The board of directors is the highest decision-making authority in a corporation. But sometimes boards struggle to make decisions. In surveys, 67 percent of directors report the inability to decide about some issues in the boardroom. Moreover, 37 percent say they have encountered a boardroom dispute threatening the very survival of the corporation.  Such a “division among the directors”—such deadlock on the board—“may render the board unable to take effective management action” and can even lead directors to “vote wholly in disregard of the interests of the corporation.” 
In 2017, deadlock on the board made it hard for Uber to appoint a CEO. According to the New York Times, “Uber’s C.E.O. selection…illustrates the high-wire act of herding eight board members…toward consensus.” Moreover, deadlock on the board led one front-runner for the job, Meg Whitman, to withdraw her name from consideration, saying “it was becoming clear that the board was still too fractured to make progress on the issues that were important to me.” 
Deadlock on the board can be so costly to U.S. corporations that most states have adopted deadlock statutes, which often give courts the power to dissolve a deadlocked corporation, a power they rarely have, except in the event of default or fraud. A substantial legal literature studies how corporations can resolve deadlock ex post. In our paper, available here, we ask how deadlock can be avoided ex ante. Can the right mix of directors ensure a board makes efficient decisions? And, if so, how should director elections be structured to help achieve the right board composition? Should director tenure be limited? And should shareholders have all the power to choose directors or should the CEO have some power as well?
To address these questions, we develop a dynamic model of board decision making in which directors’ disagreements can lead to deadlock on the board. Shareholders suffer, since a deadlocked board struggles to remove undesirable policies or executives—a deadlocked board leads to an entrenched CEO. In particular, we show that a CEO can be so severely entrenched that he is not fired even if all directors prefer a replacement. To see why, consider a firm with a bad incumbent CEO, whom the board is considering replacing with an alternative. Suppose all directors agree that the alternative is better than the incumbent, but some directors are especially biased toward him. For example, activist representatives could be biased toward an executive with a history of divestment. Then, if the alternative becomes the new CEO, the biased directors will try to keep him in place, voting down alternatives in the future, no matter how much other directors prefer them—the new CEO will become entrenched. To prevent this, other directors block the alternative today, keeping the bad incumbent CEO in place to retain the option to get their way in the future—the incumbent CEO becomes entrenched. The fear of entrenchment begets entrenchment.
We find that long director tenures, a hotly debated issue, exacerbate deadlock. In anticipation of a long tenure, directors behave strategically, blocking good candidates to preserve a strong bargaining position in the future, and thereby create deadlock.
Similarly, we find that boardroom diversity can exacerbate deadlock (its benefits notwithstanding). For example, the deadlock caused by an activist’s bias toward divestiture is not resolved by adding some executive directors biased toward investment. These directors will block divestiture-oriented policies, even if they agree that they are optimal today, just to preserve a strong bargaining position for the future. More generally, heterogeneous director biases do not cancel out—they do not yield a board that implements policies in shareholders’ interests. Rather, they can yield a board that does not implement any policies at all. This downside of diversity is in line with the empirical literature, which finds that diversity of directors’ skills and experience is negatively associated with strategic change and firm value. 
Finally, we show that it is not always optimal for shareholders to have full power to choose new directors. By ceding some power over director appointments to the CEO, shareholders can commit not to block his preferred policies in the future, and hence prevent deadlock today.
In summary, we analyze a formal model of deadlock on the board, and we show that board composition and director appointment procedures matter for deadlock, sometimes in unexpected ways. Our findings suggest implementable policy procedures, relevant for investors, directors, and managers alike: Beware of too much diversity, encourage shorter director tenures, and do not give shareholders full control over director elections.
 IFC, 2014, Conflicts in the boardroom survey, Corporate Governance Knowledge Publication, p. 2.
 Kim, Susanna, 2003, The provisional director remedy for corporate deadlock: A proposed model statute, Washington and Lee Law Review 60, pp. 113, 120.
 See “Inside Uber’s Wild Ride in a Search of a New C.E.O.” New York Times, August 29, 2017. Whitman’s description of Uber’s board mirrors the dictionary definition of deadlock: “a situation, typically one involving opposing parties, in which no progress can be made” (New Oxford American Dictionary).
 Knyazeva, Anzhela, Diana Knyazeva, and Charu Raheja, 2013, The benefits of focus vs. heterogeneity: Dissimilar directors and coordination within corporate boards, Working paper, University of Rochester; Goodstein, Jerry, Kanak Gautam, and Warren Boeker, 1994, The effects of board size and diversity on strategic change, Strategic Management Journal 15, 241–250.