The following post comes to us from Nadya Malenko of the Finance Department at Boston College.
The board of directors is a collective body, whose members have diverse expertise in various aspects of the company’s business. Therefore, communication between directors is critical to successful board functioning. In recent years, regulators, shareholders, and directors themselves have been paying increased attention to decision-making policies that could increase the quality of board discussions. Executive sessions that exclude the management, separation of the CEO and chairman positions, board retreats, and separate committees on specific topics have been put in place to promote more effective communication. As governance experts Carter and Lorsch (2004) emphasize, “If we could offer only one piece of advice, it would be to strive for open communication among board members.”
In practice, communication in boards is impeded by several factors. In many board meetings, time allocated for discussion of important issues is very limited, and managers use their power to prevent any contentious discussions. Because of that, directors have been increasingly engaging in discussions outside board meetings (Hart, 2003), which is difficult given directors’ busy schedules. According to a former board chairman, “If you want to be influential on a board it takes time, so you have to make the commitment of time. Cause you’re gonna have a lot of discussions outside of board meetings” (Stevenson and Radin, 2009). Moreover, discussions outside board meetings can result in retaliation from the manager, who may want to prevent directors from communicating behind his back. Another obstacle to communication is that directors often refrain from openly expressing their views due to reputational concerns, such as the fear to appear incompetent by voicing a controversial opinion, or to be perceived as a troublemaker. Finally, given directors’ diverse preferences, backgrounds, and areas of expertise, conveying one’s knowledge effectively and persuasively requires preparation and effort. All these factors make it personally costly for directors to communicate their position openly and effectively.
In the paper, Communication and Decision-Making in Corporate Boards, forthcoming in the Review of Financial Studies, I examine whether existing board policies mitigate or exacerbate the problem of ineffective communication in order to better understand which policies are optimal. To study this question, I develop a theory of board decision-making whose key element is that the quality of discussion is endogenous and determined by the efforts directors put into communicating their position to others. The model yields novel implications for the choice between open and secret ballot voting, the role of board diversity, executive sessions, transparency, and board committees. These implications follow from two main results, which show that when communication is personally costly, certain biases in directors’ preferences can improve the quality of board discussions. In particular, both directors’ concerns about conformity and diversity in directors’ private interests can give directors stronger incentives to incur the costs and communicate their position to others.
Anecdotal and survey evidence suggests that conflicts of interest and pressure for conformity play an important role in directors’ decisions. Conflicts of interest arise due to private benefits directors receive from certain board decisions. For example, such conflicts are likely to arise in corporate control transactions due to directors’ ownership, affiliation, or relation with the CEO. Desire for conformity, which makes directors reluctant to deviate from other board members, is perceived by many as equally important. This reluctance to take a minority stand is due to several reasons, including the influence of the CEO and directors’ reputational concerns. For example, anecdotal evidence suggests that directors who oppose the CEO without support from other board members face retaliation and pressure to resign.
The paper develops a model that incorporates the key features of board decision-making described above—costly communication, conflicts of interest, and pressure for conformity. In the model, the board contemplates a decision whose value is uncertain—for example, an acquisition. Each director has private information relevant to the decision. The board’s decision process takes place in two stages—communication, followed by decision-making. At the communication stage, each director decides whether to incur a cost to communicate his information to others. At the decision-making stage, all directors take actions (e.g., vote) based on their private information and the information inferred from the discussion, and the board’s collective decision is made. Directors can have conflicts of interest and thus prefer a decision that is not optimal for shareholders. Directors can also have a preference for conformity and thus incur a loss if their action deviates from other directors’ actions—for example, if they vote differently from the majority.
As a benchmark, I show that if decisions are made without prior communication—that is, if directors vote just on the basis of their private signals, then both conflicts of interest and conformity are detrimental for board decisions. However, the conclusions are different when prior to making the decision, directors can choose to communicate their information to others at a cost.
The first result shows that pressure for conformity at the decision-making (voting) stage gives directors stronger incentives to incur the costs and communicate their information to others prior to the vote. By encouraging better communication, conformity at the decision-making stage can have an overall positive effect on board decisions. As an illustration of this result, consider the difference between open and secret ballot voting. Anecdotal evidence suggests that in the majority of cases, directors vote by open ballot. At first glance, this seems puzzling: pressure for conformity is stronger when voting is by open rather than secret ballot because the vote of each director is observable to other directors and the CEO. Thus, the open ballot system is likely to prevent directors from using their information and honestly voting their opinions. However, my first result implies that open ballot voting can nevertheless be optimal because it improves communication prior to the vote.
To see the intuition, suppose that the board contemplates a merger supported by the manager. If voting is by secret ballot, a director with negative information about the prospects of the merger will vote against it. Even if he is the only one to vote negatively, he will not be identified by the manager and hence will not be penalized. However, given potential costs of sharing his concerns with others, the director may be reluctant to speak up during the discussion and will simply vote against. Conversely, suppose that voting is by open ballot. Unless the director shares his negative information prior to the vote, he faces the risk that he will be the only one to vote against and hence will be penalized by the manager. Realizing this, the director has strong incentives to share his concerns with others, for example, outside the board meeting. By communicating his information, he ensures that other directors also become pessimistic about the merger and will support him in his opposition to the CEO. Thus, board communication is more effective under open ballot voting.
In general, the choice between open and secret ballot (stronger and weaker pressure for conformity) depends on several factors, including the decision under consideration, the nature of directors’ information, and their private interests. In the paper, I discuss the implications of the model for voting procedures, transparency, and executive sessions, and present anecdotal evidence consistent with these implications.
The second result of the paper shows that a stronger conflict of interest relative to other board members gives a director stronger incentives to communicate his information to them. Thus, board discussions can be more effective when directors’ private interests are more diverse. For example, this implies that in the context of a takeover, a board that includes both insiders, who are biased against the takeover, and bidder representatives, who are biased in its favor, can be more effective than a fully independent board.
The intuition is the following. Consider a firm that receives a takeover bid and an inside director who is biased against being taken over because of private benefits of control. Other board members expect the director to actively participate in the discussion and present evidence that the bid undervalues the target whenever he has such evidence. Hence, if the director does not speak up, other directors infer that he privately knows that the target’s value is low. This negative inference increases the probability that the board will accept the bidder’s offer. The more biased is the director relative to other board members, the more harmful for him is their negative inference when he does not speak up. At the margin, this gives a more biased director stronger incentives to incur the costs of communication and try to credibly convey his information.
Finally, the paper has implications for the allocation of decision-making rights among directors. I show that even if directors are identical, it can be optimal to allocate full decision-making power to one director, as opposed to giving voting rights to all directors. In the presence of communication costs, such allocation of control leads to the most efficient use of directors’ diverse information, with the director in control serving as a communication link. Directors without decision-making power communicate their information to the director in control, who makes the decision by incorporating all available information. I also show that when directors are heterogeneous, it is optimal to allocate control to directors who care the least about conforming to others because such directors distort their decisions the least. This is consistent with the observed practices, where more experienced directors (who are likely to have lower costs of disagreeing with others) are more likely to be appointed to the chairman or lead director positions.
The full paper is available for download here.