In her principal speech during her recent successful campaign to become the United Kingdom's new Prime Minister, Theresa May said a lot of things that as a corporate governance thinker I have very serious doubts about. Let's start with her proposal to put consumers and employees on corporate boards of directors.
And I want to see changes in the way that big business is governed. The people who run big businesses are supposed to be accountable to outsiders, to non-executive directors, who are supposed to ask the difficult questions, think about the long-term and defend the interests of shareholders. In practice, they are drawn from the same, narrow social and professional circles as the executive team and – as we have seen time and time again – the scrutiny they provide is just not good enough. So if I’m Prime Minister, we’re going to change that system – and we’re going to have not just consumers represented on company boards, but employees as well.
Presumably, she has something along the lines of the German codetermination model in mind. And that's a really bad idea.
Is it curious that only shareholders get the vote? What about all of the corporation’s other constituencies, such as employees, creditors, customers, suppliers, etc.? Why do they not get a voice in, say, the election of directors? The traditional answer is that shareholders own the corporation. Ownership typically connotes control, of course. Consequently, despite the separation of ownership and control characteristic of public corporations, shareholders’ ownership of the corporation might be deemed to vest them with unique control rights. Recall that the nexus of contracts theory of the firm demonstrates that shareholders do not in fact “own” the corporation in any meaningful sense. By throwing the concept of ownership out the window, however, the contractarian model eliminates the obvious answer to our starting question—why are only shareholders given voting rights?
Our answer to that question relies on the analysis of organizational decisionmaking economist Kenneth Arrow set out in The Limits of Organization.[1] Recall that Arrow identified two basic modes of decisionmaking: consensus and authority. Consensus requires that each member of the organization have identical information and interests so that preferences can be aggregated at low cost. In contrast, authority-based decisionmaking structures arise where group members have different interests and information.
The analysis that follows proceeds in three steps. First, why do corporations not rely on consensus-based decisionmaking? In answering that question, we begin by imagining an employee-owned firm with many thousands of employee-shareholders. (Employees are used solely for purposes of illustration—the analysis would extend to any other corporate constituency.) After demonstrating that Arrow’s conditions cannot be satisfied in such a firm, we then turn to the more complex public firm in which employees and shareholders constitute separate constituencies to demonstrate that Arrow’s conditions are even less likely to be met in this type of firm. Second, why do corporations not permit multiple constituencies to elect directors? Finally, why are shareholders the favored constituency?
A. The necessity of authority
1. Information
Assume an employee-owned corporation with 5,000 employee-shareholders. Could such a firm function as a participatory democracy? Not if we hoped that each participant would make informed decisions. As a practical matter, of course, our employee-shareholders are not going to have access to the same sorts of information. Assuming at least some employees serve in managerial and supervisory roles, they will tend to have broader perspectives, with more general business information, while line workers will tend to have more specific information about particular aspects of the shop floor.
These information asymmetries will prove intractable. A rational decisionmaker expends effort to make informed decisions only if the expected benefits of doing so outweigh its costs. In a firm of the sort at bar, gathering information will be very costly. Efficient participatory democracy requires all decisionmakers to have equal information, which requires that each decisionmaker have a communication channel to every other decisionmaker. As the number of decisionmakers increases, the number of communication channels within the firm increases as the square of the number of decisionmakers.[2] Bounded rationality makes it doubtful that anyone in a firm of any substantial size could process the vast number of resulting information flows. Even if they were willing to try, moreover, members of such a firm could not credibly bind themselves to reveal information accurately and honestly or to follow prescribed decisionmaking rules. Under such conditions, Arrow’s model predicts that the firm will tend towards authority-based decisionmaking. Accordingly, the corporation’s employer-owners will prefer to irrevocably delegate decisionmaking authority to some central agency, such as a board of directors.
Now introduce the complication of separating capital and labor. Nothing about such a change economizes on the decisionmaking costs outlined above. Instead, as described below, labor and capital can have quite different interests, which increases decisionmaking costs by introducing the risk of opportunism. In particular, capital and labor may behave strategically by withholding information from one another.
2. Interests
Again, begin by assuming an employee-owned firm with 5,000 employee shareholders. Is it reasonable to expect the similarity of interest required for consensus to function in such a firm? Surely not. In some cases, employees would differ about the best way in which to achieve a common goal. In others, individual employees will be disparately affected by a proposed course of action. Although the problems created by divergent interests within the employee block are not insurmountable, such differences at least raise the cost of using consensus-based decisionmaking structures in employee-owned firms.
The existence of such divergent interests within the employee group is confirmed by the empirical evidence. Labor-managed firms tend to remain small, carefully screen members, limit the franchise to relatively homogeneous groups, and use agenda controls to prevent cycling and other public choice problems.[3] All of these characteristics are consistent with an attempt to minimize the likelihood and effect of divergent interests.
Now again complicate the analysis by separating capital and labor. Although employee and shareholder interests are often congruent, they can conflict. Consider, for example, the down-sizing phenomenon. Corporate restructurings typically result in substantial reductions in force, reduced job security, longer work weeks, more stress, and diminished morale.[4] From the shareholders’ perspective, however, the market typically rewards restructurings with substantial stock price increases. The divergence of interest suggested by this example looms large as a bar to the use of consensus in capitalist firms.
B. The inefficiency of multiple constituencies
The analysis to this point merely demonstrates that corporate decision making must be made on a representative, rather than on a participatory, basis. As yet, nothing in the analysis dictates the U.S. model in which only shareholders elect directors. One could plausibly imagine a board of directors on which multiple constituencies are represented. Indeed, imagination is not required, because the supervisory board component of German codetermination provides a real world example of just such a board.[5] Empirical evidence, however, suggests that codetermination does not lead to efficiency or productivity gains.[6]
Why not? In Arrow’s terminology, the board of directors serves as a consensus-based decisionmaking body at the top of an authority-based structure. Recall that for consensus to function, however, two conditions must be met: equivalent interests and information. Neither condition can be met when employee representatives are on the board.
The two factors are closely related, of course. Indeed, it is the potential divergence of shareholder and employee interests that ensures employee representatives will be deprived of the information necessary for them to function. Because of the board’s position at the apex of the corporate hierarchy, employee representatives are inevitably exposed to a far greater amount of information about the firm than is normally provided to employees. As the European experience with codetermination teaches, this can result in corporate information leaking to the work force as a whole or even to outsiders. In the Netherlands, for example, the obligation of works council representatives to respect the confidentiality of firm information “has not always been kept, causing serious concerns among management which is required . . . to provide extensive ‘sensitive’ information to the councils.”[7]
Given that providing board level information to employee representatives appears clearly contrary to shareholder interests,[8] we would expect managers loyal to shareholder interests to withhold information from the board of directors in order to deny it to employee representatives, which would seriously undermine the board’s ability to carry out its essential corporate governance roles. This prediction is borne out by the German experience with codetermination. German managers sometimes deprive the supervisory board of information, because they do not want the supervisory board’s employee members to learn it.[9] Alternatively, the board’s real work may be done in committees or de facto rump caucuses from which employee representatives are excluded. As a result, while codetermination raises the costs of decisionmaking, it may not have much effect on substantive decisionmaking.[10]
Although Arrow’s equality of information criterion is important, in this context the critical element is the divergence of shareholder and employee interests. The interests of shareholders will inevitably differ as amongst themselves, as do those of employees, but individual constituents of the corporation nevertheless are more likely to share interests with members of the same constituency than with members of another constituency. Allowing board representation for employees thus tends only to compound the problem that gives rise to an authority-based hierarchical decisionmaking structure by bringing the differing interests of employees and shareholders directly into the board room.[11] The resulting conflicts of interest inevitably impede consensus-based decisionmaking within the board. Worker representatives on corporate boards tend to prefer greater labor advocacy than do traditional directors, no doubt in large part because workers evaluate their representatives on the basis of labor advocacy, which also results in role conflicts.[12] The problem with codetermination thus is not only that the conflict of employee and shareholder interests impedes the achievement of consensus, but also that it may result in a substantial increase in agency costs.[13]
Although it is sometimes asserted that employee representation would benefit the board by promoting “discussion and consideration of alternative perspectives and arguments,”[14] the preceding analysis suggests that any such benefits would come at high cost. In addition, there is reason to doubt whether those benefits are very significant. Workers will be indifferent to most corporate decisions that do not bear directly on working conditions and benefits.[15] All of which tends to suggest that employee representatives add little except increased labor advocacy to the board.
C. Why only shareholders?
The analysis thus far demonstrates that public corporation decisionmaking must be conducted on a representative rather than participatory basis. It further demonstrates that only one constituency should be allowed to elect the board of directors. The remaining question is why shareholders are the chosen constituency, rather than employees. Answering that question is the task of this section.
The standard law and economics explanation for vesting voting rights in shareholders is that shareholders are the only corporate constituent with a residual, unfixed, ex post claim on corporate assets and earnings.[16] In contrast, the employees’ claim is prior and largely fixed ex ante through agreed‑upon compensation schedules. This distinction has two implications of present import. First, as noted above, employee interests are too parochial to justify board representation. In contrast, shareholders have the strongest economic incentive to care about the size of the residual claim, which means that they have the greatest incentive to elect directors committed to maximizing firm profitability.[17] Second, the nature of the employees’ claim on the firm creates incentives to shirk. Vesting control rights in the employees would increase their incentive to shirk. In turn, the prospect of employee shirking lowers the value of the shareholders’ residual claim.
At this point, it is useful to once again invoke the hypothetical bargain methodology. If the corporation’s various constituencies could bargain over voting rights, to which constituency would they assign those rights? In light of their status as residual claimants and the adverse effects of employee representation, shareholders doubtless would bargain for control rights, so as to ensure a corporate decisionmaking system emphasizing monitoring mechanisms designed to prevent shirking by employees, and employees would be willing to concede such rights to shareholders.
Granted, collective action problems preclude the shareholders from exercising meaningful day-to-day or even year-to-year control over managerial decisions. Unlike the employees’ claim, however, the shareholders’ claim on the corporation is freely transferable. As such, if management fails to maximize the shareholders’ residual claim, an outsider can profit by purchasing a majority of the shares and voting out the incumbent board of directors. Accordingly, vesting the right to vote solely in the hands of the firm’s shareholders is what makes possible the market for corporate control and thus helps to minimize shirking. As the residual claimants, shareholders thus would bargain for sole voting control, in order to ensure that the value of their claim is maximized. In turn, because all corporate constituents have an ex ante interest in minimizing shirking by managers and other agents, the firm’s employees have an incentive to agree to such rules.[18] The employees’ lack of control rights thus can be seen as a way in which they bond their promise not to shirk. Their lack of control rights not only precludes them from double-dipping, but also facilitates disciplining employees who shirk. Accordingly, it is not surprising that the default rules of the standard form contract provided by all corporate statutes vest voting rights solely in the hands of common shareholders.
To be sure, the vote allows shareholders to allocate some risk to prior claimants. If a firm is in financial straits, directors and managers faithful to shareholder interests could protect the value of the shareholders’ residual claim by, for example, financial and/or workforce restructurings that eliminate prior claimants. All of which raises the question of why employees do not get the vote to protect themselves against this risk. The answer is two-fold. First, as we have seen, multiple constituencies are inefficient. Second, as addressed below, employees have significant protections that do not rely on voting.
Suppose a firm behaves opportunistically towards it employees. What protections do the employees have? Some are protected by job mobility. The value of continued dealings with an employer to an employee whose work involves solely general human capital does not depend on the value of the firm because neither the employee nor the firm have an incentive to preserve such an employment relationships. If the employee’s general human capital suffices for him to do his job at Firm A, it presumably would suffice for him to do a similar job at Firm B. Such an employee resembles an independent contractor who can shift from firm to firm at low cost to either employee or employer.[19] Mobility thus may be a sufficient defense against opportunistic conduct with respect to such employees, because they can quit and be replaced without productive loss to either employee or employer. Put another way, because there are no appropriable quasi-rents in this category of employment relationships, rent seeking by management is not a concern.
Corporate employees who make firm-specific investments in human capital arguably need greater protection against employer opportunism, but such protections need not include board representation. Indeed, various specialized governance structures have arisen to protect such workers. Among these are severance pay, grievance procedures, promotion ladders, collective bargaining, and the like.[20]
In contrast, shareholders are poorly positioned to develop the kinds of specialized governance structures that protect employee interests. Unlike employees, whose relationship to the firm is subject to periodic renegotiation, shareholders have an indefinite relationship that is rarely renegotiated, if ever. The dispersed nature of stockownership also makes bilateral negotiation of specialized safeguards difficult. The board of directors thus is an essential governance mechanism for protecting shareholder interests.
If the foregoing analysis is correct, why do we nevertheless sometimes observe employee representation? An explanation consistent with our analysis lies close at hand. In the United States, employee representation on the board is typically found in firms that have undergone concessionary bargaining with unions. Concessionary bargaining, on average, results in increased share values of eight to ten percent.[21] The stock market apparently views union concessions as substantially improving the value of the residual claim, presumably by making firm failure less likely. While the firm’s employees also benefit from a reduction in the firm’s riskiness, they are likely to demand a quid pro quo for their contribution to shareholder wealth. One consideration given by shareholders (through management) may be greater access to information, sometimes through board representation. Put another way, board of director representation is a way of maximizing access to information and bonding its accuracy. The employee representatives will be able to verify that the original information about the firm’s precarious financial situation was accurate. Employee representatives on the board also are well-positioned to determine whether the firm’s prospects have improved sufficiently to justify an attempt to reverse prior concessions through a new round of bargaining.
[1] Kenneth J. Arrow, The Limits of Organization 63-79 (1974).
[2] Oliver E. Williamson, Markets and Hierarchies: Analysis and Antitrust Implications 46 (1975). In addition to the transaction costs associated with making an informed decision, which by themselves are doubtless preclusive of participatory democracy in the large firm setting, the opportunity cost entailed in making informed decisions is also high and, even more important, readily apparent. In contrast, the expected benefits of becoming informed are quite low, as an individual decisionmaker’s vote will not have a significant effect on the vote’s outcome. Our employee-owners thus will be rationally apathetic.
[3] Greg Dow & Louis Putterman, Why Capital (Usually) Hires Labor: An Assessment of Proposed Explanations, Brown Univ. Dep’t of Econ. Working Paper 96-21 at 63-64 (Sep. 1996).
[4] Michael Useem, Investor Capitalism 164-65 (1996).
[5] In Germany alone, there are at least four different statutory models of participatory management. Some other member-states of the European Union also have some form of employee representation, and there have long been proposals to develop harmonized company laws or even a European Union-wide company law that would provide for employee representation. See generally Terence L. Blackburn, The Societas Europea: The Evolving European Corporation Statute, 61 Fordham L. Rev. 695, 743-55 (1993). Unless otherwise indicated, discussion of codetermination herein focuses on the 1976 German codetermination statute, which applies to corporations having more than 2,000 employees.
Codetermination includes a dual board structure: a supervisory board that appoints a managing board, with the latter actively operating the firm. Workers are represented only on the former. The supervisory board concept is difficult to translate into terms familiar to those trained exclusively in U.S. forms of corporate governance. Its statutory mandate is primarily concerned with the appointment and supervision of the managing board. In theory, employees and shareholders are equally represented on the supervisory board. In practice, however, the board is often controlled either by the firm’s managers or a dominant shareholder. One of the employee representatives must be from management, and shareholders are entitled to elect the chairman of the board, who has the power to break tie votes. If push comes to shove, which it reportedly rarely does, shareholders thus retain a slight but potentially critical edge.
Works councils are concerned with issues affecting individual plants, rather than the whole firm. In theory at least, works councils have considerable control over shop floor and other issues that affect the personnel of a particular plant. Depending on the nature of the issue at hand, the council may be entitled to provision of information, consultation with management, or codetermination. Their statutory mandate ranges from such minor issues as individual personnel grievances to such major concerns as the introduction of new technology or plant closings.
[6] See generally Stephen M. Bainbridge, Participatory Management Within a Theory of the Firm, 21 J. Corp. L. 657 (1996) (summarizing studies).
[7] Tom R. Ottervanger & Ralph M. Pais, Employee Participation in Corporate Decision Making: The Dutch Model, 15 Int’l Law. 393, 399 (1981).
[8] One sure result of lost confidentiality will be worker demands for higher wages. This prediction is supported by an empirical study finding that provision of financial and other business information to employees of nonunionized firms had a negative effect on firm profitability, which was attributed to higher wages demanded by the informed employees. Morris M. Kleiner & Marvin L. Bouillon, Providing Business Information to Production Workers: Correlates of Compensation and Productivity, 41 Ind. & Lab. Rel. Rev. 605, 614-15 (1988). See also Stuart Ogden, The Limits to Employee Involvement: Profit Sharing and Disclosure of Information, 29 J. Mgmt Stud. 229 (1992) (stating that U.K. employers are reluctant to provide disclosure of financial information for fear of stimulating workers to make demands respecting pay and working conditions). In a unionized firm, moreover, management will be especially reluctant to inform union members on the board of information that might aid the union in collective bargaining.
[9] Klaus J. Hopt, Labor Representation on Corporate Boards: Impacts and Problems for Corporate Governance and Economic Integration in Europe, 14 Int’l Rev. L. & Econ. 203, 206 (1994).
[10] Tove H. Hammer et al., Worker Representation on Boards of Directors: A Study of Competing Roles, 44 Indus. & Lab. Rel. Rev. 661, 663 (1991) (Scandinavian experience with codetermination shows it has little substantive effect on corporate decisionmaking).
[11] The difficulty, of course, is not merely that the interests of employees and shareholders diverge, but also that different classes of employees have divergent interests. As we have seen, this seriously compounds the problem of aggregating constituency preferences.
[12] John L. Cotton, Employee Involvement 128 (1993). This conflict is exacerbated in heavily unionized industries, as representatives of a single union might sit on the boards of multiple firms within the industry. In the extreme case, the demise of one firm might redound to the greater good of the greatest number by benefiting union members who work at competing corporations. This creates the potential for perverse incentives on the part of union representatives on the board.
[13] The most obvious problem is the possibility that employee representation will permit management to pursue its own self-interest at the expense of both shareholders and employees by playing worker and shareholder representatives off against each other. Legal and market accountability mechanisms constrain this tendency, but because they are not perfect there remains the possibility that self-interested managers may throw their support behind the side of the board whose interests happen to coincide with those of management in the issue at hand. See William J. Carney, Does Defining Constituencies Matter?, 59 U. Cin. L. Rev. 385, 420-24 (1990).
This conflict is well-known, of course, but there is a more subtle problem that is often overlooked. Corporate employees have an incentive to shirk so long as their compensation does not perfectly align their incentives with those of the firm’s shareholders. In turn, knowing of this phenomenon, the firm’s shareholders should expect management to reduce the compensation of the firm’s employees by the amount necessary to off-set the expected degree of employee shirking. Because ex ante wage adjustments rarely are fully compensatory, due to bounded rationality and the resulting use of incomplete contracts, the firm’s shareholders should expect management to monitor the employees and punish ex post those who shirk. Benjamin Klein, Contracting Costs and Residual Claims: The Separation of Ownership and Control, 26 J. L. & Econ. 367, 368 n.2 (1983). Would it not seem odd that those who are to be monitored should be allowed to choose the monitors? One of the accountability mechanisms that aligns managerial and shareholder interests is monitoring by the board of directors. Allowing employee representation on the board necessarily reduces the likelihood that the board will be an effective monitoring device. Because shareholders “could seek profits by getting highly motivated managers who sweat the labor force,” workers have an interest in supporting rules that free management from accountability to shareholders. Mark J. Roe, Strong Managers, Weak Owners: The Political Roots of American Corporate Finance 44 (1994). Managerial shirking of its monitoring responsibilities thus will often redound to the workers’ benefit, which suggests that employee representatives on the board of directors are less likely to insist on disciplining lax managers than are shareholder representatives. If employees are entitled to voting representation on the board of directors, monitoring by the board and its subordinate managers will be less effective, which will cause agency costs to rise.
The validity of this prediction is confirmed by the German experience with codetermination. Conflicts of interest faced by employee representatives on the supervisory board remain a serious, but unresolved concern. Employee representation slows the finding of a consensus on the supervisory board and creates a built-in polarization problem. Not surprisingly, it is standard practice for employee and shareholder representatives to have separate pre-meeting caucuses. Klaus J. Hopt, Labor Representation on Corporate Boards: Impacts and Problems for Corporate Governance and Economic Integration in Europe, 14 Int’l Rev. L. & Econ. 203, 206-08 (1994).
[14] Robert Howse & Michael J. Trebilcock, Protecting the Employment Bargain, 43 U. Tor. L. Rev. 751, 769 (1993).
[15] See Michael P. Dooley, European Proposals for Worker Information and Codetermination: An American Comment, in Harmonization of the Laws in the European Communities: Products Liability, Conflict of Laws, and Corporation Law 126, 129 (Peter E. Herzog ed. 1983) (opining: : “As to the majority of managerial policies concerning, for example, dividend and investment policies, product development, and the like, the typical employee has a much interest and as much to offer as the typical purchaser of light bulbs.”).
[16] See, e.g., Frank H. Easterbrook and Daniel R. Fischel, The Economic Structure of Corporate Law 66-72 (1991). An alternative answer based on the model used herein returns to the divergence of interests within constituency groups. Although investors have somewhat different preferences on issues such as dividends and the like, they are generally united by a desire to maximize share value. Board consensus therefore will be more easily achieved if directors are beholden solely to shareholder interests, rather than to the more diverse set of interests represented by employees.
A related but perhaps more telling point is the problem of apportioning the vote. Financial capital is fungible, transferable, and quantifiable. Control rights based on financial capital are thus subject to low cost allocation and valuation. In contrast, the human capital of workers meets none of these criteria. While one-person/one-vote would be a low cost solution to the allocation problem, it appears highly inefficient given the unequal distribution of reasoning power and education. If the most competent people and/or those with the most at stake should have the most votes, some more costly allocation device will be necessary.
[17] The superiority of shareholder incentives is a relative matter. Shareholders may have better incentives than other constituencies, but the phenomenon of rational apathy nevertheless limits the extent to which shareholders can be expected to act on those incentives.
[18] According to the Coase Theorem, rights will be acquired by those who value them most highly, which creates an incentive to discover and implement transaction cost minimizing governance forms. See Ronald H. Coase, The Problem of Social Cost, 3 J. L. & Econ. 1 (1960). Although shareholders and employees obviously do not bargain, a basic premise of the law and economics account is that corporate law provides them with a set of default rules reflecting the bargain they would strike if they were able to do so.
[19] See generally Oliver Williamson, Corporate Governance, 93 Yale L.J. 1197 (1984). This is not to say that exit is cost-less for either employees or firms. All employees are partially locked into their firm. Indeed, it must be so, or monitoring could not prevent shirking because disciplinary efforts would have no teeth. The question is one of relative costs.
[20] As private sector unions have declined, the federal government has intervened to provide through general welfare legislation many of the same protections for which unions might have bargained. The Family & Medical Leave Act grants unpaid leave for medical and other family problems. OSHA mandates safe working conditions. Plant closing laws require notice of layoffs. Civil rights laws protect against discrimination of various sorts. Even such matters as offensive horseplay have come within the purview of federal sexual harassment law.
[21] Brian E. Becker, Concession Bargaining: The Impact on Shareholders’ Equity, 40 Indus. & Lab. Rel. Rev. 268 (1987).
