number of meetings they attend, they have a stronger incentive to attend meetings on days on which they have more than one meeting scheduled. I use the number of board- and committee-meetings scheduled for men and women directors on a particular day as instruments to assess the attendance of men and women directors. The finding that boards are most active when a critical mass of women directors is in attendance holds when including these instrumental variables in the analysis.

To understand whether these patterns documented are consistent with the impact/correlation of the gender of directors on financial performance, I examine the relation between the gender composition of the GBC boards and financial performance. A panel data of the universe of the 34 GBCs for the years 1997-2009 is examined. Consistent with the above-mentioned findings, ROE and the net profit margins are found to be significantly larger in companies that have at least three women directors. These findings imply that indeed, a critical mass of women directors not only catalyzes board activeness, it also seems to catalyze financial performance.

I examine how critical masses of men and women directors impact upon the likelihood that individual men and women directors take an action (i.e., that they request an update or take an initiative). I adjust for the likelihood that each gender of directors takes an action using the Horvitz–Thompson estimator. This estimator adjusts for the fact that usually women directors constituted a smaller fraction of all attending directors, and therefore, all else equal, the likelihood that women director take an action is smaller. The findings show that on the level of the individual director, a critical mass of women directors increases by 193% the likelihood that a women director take an action, and by 95% the likelihood that a man director take an action (the former results are more significant than the latter).

The study also finds that women directors are significantly more likely than men directors to take actions pertaining to supervisory issues. In board meetings, (in which in contrast to board-committees no additional appointment mechanism prevails, and therefore both genders of directors have equal opportunities to take actions) each gender of directors was found to have a relative penchant for certain types of tasks: Relative to men directors, women directors were found to take frequently actions pertaining to supervisory issues. The flip side of these findings is that men were significantly more likely to take actions pertaining to managerial issues.

In addition, I examine separately one type of period in which boards are of particular importance—periods boards are in the process of replacing their CEOs (“gap periods”). First, I conduct a traditional finance analysis, which examines a panel data set for 34 GBCs for the years 2001-2009. This analysis demonstrates that with respect to CEO turnover, “above the surface” the impact of the gender composition of GBC boards is similar to that of US firms as has been documented by Adams and Ferreira (2009). I find evidence suggesting that firms with weak financial performance, which also had boards that included a critical mass of three women directors, were more likely to replace their CEO. Next, the impact of gender “beneath the surface” is examined using the minutes-data. This study shows that on the individual level, women directors were particularly likely to be active during the gap periods. On the board level, during these gap periods, once again, the board was found to be particularly active if a critical mass of women directors was in attendance.

Last, I find that the larger the percentage of women in attendance at meetings of board-committees, the longer discussions their discussions—a finding which may imply that women foster communication, particularly in small teams. Taken together, the findings of this study stress that gender-balanced boards exert more effort than non-gender-balanced boards, and have a more diverse set of skills.

The full paper is available for download here.