Fortune editor-at-large Justin Fox asks: "Why do companies have boards of directors, anyway?" Why aren't corporations run using "direct democracy—putting all major corporate decisions, including the choice of a new CEO, to a shareholder vote"? Or by "absolute corporate monarchy—allowing management to make all the decisions without oversight"? Good questions, all. On close examination, however, the modern system of corporate governance in which the corporate hierarchy is topped by a small committee of independent board members, who typically lack both day-to-day management power and any significant equity stake in the corporation, turns out to be a highly efficient solution to the problems of making decisions in complex organizations.
Why Not Direct Shareholder Democracy?
Nobel laureate economist Kenneth Arrow's seminal work on organizational decision making identified two basic decision-making structures: "consensus" and "authority." Consensus is utilized where each member of the organization has identical information and interests and will therefore select the course of action preferred by all the other team members. In contrast, authority-based decision-making structures arise where team members have different interests and amounts of information. They are characterized by the existence of a central agency to which all relevant information is transmitted and which is empowered to make decisions binding on the whole.
It is very hard to imagine a modern public corporation that could be effectively run using consensus-based decision-making mechanisms. At the most basic level, the mechanical difficulties of achieving consensus amongst thousands of shareholders impede direct democracy.
Even if those collective action problems could be overcome, however, active shareholder participation in corporate decision making still would be precluded by the shareholders' widely divergent interests and distinctly different levels of information. As to the former, while neoclassical economics assumes that shareholders come to the corporation with wealth maximization as their goal, and most presumably do so, once uncertainty is introduced it would be surprising if shareholder opinions did not differ on which course will maximize share value. As to the latter, shareholders lack incentives to gather the information necessary to actively participate in decision making and thus are rationally apathetic.
Consequently, it is hardly surprising that, among public companies, corporate governance precisely fits Arrow's model of an authority-based decision-making system. Shareholders lack the information and the incentives necessary to make sound decisions on either operational or policy questions. Under these conditions, Arrow predicts, it is "cheaper and more efficient to transmit all the pieces of information to a central place" and to have the central office "make the collective choice and transmit it rather than retransmit all the information on which the decision is based," which is precisely what a board-centered system of corporate governance does.
Why Not Monarchy?
The more interesting question Fox poses can be rephrased as: Why a board rather than an individual autocrat? In theory, corporate governance has always been board-centered. As the Delaware General Corporation Law, for example, has long put it: the corporation's "business and affairs . . . shall be managed by or under the direction of the board of directors." The statutory model of corporate governance thus contemplates not a single hierarch, but rather a multi-member body that typically will act by consensus.
In practice, however, corporations were de facto run by an imperial CEO or a team of senior managers, with little or no interference from other stakeholders. Shareholders were essentially powerless and typically quiescent. Boards of directors were little more than rubber stamps.
Today, corporate governance looks very different. In recent years, several trends coalesced to encourage more active and effective board oversight. Much director compensation is now paid in stock, for example, which helps align director and shareholder interests. Courts have made clear that effective board processes and oversight are essential if board decisions are to receive the deference traditionally accorded to them under the business judgment rule, especially insofar as structural decisions are concerned (such as those relating to management buy-outs). Third, director conduct is constrained by an active market for corporate control, ever-rising rates of shareholder litigation, and, some say, activist shareholders. As a result, modern boards of directors typically are smaller than their antecedents, meet more often, are more independent from management, own more stock, and have better access to information. Real world practice thus increasingly has come into line with the statutory model of board primacy.
The economic advantages associated with board primacy vis-à-vis CEO primacy come into play when we consider the three major functions of the board of directors: (1) providing an in-house set of specialists and access to a network of outside resources; (2) broad policy making; and (3) monitoring and disciplining top management. A sizeable body of research suggests that groups are superior to individuals in all three areas.
Networking is the easiest context in which to make the case for groups, of course. Unlike a single CEO, who often came up through the ranks, specializing in a particular area of the firm's business, a diverse board comprised mainly of outsiders links the firm to a network of other companies who may be useful suppliers or customers. Likewise, such a board often includes persons with specialized expertise, such as lawyers or bankers, who can bring valuable outsider perspectives to the table.
As for policy making and monitoring, both of these involve the exercise of critical evaluative judgment. A wealth of research in psychology, sociology, and behavioral economics confirms that groups often make better decisions than individuals. Even more strikingly, the conditions under which groups outperform individuals in laboratory settings have important similarities to board decision making.
Two examples might suffice. In the 1930s, Marjorie Shaw conducted a classic experiment in which four-person teams of undergraduates solved various problems with single, self-confirming solutions (so-called "Eureka" problems.) One set of problems involved three variants on the classic missionaries and cannibals game. The other set of problems required subjects to solve two word puzzles and another involving spatial relationships. The proportion of correct solutions in a sample of groups was significantly higher than a sample of individuals working alone.
Some 50 years later, Frederick Miner devised another classic experiment to test the ability of groups to exercise evaluative judgment vis-à-vis that of individuals. Miner's experiment required 69 self-selected groups, each composed of 4 undergraduate business students, to solve the so-called winter survival exercise. This exercise, which is variously attributed, has become something of a benchmark standard in the field. The subject group is told that they are survivors of an airplane crash at a remote location. They first must decide whether to walk out or remain at the crash site. They then must rank the utility of 15 survival aids. Miner found that group rankings were more accurate than those of the average individual subject.
Although Miner's experiment may not seem relevant to corporate governance, it has certain instructive features. First, it used business students, who presumably resemble corporate directors more closely than other plausible experimental subjects. Second, the subjects knew one another before becoming members of the group and were allowed to form their own groups, both of which somewhat replicate the process by which boards form. Finally, and most importantly, the subjects shared a single goal (i.e., survival). Granted, the experiment thus did not require them to aggregate preferences as to which there might be value differences, but rather to pool their collective knowledge and use that knowledge to evaluate alternatives in light of the shared goal. If we assume that directors generally share a primary goal of shareholder wealth maximization, however, this experimental condition also replicates corporate governance.
Even the limitations of Miner's study bear some resemblance to corporate boards, at least insofar as outsider dominated boards are concerned. Students subjects bring to the task only general human capital, and the experimental design cannot capture any equivalent to firm-specific human capital. Yet, outside directors accumulate significant firm-specific knowledge and human capital only after very long tenure. The subjects also had relatively little stake in the outcome, but many outside directors likewise lack a direct economic stake in the firm.
A wealth of other studies confirm these results, suggesting that decision making by groups will generally be preferable to decision making by an individual autocrat when it comes to the sort of critical evaluative judgments boards are asked to make. This is not to say that group decision making is always preferable. The old joke about the camel being a horse designed by a committee captures the valid empirical observation that individuals are often superior to groups when it comes to matters requiring creativity. But while individuals may well be better at devising a brilliant plan, individuals often become wedded to their plans and fail to see flaws that others might identify. As such, you may want an individual to write a symphony, but a group to evaluate it. Group decision making checks individual overconfidence by providing critical assessment and alternative viewpoints.
As we have seen, most of what boards do requires the exercise of critical evaluative judgment, but not creativity. Even the board's policymaking role entails judgment more than creativity, because the board is usually selecting between a range of options presented by subordinates. The board thus serves to constrain subordinates who have become wedded to their plans and ideas, rather than developing such plans in the first instance.
Why Not a Management Team?
Fox points out that "the chief decisionmaking group at most large corporations is top management, not the board." True, day-to-day decision making may be vested in the hands of a top management team, but ultimate power to hire and fire remains with the board.
In theory, the CEO could be monitored by his or her subordinates. Indeed, economist Eugene Fama contends that lower level managers in fact monitor more senior managers. Such up-stream monitoring, however, does not take full advantage of specialization. One response to the agency costs inherent in the corporate separation of ownership and control is separation of "decision management"—initiating and implementing decisions—from "decision control"—ratifying and monitoring decisions. Such separation is a defining characteristic of the central office typical of M-form corporations. The M-form corporation replaces the simple pyramidal hierarchy with a more complex structure in which the central office has certain tasks and the operating units have others, which allows for more effective monitoring through specialization, sharper definition of purpose, and savings in informational costs. In particular, the central office's key decision makers—the board of directors—specialize in decision control. Because executives specialize in decision management, expecting them to monitor the CEO thus calls on the former to perform a task for which they are poorly suited.
A different critique of Fama's hypothesis is suggested by evidence with respect to meeting behavior from research on group decision making. In mixed status groups, higher status persons talk more than lower status members. Managers, for example, talk more than subordinates in business meetings. Such disparities result in higher status group members being more inclined to propound initiatives and having greater influence over the group's ultimate decision.
One function of the board of directors thus is providing a set of status equals for top managers. As such, corporate law's insistence on the superiority of the board to management begins to make sense. To the extent law shapes social norms, corporate law empowers the board to constrain top management more effectively by creating a de jure status relationship favoring the board.
New Institutional Economics has taught us that the firm must be viewed as an institution—more precisely, as a set of institutions—rather than as a mere production function. Specifically, the corporation consists of a set of production teams embedded within a hierarchical structure. At the apex of that hierarchy stands not an individual, but yet another team—the board of directors.
Team production is imperfect, whether the product is a manufactured good or a corporate decision. Teams are subject to unique cognitive biases, such as groupthink, and unique sources of agency costs, such as social loafing. With respect to the exercise of critical evaluative judgment, however, groups have clear advantages over autonomous individuals. Not only do groups clearly outperform average individuals in a given sample, there is considerable evidence that the process of group interaction has synergistic effects allowing groups to outperform even the best decision makers in the sample.
The author teaches law at UCLA and is a TCS Contributing Editor.
 In this game, subjects were given three disks representing missionaries and three disks representing cannibals. The missionaries and cannibals are on one side of a river. The decision maker must get all six to the other side of the river using a boat that can only carry two discs at a time. All missionaries and one cannibal can row. Cannibals must never outnumber missionaries in any location for obvious reasons, albeit politically incorrect ones. See Marjorie E. Shaw, Comparison of Individuals and Small Groups in the Rational Solution of Complex Problems, 44 Am. J. Psych. 491, 492-93 (1932) (describing problems).
 Frederick C. Miner, Jr., Group v. Individual Decision Making: An Investigation of Performance Measures, Decision Strategies, and Process Losses/Gains, 33 Org. Beh. and Human Performance 112 (1984).
 For more on the merits of group decision making, see my article Why a Board? Group Decision Making in Corporate Governance, 55 Vanderbilt Law Review 1 (2002).