Reform of corporate governance is hindered by incremental approaches to improvement derived from management. A more powerful paradigm emerges from considering imperatives inherent in the board’s organizational position rather than from analyses more suitable to management than governance. This author argues that proper governance is not an expression of management “one step up,” but of ownership “one step down,” its role one of enlightened command rather than helpful advice.

It would seem from recent laws, codes, seminars and literature that the problems of corporate governance are transparency, independence, and composition. Yes, they are, but in the way that the problems of a horse-drawn buggy are the poor braking system, the loose steering, and the manure deposits. More to the point, the problem of the buggy is that it isn’t an automobile. If we can get past the buggy to an auto, we can then address a more progressive form of those braking, steering, and exhaust issues. Today’s objects of governance reform, particularly independence and composition, are only symptoms of deeper problems in our conception of corporate governance. Corporate governance cannot be optimized by addressing these issues by new laws and codes, at least those of the current variety, no matter how aggressively enforced.

America’s Sarbanes-Oxley Act, UK’s Higgs Review, and Ontario’s Bill 198 do not promote good governance and, therefore shareholder value, as pointed out recently by Boardroom’s Brian Lechem, so much as they protect investors from unseemly conduct. Like speed limits and stop signs they are useful to protect, but in no way do they constitute guidance for skillful, wise driving. One look at the unending stream of codes, particularly since Cadbury’s 1990 report, is convincing evidence that there will continue to be more. While new prescriptions of practice are not without value, they have sharpened boards’ interest in lawful compliance more than in better governance. (Nell Minow of The Corporate Library opined that, despite the current focus on independence and compliance, the real issue is effective representation of shareholder interest.) Indeed, the effect of more laws and codes might well be the “infection by regulatoritis” predicted by CD Howe Institute’s Jack Mintz.

So just as improvements in the components of a horse and buggy system can never yield a car, piecemeal improvements-even if improvements they truly are-will not take us to governance excellence. I believe the familiar paradigm of corporate governance that pervades virtually all the current dialogue is flawed at the outset. Unless that paradigm is changed, governance improvements will continue to be trapped in fixing up the buggy, dooming good intentions to forever designing better parts for an ill-designed whole.

What is sorely needed is to follow the positional requirements of governance to their logical conclusions. By positional, I mean a design of the board’s job derived from its position in the scheme of things, rather than an approach consisting of adjustments to traditional practices. Bill Dimma noted that while going back to basics sounds like motherhood, it can rectify what mistakenly appeared to be obvious. I ask the reader’s indulgence, then, while I go back to some fundamentals. To explain what I mean by positional analysis, it is first useful to explain what it is not.

Governance Considered from Traditional Taxonomies

Management and the conduct of enterprise have been studied and taught for centuries, but with particular rigor in the last hundred years. As in the development of any field, ways have been found to break complexities down into manageable, researchable, and teachable parts. Here are but three conceptually legitimate, practically useful ways.

Subdisciplines. For the successful conduct of enterprise, a number of skills must be brought together, each sufficiently separable to constitute a legitimate field of study. They include marketing, finance, budgeting, human resources, plant design, and transportation. Business schools have long divided coursework into such discrete disciplines. Whole texts and, indeed, whole careers have been devoted to each of them. It is easy to visualize this way of viewing these managerial components as a simple list. Governance subdisciplines are less formally entrenched (including perhaps mergers, public offerings, executive compensation, etc.).

Environmental context. A company exists in an external context with which it must deal successfully if it is to survive. For example, a company must deal with labor, capital, product/service, and raw materials markets; it must deal with physical, legal, economic, and political environments. It is easy to visualize this contextual environment as a forcefield in which these factors surround the company.

Sequence of functions. Management has been analyzed as a never ending sequence of planning, staffing, directing, controlling, and evaluating. Each step can be analyzed as a separate skill set. It is easy to visualize this repeating sequence as a wheel with never-ending cycling.

However, as useful as these conceptual framings are, not one of them addresses the role of any specific position in the organization. We can glean no help from them in deciding, say, what the CEO’s role is with respect to planning versus the VP of Marketing’s role. We know no more about the role of the Plant Manager versus the Accounts Receivable Manager with respect to dealing with personnel issues. That is, not one of these useful categories is position-determined.

That fact normally presents no problem and, in fact, is hardly noticed. Position-determined requirements are established in each company by the pattern of delegation. For any position, that pattern is decided by the level just above or higher. Ed and Edna Employee confront no ambiguity about their roles as analyzed in whatever manner, for his or her supervisor will make clear what part of any one of these or other aspects of the total belongs to them.

That solution works smoothly in a well-managed company. The ways in which labor is divided and the nature of the internal delegation system are major components of any manager’s job. Each manager knows what his or her total accountability includes (the undivided accountability, prior to further delegation), then chooses what to parcel out to subordinates. Each level of management makes sure that subordinate positions know which parts of the manager’s undivided accountability is fulfilled by their jobs. Thus subordinates’ jobs are extensions of and derived from the accountability of the job just above them. That solution works well until we reach the top of the organization.

The board’s undivided accountability is for everything, but clearly most of that overwhelming total is delegated to the CEO. Most, but not all, else the board becomes a hollow flow-through of authority from shareholders to CEO, hardly a description of a decision-making role. That is, although granting a massive amount of decision authority to the CEO, the board must retain certain specific decisions for itself. The board, then, using the previously cited categories, has direct, undelegated responsibility for some part of, say, (a) planning, organizing, controlling, and other tasks (b) capital, labor, legal, and other market environments, and (c) personnel, compensation, marketing and other functions.

But unlike every other person from the CEO on down, there is no superior to tell the board which parts of the total will belong to it alone. It is not legitimate for the board to get its instructions from the CEO on just what part of these decisions it retains (a principle that would be obvious inside the company, but reverse delegation is not uncommon in the boardroom). Doing that would constitute an inversion of the proper sequence of authority.

If shareholders are few and accessible, they would have the right as owners to instruct the board on such matters. But for many companies, shareholders are neither few nor accessible and although AGMs might conceivably address this issue, such clarity or expertise from shareholders is ordinarily unattainable. Typically, then, the board holds the only position in the company for which the specific role responsibility (that part not delegated to others) is not clearly determined by a higher authority. (Let me reaffirm here that the board is accountable for everything and any apportioning of what it keeps versus what it delegates affects only its hands-on, direct duties, not the overall accountability.) There is no other legitimate option: the board must as a group make the decision itself, for it has no boss to turn to.

Management’s authority is what is left after the board decides which decisions or levels of decisions it retains. Therefore, the act of designing its own job also answers what the board will delegate to the CEO. Whether it wishes this burden or not, the board as shareholder representative must make the call itself. In other words, directors must carve out the nature of governance in each company’s specific situation. Clearly this kind of board-as-first-mover is a completely different role from being an advisor or accepting that the governance role should be driven, as Kenneth Lay put it, by “what I as CEO expect of a board.”

Governance Considered from the Board’s Position

The weakness of the “best practices” approach to governance improvements lies in the origins of practices themselves. Tradition has led to boards doing certain things and not doing other things. What boards do and don’t do has arisen more from personalities and the curious management-governance inversion as from any grand design about the nature of governance. (By inversion, I mean the de facto superiority of management over governance.) Best practices merely improves endeavors that may or may not be part of a more carefully designed board job. In other words, conscientious effort can be expended learning to do the wrong things better.

Appling the kinds of analyses to governance as those that work quite well in management skips the step of deriving an appropriate role founded on where the board sits in the scheme of things. So it is first necessary to be very clear what a board is for before great advances can be made in what it does.

So a purely best practices approach can be no better than the position analysis (or lack of one) that should have preceded it. Best merger practices would only confuse a clerk’s job. Best filing practices can only clutter a CEO’s job. So the relevant point of departure is to place the board in the scheme of things, then to deduce what must be true from the nature of that position. Only then can the specific tasks compelled by that nature be profitably improved by continually better practices.

The starting point seems unequivocal: The board’s position is squarely between shareholders and management. That position is easily visualized as the familiar three link chain shown on the standard organization chart connecting shareholders to board to management. The positional analysis is founded on the accountability and authority implications of this simple three part sequence.

It is important, then, that we recognize what this chain of command is and what it isn’t. First, it is definitely a hierarchy with authority flowing from top to bottom and accountability flowing back the other way. There are no individuals shown, for no individuals are part of the simple base. For example, there is no chair because the chair is a only a single member of the board. It is the board that has independent existence. Further, there is no CEO shown, for having a chief executive-regardless how ubiquitous the role is and how inadvisable it would be to operate without one-is, nevertheless, a choice the board makes to simplify the authority and accountability of management.

Consequently, I believe that honoring the three link chain of moral and legal authority leads to a number of conclusions about the nature of proper corporate governance so compelling that practices inconsistent with them must be eliminated regardless of how entrenched they are in the conventional wisdom. Here are a few of the conclusions:

Shareholders-not stakeholders-own the company. If, due to the public policy of a sovereign power, ownership is differently defined (as is implicit in the German system), then the ownership is not shareholders alone. But I will refer to shareholders as the owners, since in most jurisdictions (including Canada) there is little legitimate alternative. This assertion does not relieve the company from obligations to other stakeholders, including ethical obligations to communities, employees, and even future generations. It merely reserves ownership status for shareholders. They are the source of ownership authority and, hence, the group to whom the board is directly accountable and with whom the agent-principal bond is shared.

The board-and only the board-is the shareholder representative. The board governs that which owners own, making it the “virtual ownership” on site. As shareholder representative, the board’s primary obligation and its undying fealty must be to shareholders, not to stakeholders in general, not to managers and certainly not to directors’ personal interests. That assertion does not release boards from ethical behavior and obligations to other non-owner stakeholders any more than the fact that I am sole owner of my automobile releases me from obligations to others as I drive.

Only the board is accountable to shareholders. The current flurry of making the audit committee, chair, CEO, CFO and others accountable to shareholders does not increase accountability in the system, but undermines it by parceling out board accountability and by providing routes around the board link in the chain. When the law directs that a unit subordinate to the board (say, the audit committee, as does Sarbanes-Oxley) be this or do that, it has taken the board off the hook.

The only board authority is group authority. The board holds its awesome authority as a group, not as individuals. No individual or subgroup exercises legitimate authority except with active or passive permission of the group. Therefore, the group is as fully accountable for officer and committee decisions as if it were making their decisions itself.

Group authority and accountability can be fragmented by committees. Any external focus on the jobs and responsibilities of committees tends to jump past and often ignore the overall accountability of the board, as if committees are more important in the scheme of governance than the board itself. Focused instead on the board’s group obligation, we would have to conclude that a board can operate without any committees as long as it fulfills its accountability. In practice (and reflected in many codes), however, the concentration is more on committee functions than on board accountability. Glorianne Stromberg, formerly a commissioner of the Ontario Securities Commission, worried that committees might “fragment the decision rights of the board as a whole.” The complement of her worry about decision authority is that holistic board accountability suffers commensurately. It was diagnostic of the prevalent mentality that in the early weeks of the Enron disaster, virtually all press attention was focused on auditors, audit committee, and executives, rather than the board that was accountable for all their actions.

The board is a permanent, active link in the chain of authority. A link in the chain of command is in that position at all times, not simply when things go wrong. The board is not a standby authority that springs into action only when the company is faltering. If the company is failing, it is doing so with full board accountability for its having done so and with authority the board has actively granted or passively allowed. The board is either a link in the chain or it is not. No one would allow a CEO to approach his or her job in this intermittent manner. The job of a driver is to steer the vehicle, not simply to get it out of the ditch when it goes off course.

As a link in the chain of authority, the board is a commander, not an advisor. The board does not exist to advise or assist management, but to empower, charge, and evaluate management. A board might “ask good questions” or advise, but these do not constitute its job. They wouldn’t for a CEO with respect to his or her subordinates and they don’t for a board. A governance-destroying CEO-centrism quickly turns the board’s commanding role into the feckless one of advisor. The board does not exist to react to CEO requests, to have its agenda management-driven, or to be either management’s adversaries or its cheerleaders any more than the CEO’s job exists for these reasons with respect to his or her subordinates.

The CEO works for the board. CEOs typically manage boards far more than boards govern CEOs. Some CEOs treat their boards as just another department to be managed, as Robert Monks observed, as one of the external forces in his or her environmental context. However, since this inversion of rightful authority can only occur if the board allows it, it always signifies board default on its responsibility.

The CEO works for only the board. Even when companies have assigned the chair and CEO roles to two persons, there is a tendency for the chair to stand between the board and the ostensible CEO. I say “ostensible” because the chair becomes the de facto CEO, for the titled CEO then reports to the chair rather than or in addition to the board. When this is the case, the roles have not been separated at all, only obscured. In like manner, the CEO does not work for board committees, but only for the board. Board committees may help the board fulfill its duties, but may not exercise command authority.

The chair works for the board. The board’s position in the chain of authority must imply that all persons or groups (except for shareholders) work for the board; the chair is no exception. If the chair performs poorly, the board is accountable for his or her doing so. If the chair and CEO are in competition, as seems to be a pervading worry, the appropriate role of chair has been misconstrued by the board. The chair should exemplify what Robert Greenleaf called a servant-leader. If the chair personally exercises governance authority, the board is diminished, and frequently rendered subordinate to the chair.

The chair and CEO are separate roles. The chair and CEO roles are not, then, hierarchically related. They both work for the board doing completely separate jobs that require completely different skills. The chair is a staff position (helping the board fulfill its authoritative duties) and the CEO is a line position (actually exercising delegated executive authority). With a more demanding governance role for the board, the non-executive chair job can no longer be seen, as some now do, as simply an emasculated CEO. Of course, the need for a lead director evaporates, since it is but a patchwork solution for chair positions that don’t work when needed most.

The board’s job derives from above, not below. CEOs do not construct their jobs as summaries of jobs below them. They construct the jobs below them as delegated divisions of the total executive accountability with which they themselves are charged. So the CEO’s job is not an extension upward of sub-CEOs’ jobs, but theirs extensions downward from the CEO’s. Similarly, the board’s job is an arm of ownership, not of management or put another way, management is derived from governance, not governance from management. Accordingly, the board’s primary relationship should be with owners, not with management.

The board must add value, but to ownership, not to management. It is common to hear that boards should add value, but the value referred to is usually value to management in the form of advice and contacts. But governance exists to add value to owners’ voice in what they own, not to add value to management. If CEOs want help, they are surely capable of arranging for it without causing the governance function to be compromised to get it.

The position of governance between owners and operators compels these conclusions, but there are numerous others. Board meetings must become truly the board’s meetings rather than management’s meetings for the board. Chair and CEO roles in publicly traded companies must be held by different persons-an evolution in which the U.K. has moved ahead of North America. But chairs must become chief governance officers rather than òberCEOs-failure to do which has frequently rendered the U.K. role separation one in title only. Regulators, investors, directors, and executives must realize that the proper nature of governance is not management one step up, but ownership one step down. Where independence from management is an issue (it might not be in closely held companies), we must finally realize that independence in the parts can never overcome lack of independence in the whole – for example, no amount of audit or compensation committee independence can counterbalance lack of board independence. The list goes on, including such current topics as the inclusion of inside directors in the work of the board.

The unique position of governance deserves a sound scaffold of ideas-governance theory, if you will-that would provide a consistent operating system for boards. Such a careful recrafting of corporate governance is what Caroline Oliver and I have argued in our 2002 book, Corporate Boards That Create Value. For too long, governance, while de jure superior to management in authority, has been de facto inferior to management in methodology, yielding the familiar inversion of authority at the pinnacle of organization.

One of the problems of going back to fundamentals is that it can appear so naïve. The more needed it is, the more the picture that emerges is not an accurate reflection of the “real world.” That is certainly true of the analysis presented here. On the other hand, corporate governance reform-like all true reform-is not a matter of reproducing the world as we have known it, but growing beyond the difficulties inherent in that world by rebuilding the foundation. Until governance systems begin to take boards’ positional imperatives seriously, best practices and evolving codes will continue not to reform governance but merely to tinker with it.