Carver's Policy Governance® Model in Nonprofit Organizations
by John Carver and Miriam Carver
Over the last decade or two, there has been increasing interest in the composition, conduct, and decision-making of nonprofit governing boards. The board-staff relationship has been at the center of the discussion, but trustee characteristics, board role in planning and evaluation, committee involvement, fiduciary responsibility, legal liability, and other topics have received their share of attention. Nonprofit boards are not alone, for spirited debate about the nature of business boards has been growing as well. Whatever the reasons for this intense interest in governance, the Policy Governance model for board leadership, created by the senior author, is frequently a primary focus of debate.
The Nature of Governance and the Need for Theory
The Policy Governance model is, at the same time, the most well-known modern theory of governance worldwide and in many cases the least understood. It applies to governing boards of all types—nonprofit, governmental, and business—and in all settings, for it is assembled from universal principles of governance. In this article, we will focus exclusively on its use in nonprofit boards, though many descriptions of its application in business (for example, Carver, 2000a, 2000c) and government (for example, Carver, 1996a, 1997d, 2000b, 2001; Carver and Oliver, 2002) are available elsewhere.
Governing boards have been known in one form or another for centuries. Yet throughout those many years there has been a baffling failure to develop a coherent or universally applicable understanding of just what a board is for. While comparatively little thought has been given to developing governance theory and models, we have seenmanagement of nonprofit organizations transform itself over and over again. Managers have moved through PERT, CPM, MBO, TQM, and many more approaches in a continual effort to improve effectiveness. Embarrassingly, however, boards do largely what they have always done.
We do not intend to demean the intent, energy, and commitment of board members. There are today many large and well known organizations that exist only because a dedicated group of activists served as both board and staff when the organization was a "kitchen table" enterprise. Board members are usually intelligent and experienced persons as individuals. Yet boards, as groups, are mediocre. "Effective governance by a board of trustees is a relatively rare and unnatural act . . . . trustees are often little more than high-powered, well-intentioned people engaged in low-level activities" (Chait, Holland, and Taylor, 1996, p. 1). "There is one thing all boards have in common . . . . They do not function" (Drucker, 1974, p. 628). "Ninety-five percent (of boards) are not fully doing what they are legally, morally, and ethically supposed to do" (Geneen, 1984, p.28). "Boards have been largely irrelevant throughout most of the twentieth century" (Gillies, 1992, p. 3). Boards tend to be, in fact, incompetent groups of competent individuals.
An extraterrestrial observer of board behavior could be forgiven for concluding that boards exist for several questionable reasons. They seem to exist to help the staff, to lend their prestige to organizations, to rubber stamp management desires, to give board members an opportunity to be unappointed department heads, to be sure staffs get the funds they want, to micromanage organizations, to protect lower staff from management, and sometimes even to gain some advantage for board members as special customers of their organizations, or to give board members a prestigious addition to their resumes.
But these observations—accurate though they frequently are—simply underscore the disclarity of the board's rightful job. Despite the confusion of past and current board practices, we begin in this article with the assertion that there is one central reason to have a board: Simply put, the board exists (usually on someone else's behalf) to be accountable that its organization works. The board is where all authority resides until some is given away (delegated) to others. This simple total authority-total accountability (within the law or other external authorities) is true of all boards that truly have governing authority.
The Policy Governance model begins with this assertion, then proceeds to develop other universally applicable principles. The model does not propose a particular structure. A board's composition, history, and peculiar circumstances will dictate different structural arrangements even when using the same principles. Policy Governance is a system of such principles, designed to be internally consistent, externally applicable, and—to the great relief of those concerned with governance integrity—logical. Logical and consistent principles demand major changes in governance as we know it, because these principles are applied to subject matter that has for many years been characterized by a hodgepodge of practices, whims of individuals, and capricious decision making.
Such a change is a paradigm shift, not merely a set of incremental improvements to the status quo. Paradigm shifts are difficult to cope with, since they often render previous experience unhelpful; they demand a significant level of discipline to be put into effect. But if there is sufficient discipline to use the Policy Governance model in its entirety, board leadership and the accountability of organizations can be transformed.
It is important that we underscore this point. Using parts of a system can result in inadequate or even undesirable performance. It is rather like removing a few components from a watch, yet expecting it still to keep accurate time. Unlike the traditional practices to which boards have become accustomed, the Policy Governance model introduces an integrated system of governance (Carver and Carver, 1996; Carver, 1997).
Greater effectiveness in the governing role requires board members first to understand governance in a new way, then to be disciplined enough to behave in a new way. Boards cannot excel if they maintain only the discipline of the past any more than managers of this new century can excel if they are only as competent as those of the past. Does this ask too much of boards? Perhaps it does ask too much of many of today's board members. Yet there are other board members—or potential board members who thus far have refused to engage in either the rubber-stamping or the micromanaging they see on boards—who would rejoice in greater board discipline.
The Policy Governance model requires that boards become far more enlightened and more competent as groups than they have been. If that means losing some board members as the composition of boards goes through change, then the world will be the better for it. The Policy Governance model is not designed to please today's board members or today's managers. It is designed to give organizations' true owners competent servant-leaders to govern on their behalf.
Board as Owner-Representative and Servant-Leader
In the business sector, we can easily see that a board of directors is the voice of the owners (shareholders) of the corporation. It is not always apparent that nonprofit organizations also have owners. Certain nonprofits, such as trade associations or professional societies, are clearly owned by their members. Beyond such obvious cases of ownership, however, it is useful to conceive that community-based agencies in the social services, health, education, and other fields are "owned" by their communities. In neither trade associations nor community agencies is there is a legal equivalent of shareholders, but there is a moral equivalent that we will refer to as the "ownership." Looking at ownership in this very basic way, it is hard to conceive of any organization that isn't owned by someone or some population, at least in this moral sense.
The Policy Governance model conceives of the governing board as being the on-site voice of that ownership. Just as the corporate board exists to speak for the shareholders, the nonprofit board exists to represent and to speak for the interests of the owners.
A board that is committed to representing the interests of the owners will not allow itself to make decisions based on the best interests of those who are not the owners. Hence, boards with a sense of their legitimate ownership relationship can no longer act as if their job is to represent staff, or other agencies, or even today's consumers (we will use that word to describe clients, students, patients, or any group to be impacted). It possible that these groups are not part of the ownership at all, but if they are, it is very likely they constitute only a small percentage of the total ownership.
We are not saying that current consumers are unimportant, nor that staff are unimportant. They are critically important, just as suppliers, customers, and personnel are for a business. It is simply that those roles do not qualify them as owners. They are due their appropriate treatment. To help in their service to the ownership, Policy Governance boards must learn to distinguish between owners and customers, for the interests of each are different. It is on behalf of owners that the board chooses what groups will be the customers of the future. The responsible board does not make that choice on behalf of staff, today's customers, or even its own special interests.
Who are the owners of a nonprofit organization? For a membership organization, its members are the owners. For an advocacy organization, persons of similar political, religious, or philosophical conviction are the owners. There are many variations. But for purposes of this paper, we will assume a community organization, such as a hospital, mental health or family service agency, for which we can confidently say that the community as a whole is the legitimate ownership. In this case, it is clear that in a community organization, the board must be in a position to understand the various views held in the community about the purpose of the organization. In short, if the community owns the organization, what does the community want the organization for?
Traditionally, boards have developed their relationships largely inside the organization—that is, with staff. Policy Governance demands that boards' primary relationships be outside the organization—that is, with owners. This parallels the concept of servant leadership developed by Greenleaf (1977, 1991), in that the board is first servant, before it is leader. It must lead the organization subject to its discoveries about and judgments of the values of the ownership.
We have thus far referred repeatedly to the board and very little to board members; that is intentional. Since we are now establishing the starting point for governance thinking, it is important that we start with the body charged with authority and accountability—the board as a group, not individual board members. It is the board as a body that speaks for the ownership, not each board member except as he or she contributes to the final board product. So while we might derive roles and responsibilities for individual board members, we must derive them from the roles and responsibilities of the board as a group, not the other way around. Hence, board practices must recognize that it is the board, not board members, who have authority.
The board speaks authoritatively when it passes an official motion at a properly constituted meeting. Statements by board members have no authority. In other words, the board speaks with one voice or not at all. The "one voice" principle makes it possible to know what the board has said, and what it has not said. This is important when the board gives instructions to one or more subordinates. "One voice" does not require unanimous votes. But it does require all board members, even those who lost the vote, to respect the decision that was made. Board decisions can be changed by the board, but never by board members.
The Necessity for Systematic Delegation
On behalf of the ownership, the board has total authority over the organization and total accountability for the organization. But the board is almost always forced to rely on others to carry out the work, that is, to exercise most of the authority and to fulfill most of the accountability. This dependence on others requires the board to give careful attention to the principles of sound delegation.
Since the board is accountable that the organization works, and since the actual running of the organization is substantially in the hands of management, then it is important to the board that management be successful. The board must therefore increase the likelihood that management will be successful, while making it possible to recognize whether or not it really is successful. This calls upon the board to be very clear about its expectations, to personalize the assignment of those expectations, and then to check whether the expectations have been met. Only in this way is everyone concerned clear about what constitutes success and who has what role in achieving it.
At this point, we wish to introduce the chief executive (CEO) role. (Policy Governance works in the absence of a CEO role, but the governing job is more difficult than with a CEO.) We are not concerned whether the CEO is called executive director, director-general, president, general manager, superintendent, or any other title. We are, however, concerned how the role is defined and we will use the term "CEO" to reflect the role definition we recommend.
We recommend that the board use a single point of delegation and hold this position accountable for meeting all the board's expectations for organizational performance. Naturally, it is essential that the board delegate to this position all the authority that such extensive accountability deserves. The use of a CEO position considerably simplifies the board's job. Using a CEO, the board can express its expectations for the entire organization without having to work out any of the internal, often complex, divisions of labor. Therefore, all the authority granted by the board to the organization is actually granted personally to the CEO. All the accountability of the organization to meet board expectations is charged personally to the CEO. The board, in effect, has one employee.
It is important that boards maintain a sense of cause and effect with respect to their CEOs. The board creates the CEO; the CEO does not create the board. As the board contemplates its accountability to the ownership, it decides that creating a CEO role will be a key method in fulfilling that accountability. It is true that a founding father or mother will sometimes be the inspiration for a new organization, so that the board then created occurs after rather than before the founder. If the founder becomes the new CEO, it will seem that the CEO is parent to the board. Boards established in this way make a grave error when they mistake an accident of history for a proper view of their accountability. The CEO role, as such, is even in these cases created and governed by the board (see Carver, 1992).
Consequently, in every case, the board is totally accountable for the organization and has, therefore, total authority over it—including over the CEO. We can say that the board is accountable for what the CEO's job is and that the CEO do the job well. But we cannot say the CEO is accountable for what the board's job is and that the board do its job well. Unfortunately, much of current nonprofit practice supports this board-staff inversion. CEOs are expected to tell their boards what to talk about (provide agendas), to pull their boards together when there is dissension, and to orient new board members to their job. Nowhere else in an organization are subordinates responsible for the conduct of the superiors. Yet virtually all nonprofit literature on governance falls into this fallacy of CEO-centrism. "Thus, we argue, the board's performance becomes the executive's responsibility," say Herman and Heimovics (1991, p. xiii), a position we contend excuses and prolongs board irresponsibility.
We have said being accountable in leadership of the organization requires the board (1) to be definite about its performance expectations, (2) to assign these expectations clearly, and then (3) to check to see that the expectations are being met. Traditional governance practices lead boards to fail in most or all of these three key steps.
Board expectations—which are instructions—when they are stated at all, tend to be unclear, incomplete, or a mixture of whole board and individual board member expressions. Board members form judgments of staff performance on criteria the board (as a whole body) has never stated. Regular financial reports report against few or no criteria. Staff members can be seen taking notes of what individual board members say, as if it matters and as if they work for the board members rather than the CEO. Boards decide whether CEO's budgets merit approval when they have never stated the grounds for approval and disapproval. Virtually every board meeting—other than in Policy Governance boards—is testimony to carelessness of delegation and role clarity.
Traditional governance allows boards to instruct staff by the act of approving staff plans, such as budgets and program designs. When the board has approved a staff recommendation, doesn't the resulting approved document become a clear board instruction? Actually, it does not. For example, when a board approves the CEO's personnel policies or budget, does it really mean as an instruction every tiny segment of that document? Does every budget line and the smallest issues of a program plan become a criterion on which the CEO will be judged? Certainly not. Even the most micromanaging board does not go that far. But to what level of detail should the CEO treat the approved document as being a board instruction, therefore a criterion for evaluation? The tradition-blessed habit of board approvals is a poor substitute for setting criteria, then checking that they have been met. Board approvals are not proper governance, but commonplace examples of boards not doing their jobs.
What about the clear assignment of expectations to a person or persons? In conventional practice, boards' delegation to a CEO is frequently compromised by delegating the same responsibilities more than once or by delegating to around the CEO to sub-CEO staff. An example of the former is when a board charges the CEO and a board finance committee for financial decisions. Delegating around the CEO occurs either when a board gives instructions to the financial officer or other person who reports to the CEO or when a board itself judges the performance of sub-CEO staff.
Finally, in the absence of clear instructions or clear assignment, evaluating performance is an exercise in futility. Yet boards receive volumes of information that purports to monitor organizational performance. The sheer amount of information masks the fact that proper monitoring is still not occurring. Because monitoring performance is the systematic disclosure of whether board expectations have been met, monitoring that is fair and incisive can only occur after clearly stated and clearly assigned board expectations.
Using the Ends/Means Distinction
The point was made earlier in this paper that the board is accountable that the organization works. Clearly, the word "works" must be defined; defining it establishes the board's expectations for the organizations, the performance that will constitute success. The board need not control everything, but it must control the definition of success. It is possible to control too much, just as it is possible to control too little. It is possible to think you are in control when you are not. The zeal of a conscientious board can lead to micromanagement. The confidence of a trusting board can lead to rubber stamping. Defining success is a matter of controlling for success, not for everything. How can a board control all it must, rather than all it can?
Boards have had a very hard time knowing what to control and how to control it. Policy Governance provides a key conceptual distinction that enables the board to resolve this quandary. The task is to demand organizational achievement in a way that empowers the staff, leaving to their creativity and innovation as much latitude as possible. This is a question of what and how to control, but it is equally a question of how much authority can be safely given away. We argue that the best guide for the board is to give away as much as possible, short of jeopardizing its own accountability for the total.
What is there to control? In any organization, there are uncountable numbers of issues, practices, and circumstances being decided daily by someone. The Policy Governance model posits that all of these decisions can be classified as those that define organizational purpose, and those that don't. But the model calls for a very narrow and careful definition of purpose: it consists of what (1) results for which (2) recipients at what (3) worth.
Let us define these more fully: Some decisions directly describe the intended consumer results of the organization, for example, reading skills, family harmony, knowledge, or shelter from the elements. Some decisions directly describe the intended recipients of such results, such as adolescents, persons with severe burns, or low income families. Some describe the worth of the intended results, such as in dollar cost or priority against other results.
In Policy Governance, this triad of decisions is called "ends." Ends are always about the changes for persons to be made outside the organization, along with their cost or priority. Ends never describe the organization itself or its activities. For example, the professional and technical activities in which the organization engages are not ends. In a school, for example, which students should acquire what knowledge at what cost are ends issues. Ends are about the organization's impact on the world (much like cost-benefit) that justify its existence.
Any decision that is not an ends decision is a "means" decision. In that same school, the choice of reading program, teachers' credentials, and classroom arrangement are means issues. Most decisions in an organization are means decisions; some are very important means. But even if a decision is extremely important, even if it is required by law, even if it is critical to survival, unless it passes the ends test (designation of consumer results, which consumers, or the worth of consumer results), it is not an ends decision. Hence, means include personnel matters, financial planning, purchasing, programs, services and curricula, and even governance itself. No organization was ever formed so it could be well governed, have good personnel policies, a fine budget, sound purchasing practices, or even nicely planned services, programs or curricula.
The ends/means distinction is critical. Many boards claiming to use the model routinely confuse the Policy Governance meaning of ends and means, thereby sacrificing much of the benefit the model can give. For example, means is not synonymous with "administration" as some have misinterpreted (Herman and Heimovics, 1991, p. 44). Ends is not synonymous with "strategic plan," as others have misinterpreted (Murray, 1994). The ends/means distinction is not comparable to any other distinction used in management or governance; it is not parallel to policies/procedures, strategies/tactics, policy/administration, or goals/objectives. Indeed, ends may include very small and specific decisions about a single consumer, while means may include very important programmatic decisions as well as how a board constructs its committees. The ends/means distinction is exclusively peculiar to Policy Governance (with the possible exception of Argenti, 1993) and, therefore, is governed by Policy Governance principles. In Policy Governance,means are means simply because they are not ends.
Are ends the same as mission? Unfortunately, the answer is usually "no," because mission statements have not traditionally had to conform to the definition we have given ends. Consider the following mission statement of a mental health center: "The mission of the XYZ Center is to be a responsible employer, providing quality mental health services in a cost-efficient manner." This statement—quite acceptable in traditional governance—is entirely means, no ends. This organization can fulfill its mission even if consumers' lives are not any better. In contrast, consider this broad statement of ends: "The XYZ Center exists so that people with major mental illness live productive lives in an accepting community at a cost comparable to other providers." In the latter, unless the targeted group are benefited in the required way, the organization is not successful, no matter how good an employer it is and no matter how much "quality" its services have. Notice that the cost component in the first statement is the cost of staff activity (services), while in the second statement it is the cost of consumer results.
No matter how central ends are to the organization's existence, however, because the board is accountable for everything, it is accountable for means as well. Accordingly, it must exercise control over both ends and means, so having the ends/means distinction does not in itself relieve boards from any responsibility. The ends/means distinction does, however, make possible two entirely different ways of exercising control, ways that—taken together—allow the board to have its arms responsibly around the organization without its fingers irresponsibly in it, ways that for the staff maximize accountability and freedom simultaneously. The board simply makes decisions about ends and means—that is, it controls the organization's ends and means—in different ways, as follows:
- Using input from the owners, staff, experts and anyone in a position to increase the board's wisdom, the board makes ends decisions in a proactive, positive, prescriptive way. We will call the board documents thus produced "Ends policies."
- Using input from whoever can increase board wisdom about governance, servant leadership, visioning, or other skills of governance and delegation, the board makes means decision about its own job in a proactive, positive, prescriptive way. We will call the board documents thus produced "Governance Process policies" (about the board's own job) and "Board-Staff Linkage policies" (about the relationship between governance and management). Both of these categories are means, but they concern means of the board, not the staff.
- Using input from whoever can increase its sense of what can jeopardize the prudent and ethical conduct of the organization, the board makes decisions about the staff's means in a proactive, but negative and boundary-setting way. Because these policies set forth the limits of acceptable staff behavior, that is, the unacceptable means, we will call the board documents thus produced "Executive Limitations policies."
At this point in our argument, we have used the ends/means concept to introduce new categories of board policies. These categories of board policies are exhaustive, that is, no other board documents are needed to govern except bylaws. (Articles of incorporation or letters patent are required to establish the nonprofit as a legal entity, but these are documents of the government, not the board.) We will not discuss bylaws here, except to say they are necessary to place real human beings (board members) into a hollow legal concept (the corporate "artificial person") (Carver, 1995). However, so that we might continue to discuss the concepts represented by the words "ends" and "means," yet distinguish the titles of policy categories, we will capitalize Ends, Executive Limitations, Governance Process, and Board-Staff Linkage.
The negative policies about operational means requires further discussion. Here is the logic: If the board has established Ends and has determined through monitoring that those Ends are actually accomplished, it can be argued that the staff means must have worked. In other words, the means by which Ends were accomplished, though interesting, is of little importance to the board. This logic is largely accurate, but there is an important problem with it. Some means can be unacceptable even if they do work. Means that are effective, but still "unacceptable" are ones that are improper treatment of people or assets, that is, means that are imprudent or unethical. Consequently, although there is no reason for a board to control staff means decisions for reasons of effectiveness, there is reason to control staff means for reasons of prudence and ethics.
Whoever is directly responsible for producing ends must decide which means to use. That is, one must be prescriptive about one's own means. But the board is not charged with producing ends, only with defining them. It is to the board's advantage to allow the staff maximum range of decision-making about means, for skill to do so is exactly why staff were employed. If the board determines the means of its staff, it can no longer hold the staff fully accountable for whether ends are achieved, it will not take advantage of the range of staff skills, and it will make its own job more difficult. Happily, it is not necessary for the board to tell the staff what means to use. In Policy Governance the board tells the staff or—more accurately—the CEO what means not to use!
Therefore, it is the board's job to examine its values to determine those means which it does not want in its organization, then to name them. The board can then tell its CEO that as long as the Ends are accomplished and the unacceptable means do not occur, the CEO can make all further decisions in the organization that he or she deems wise. It is in this way that extensive, albeit explicitly circumscribed, authority is granted to the CEO. Effectiveness demands a strong CEO; prudence and accountability to the board demand that the CEO's power be bounded.
This unique delegation technique has a number of advantages. First, it recognizes that board interference in operational means makes ends harder and more expensive to produce. Therefore, delegation which minimizes such interference is in the board's interest. Second, it accords to the CEO as much authority as the board can responsibly grant. Therefore, there is maximum empowerment inside the organization to harness for ends achievement. Third, it gives room for managerial flexibility, creativity and timeliness. Therefore, the organization can be agile, able to respond quickly to emergent opportunities or threats. Fourth, it dispels the assumption that the board knows better than the staff what means to use. Therefore, the board does not have to choose between overwork and being amateurs supervising professionals. Fifth, in this system all means that are not prohibited are, in effect, pre-approved. Therefore, the board is relieved from meticulous and repetitive approval of staff plans. Sixth, and perhaps most importantly, by staying out of means decisions, except to prohibit unacceptable means, the board retains its ability to hold the CEO accountable for the decisions that take place in the system.
Thus, when we say a board is responsible that its organization works, we simply mean that the organization (1) accomplishes the intended results for the intended people at the intended cost or priority—expressed in the board's Ends policies; and that it (2) avoids unacceptable methods, conduct, activities, and circumstances—unacceptable means expressed in the board's Executive Limitations policies.
Expressing Expectations in Nested Sets
We have established that Policy Governance boards express their expectations for themselves and for their organizations in four categories of board policies: Ends, Executive Limitations (the unacceptable means), Governance Process, and Board-Staff Linkage (the latter two are board means divided into two parts). The separation of organizational values into these categories is a major organizing principle for governing boards. These four categories completely embrace all possible organizational values (except those more pertinent to articles of incorporation/letters patent and bylaws)—no other policies or documents are needed. But another feature must be added to enable the board to address its desired level of specificity within these categories.
To ensure precision as well as completeness in policy-making, Policy Governance provides an additional principle, one which recognizes the varying sizes of issues and values. One Ends statement of a nonprofit board may be that persons without shelter should have adequate housing. Another may be that families with school age children should have housing that allows children of different genders to sleep in separate rooms. It is easy to see that the second example is more detailed, or "narrower," than the first. Notice that these two statements can be pictured as a set of nested bowls, in that the first is a broader value that includes the second one within it. Even more detailed choices exist within the second level, and so on to third, fourth, and more bowls until the specificity reaches a level where Mr. Smith rather than Mr. Jones gets a particular amount of shelter next week.
Now let's illustrate the "nested bowls" concept with an example of unacceptable means. One means value of a nonprofit board may be that the CEO not allow anything imprudent, illegal or unethical. Another may be that unbonded persons may not have access to material amounts of funds. The first example is a broader prohibition than the second, but less specific. Even more detailed "bowls" exist, of course, such as a further proscription against access to more than $5,000 on any one occasion or more than $8,000 cumulatively over a one year period.
Board values about ends and unacceptable means, as well as the board's own means, then, can be stated broadly, or more narrowly. The advantage of stating values broadly is that such a statement is inclusive of all smaller statements. The disadvantage, of course, is that the broader the statement, the greater is the range of interpretation that can be given to it. To take advantage of the fact that values or choices of any sort can be seen as nested sets, the Policy Governance board begins its policy making in all four categories by making the broadest, most inclusive statement first.
The board then considers the range of interpretation that such a statement allows, and determines whether it is comfortable with the statement being given any interpretation that is reasonable. If the board would be uncomfortable delegating such a range, that is a signal that the board must define its words more narrowly, moving into more detail one level at a time. At some point, the board will have narrowed its words to the point that it can accept any reasonable interpretation of those words. Now the board has reached the point of delegation.
As an example, consider an Executive Limitations policy in which the board is putting certain financial conditions and activities "off limits." At the broadest level, the board might say: "With respect to actual, ongoing financial condition and activities, the CEO shall not allow the development of fiscal jeopardy or a material deviation of actual expenditures from board priorities established in Ends policies." That covers the board's concerns about the organization's current financial condition at any one time, for there is likely nothing else to worry about that isn't included within this "large bowl" proscription.
However, most boards would think such a broad statement leaves more to CEO interpretation—even if reasonable interpretation—than the board wishes to delegate. Hence, the board might add further details, such as saying the CEO shall not: (1) Expend more funds than have been received in the fiscal year to date except through acceptable debt. (2) Indebt the organization in an amount greater than can be repaid by certain, otherwise unencumbered revenues within 60 days, but in no event more than $200,000. (3) Use any of the long term reserves. (4) Conduct interfund shifting in amounts greater than can be restored to a condition of discrete fund balances by unencumbered revenues within 30 days. (5) Fail to settle payroll and debts in a timely manner. (6) Allow tax payments or other government ordered payments or filings to be overdue or inaccurately filed. (7) Make a single purchase or commitment of greater than $100,000, with no splitting of orders to avoid this limit. (8) Acquire, encumber or dispose of real property. And (9) Fail to aggressively pursue receivables after a reasonable grace period.
A given board might go into less or more detail than in this example. But in any case, these principles stay intact: The language moves from a broad level toward a lesser level (we showed two levels in the example just given). The values that become policy are generated by the board's deliberations, not approved from a staff recommendation. The board, not the staff, decides what to say and where to stop. No matter where the board stops, the CEO is granted authority to use any reasonable interpretation of the board's words. The board can shrink, expand, or change the content of the policy at any time, as long as it does not judge performance retroactively.
This view of organizational issues—as values that can be specified moving methodically from the broadest to more narrow levels—allows the board to manage the amount delegated. The board is always clear about the authority being given away. The recipient of the board's delegation is always clear about the amount of accountability expected in return. There is a continuum of sizes of issues upon which, in Policy Governance, the board owns the broadest level, then successively smaller levels until it decides to delegate, after which it is safe to allow the remaining decisions to be made by others.
It is often observed by other governance authors that the distinction between what is board work and what is executive work is a naïve distinction. There is no universal rule, they contend, to mark where board policy stops and administration begins. Indeed, they are right as far as traditional governance is concerned, for the conventional approach to the board job is unable to make a policy-administration distinction that holds up in the real world. Policy Governance, however, introduces entirely different, more powerful conceptual tools— rigorous "one voice" clarity of delegation using descending levels of board control within the ends/means context. Even though there is still no predetermined or fixed point where board work automatically becomes executive work, each board using the principles we are describing can establish and, when necessary change, a distinct point of delegation applicable to its own organization. It is at that point, by the values of that board, for that organization, for that time, that governance stops and "sub-governance" begins.
To summarize the policy development sequence, Policy Governance boards develop policies which describe their values about Ends, Executive Limitations, Governance Process, and Board-Staff Linkage. Each policy type is developed from the broadest, most inclusive level to more defined levels, continuing into more detail until the board reaches the point at which it can accept any reasonable interpretation of its words from its delegatee. A step-by-step guide to such development of policy documents is available (Carver and Carver, 1997). Ends and Executive Limitations are delegated to the CEO, who is held accountable by the board for accomplishing any reasonable interpretation of the boards expectations in these areas. Governance Process and Board-Staff Linkage policies are delegated to the board Chair, who is given the authority to ensure that the board governs in accordance with its own expectations of itself, using any reasonable interpretation of the policy language.
Board Discipline, Mechanics, and Structure
It is clear that the Policy Governance model requires a board to govern in an organized, planned and highly disciplined manner. Boards which are accustomed to talking about issues simply because they interest individual board members will find agenda discipline to be a major challenge, as will boards that rely on their staffs to supply their agendas. Not everything is appropriate for board discussion just because it is interesting or even because the staff wants the board to make the decision. Matters that have been delegated to the CEO should not be decided by the board or by board committees, for in making such decisions, the board renders itself unable to hold the CEO accountable.
Policy Governance boards know that their job must result in the production of three deliverables. (1) The first deliverable is a systematic linkage between the organization and the ownership. This is not public relations. The board connects with the ownership in order to ascertain the range of ownership values about the purpose of the organization. If the board is to make Ends decisions on behalf of the owners, it must know what the owners in all their diversity think. (2) The second deliverable is written governing policies in the four areas, using the principles we have described. (3) The third deliverable is the assurance of organizational performance, that is, performance which can be shown to be a reasonable interpretation of the board's Ends and Executive Limitations policies.
We use "deliverables" to mean job products, outputs, or values-added. Since these summarize the purpose for the board's job, producing these deliverables is what board meetings are for. In fact, the list of job outputs can be considered to be a perpetual job description, for every agenda is an instance of the board's working to perform its job. A board can decide how much, in what detail, and at what level of excellence it will pursue its perpetual agenda in the ensuing year. By doing so, it takes control of its own agenda, rather than allowing its agenda to be staff-driven. Establishing its own job description and the longterm or midterm agenda is recorded as one of the board's Governance Process policies. As we shall shortly point out, if the board sketches its annual agenda only broadly, the specifics will be filled in by the board Chair, who is charged with taking care of Governance Process details.
Accordingly, the board must plan meetings that enable and guarantee the production of these deliverables. Being entertained or intrigued by staff jobs is no substitute for the board's accomplishment of its own job. While the board is entitled to any information it wants, it must be aware that collecting information about staff activities and even conscientiously listening to many staff reports does not substitute for governance. Let us again reiterate that the board, not the staff, is responsible that a board's meetings fulfill its governance responsibilities.
In taking responsibility for its own performance, the board confronts the difficulty of acting responsibly as a group of equals. Since the board is by definition a group of peers, no one has authority over anyone else. The first action of a group of peers is to create a position of Chairperson—a first among equals—to help it stay on task. Although it is important that each board member continue to take responsibility for the board's group behavior, the board grants the Chair extra authority required to make rulings that keep the board on track. To stay consistent with the superior role of the board as a group, however, in Policy Governance the Chair only has authority that is within a reasonable interpretation of the board's policies on Governance Process and Board-Staff Linkage. Hence, the Chair is truly the servant-leader of the board (Carver, 1999).
It is usual for nonprofit boards to expect the Chair to supervise the CEO, but in Policy Governance there is no need for the Chair to have authority over the CEO. Only the board has authority over staff operations, and it exercises that authority through carefully crafted policies. It is not only unnecessary, but harmful for the Chair to tell the CEO what the board wants, for the board speaks for itself. Consequently, both the Chair and the CEO work for the board as a whole, but their roles do not overlap because they are given authority in different domains. The Chair's job is to see to it that the board gets its job done—as described in Governance Process and Board-Staff Linkage policies. The CEO's job is to see to it that the staff organization gets its job done—as described in Ends and Executive Limitations policies.
Board Treasurers, as commonly used, threaten CEO accountability as well as the one voice principle. Treasurers are typically expected to exercise individual judgment about the financial dealings of the organization. But Policy Governance boards do not allow Treasurers to exercise authority over staff. (Rendering an official judgment of performance against one's own individual criteria has the same effect as exercising authority.) By creating a role with supervisory authority over the CEO with respect to financial management, the board cannot then hold the CEO accountable for that topic. The board should accept responsibility for financial governance (setting policy, then comparing performance) and require the CEO to be accountable for managing finances so that performance compares favorably to policy. The typical use of a Treasurer, when a Policy Governance board is required by law to have one, is to assist the board in making financial policy, never to judge CEO compliance against the Treasurer's own expectations. For more thorough treatment of the board's role in financial oversight, including commentary on the Treasurer and finance committee, see Carver (1991, 1996b).
In keeping with the "one voice " principle, the board can allow no structures or practices in which board members or board committees exercise authority over staff, any function of staff, or any department of staff. Typical nonprofit boards have a myriad of traditions that violate the one voice principle, such as placing the Chair between the board and the CEO. So it is common for boards to underestimate the amount of board member interference in operations. Such interference, even when well-intended, undermines the board's ability to hold the CEO accountable, for the CEO can argue that his or her actions were taken in compliance with a board member instruction.
Advice is a concept often carelessly used in nonprofit boards. This seemingly innocuous and well-intended practice can have the same deleterious effect as direct instruction by individuals or committees. It is common for the board, board committees, or individual board members to give advice to staff. But advice, if it is really advice, can be rejected. If staff has any doubt that advice given by the board or one of its components cannot safely be turned down, the clarity of board-to-staff delegation will be undermined. Policy Governance boards refrain from giving advice or allowing their members to give advice unless advice is requested. This protects the board's ability to hold the CEO accountable for his or her own decisions. The CEO and any of the staff can request advice if they need it, and they can request it from wherever they wish.
Traditional boards frequently create committees to assist or advise the CEO or staff, such as committees on personnel, finance, program, property maintenance, and other such staff means issues. In Policy Governance, such committees are illegitimate. They constitute interference in the CEO's sphere of authority and accountability, and damage the board's ability to hold the CEO accountable.
If, for example, the staff wishes to have an advisory committee, it is perfectly free to create one, then to use the advice or not as it deems wise. If, however, the board controls the mechanism of advice, a very different relationship between advisors and advisees is established. The wisest route is for the board to govern and leave advice and advisory mechanisms to the staff's own initiative. This way the staff gets all the advice it needs, role clarity and accountability are maintained, and board members are frequently spared unnecessary work.
Policy Governance boards use committees only to help the board to do its own job. Hence, a committee which explores methods of ownership consultation about Ends options is legitimate, as is a committee that studies possible sources of fiscal jeopardy that the board might address in an Executive Limitations policy. But a human resources committee that advises on or intervenes in personnel issus is not. To request advice or assistance with one's own job is acceptable and does not compromise accountability, but to foist help or advice on subordinates is not only unnecessary but destructive of accountability as well.
Policy Governance takes seriously the normally rhetorical assertion that boards be visionary and provide long term leadership. The discipline required for this challenge cannot be overstated. In fact, Policy Governance has been criticized as a "heroic board" model that is romantically idealistic! Yet boards do, in fact, have a critical job to do; no amount of helping staff can substitute for getting its own job done. Boards must persevere with the arduous, complex task of describing purpose and ethics/prudence boundaries. Forming those values into clear policies is far harder than telling the staff how to do its job. Speaking proactively for the ownership requires strong commitment not to take reactive refuge in rituals, reports, and approvals.
This requires board member expertise relevant to governance, not management. Board members should no longer be recruited based on their having skills that mirror the skills of staff. Governance excellence requires members who can think conceptually and with a long term perspective, able to welcome a diversity of opinions but abide by group decisions. They must be able to speak on behalf of the ownership rather than merely from their own or some splinter group perspective. They must place organizational accountability above personal gratification. They must be able to view the board's task of assuring performance at arm's length—through setting expectations (using the ends/means principle and values viewed as descending "bowls"), delegating pointedly (to a CEO if possible), and monitoring. And it is to the function of monitoring or evaluation that we turn now.
Evaluation of performance is not extraneous to the board's job. It is as integral to the board's job as it is to any manager's. But, as we have shown, proper evaluation is impossible unless the board has first stated its expectations and assigned them to a specific delegatee. That is, evaluation of staff performance cannot occur appropriately unless the board has done its job first.
Moreover, if the board has a CEO, the results of proper evaluation of organizational success is the only fair evaluation of CEO performance. Since the CEO's job is to see to it that the organization meets the board's expectations, there is nothing more and nothing less to evaluate when assessing the CEO. Thus, the board's evaluation of organizational performance is the same as board evaluation of CEO performance (Carver, 1997a). Monitoring the evaluative data, as we shall see, is an ongoing activity—perhaps as frequently as monthly—and the board may wish to have a formal evaluation of the CEO once each year. However, the CEO's formal evaluation is only a summary of the accumulated monitoring data, not something in addition.
But let us consider the monitoring or evaluative information itself. Not all information is useful in monitoring performance. There are two types of information that are useful for other purposes, but not for monitoring: one is information for board decisions, the other is information simply to satisfy board members' casual interest. To examine evaluation or monitoring, we must first separate out these two types of information, for they do not qualify as monitoring against pre-established criteria.
First, information for board decisions is needed in order for the board to make wise policy in the first place. To create policies that are both realistic and demanding, boards require information from a variety of sources. These sources include staff, owners, experts, associations to which the board may belong, and others. This information is required for the board's own decision-making and does not judge staff accomplishment. Boards should invest a great deal of energy in gathering wisdom, spending perhaps half their time in becoming educated. So information for board decisions is essential for board performance, but not for monitoring staff performance.
Second, information for board interest is information about the organization or its environment that is not useful for board decision-making, but is of political, social, or technical interest to board members. This information does not include data that directly measure the degree of staff performance on board expectations, for that would qualify it to be called true monitoring information. This kind of information is incidental to the board's job of monitoring, but comprises most of what most traditional boards receive. There is nothing wrong with boards getting all the incidental information they want, but there is something very wrong with the delusion that they are at that time doing their job. In traditional governance, most staff reports, including most financial reports and reports that purport to be "evaluation" are incidental information simply because they are not data compared with previously stated board criteria.
Monitoring or evaluative information must speak directly to whether board expectations are being fulfilled. Consequently, it is always related to expectations set by the board in its Ends and Executive Limitations policies. This discipline not only makes it unnecessary for the board to trudge through the mountains of data staff are able to assemble, but it keeps evaluation fair. After all, it is only right that the CEO should know ahead of time the criteria on which he or she will be judged. Since monitoring information is only that information that describes actual performance compared to expected performance, it is evident that most reports collected, examined and approved by traditional boards constitute interesting information, but cannot be said to be effective monitoring reports. For example, boards that gravely approve (or accept) financial statements thinking they have thereby exercised fiduciary responsibility are simply engaging in a meaningless ritual, for without criteria they don't even know what in those reports would have been disapprovable.
When monitoring is defined as we have done here, reports tend to be straightforward and transparent. Each board member can follow the link from board criteria to management data, for the report is not cluttered with incidental information. Monitoring is not nearly as difficult or time-consuming when boards know what performance they are expecting to see proven. Monitoring is thus more exact and, simultaneously, requires negligible board meeting time. In fact, we recommend that monitoring data be mailed to board members, thereby preserving valuable meeting time for board education and deliberation. Getting monitoring largely out of board meetings allows those meetings to focus on creating the future rather than reviewing the past, because inspection of the past is now safely routinized. For each Ends and each Executive Limitations policy, the board will have set a frequency and a method of monitoring, after which the process runs automatically. The choice of method will be a report from the CEO, judgment by a disinterested party (for example, an auditor), or—less frequently—direct board inspection of organizational practices or circumstances. It turns out to be rare that monitoring needs to be discussed in the board meeting, except for board members to affirm that they have received and read the mailed reports.
To illustrate the nature of what is reported in a Policy Governance monitoring report, we will use two items from an Executive Limitations policy already shown. In that policy, among other unacceptable means, the CEO was told he or she cannot (1) expend more funds than have been received in the fiscal year to date except through acceptable debt and (2) indebt the organization in an amount greater than can be repaid by certain, otherwise unencumbered revenues within 60 days, but in no event more than $200,000. Here is what the monitoring data might look like for these two provisions: Item 1: Through the end of May, $3,694,800 has been expended. Receipts in the same period were $3,654,728. The shortfall of $40,072 was offset by a $60,000 short term loan. Item 2: Total debt is a 45 day working capital loan for $60,000 incurred on May 25. Revenues of $75,000 from our foundation grant, guaranteed by letter of May 5, are not otherwise encumbered and will be used, in part, to retire the debt prior to due date.
Notice that the data are rather bare-bones, only enough to answer the question, unobscured by incidental information. Board members should adopt a "prove it to me" attitude, so if the information submitted is insufficient to convince them, then more detail can be added. But the detail must be such that directly address the criteria. For example, what data prove the "not otherwise encumbered" statement? Obviously, the complexities of some organizations will cause the monitoring data to have more facets than in our simple example. Even then, however, the reported data should be as brief as possible and maintain a razor-sharp connection to the policy-based criteria being monitored. If more interesting, explanatory information, other than that directly addressing the criteria, is desired by the board or offered by the CEO, it should not clutter the monitoring report, but be distributed separately. Board members can know anything they wish, but they should never be in doubt about what is disclosure of performance on the board's criteria and what is not.
Using similar criterion-focused reasoning, when the board seeks to evaluate itself, it compares its actual behavior and accomplishment with the behavior and accomplishment it committed to in its Governance Process and Board-Staff Linkage policies (Carver, 1997b). Policy Governance boards tend to self evaluate on a frequent basis—we recommend every meeting—because a more sophisticated system requires continual tending.
Because in Policy Governance the board is in charge of its own job, board meetings become the board's meetings rather than management's meetings for the board. Board meetings occur because of the need for board members to learn together, to contemplate and deliberate together, and to decide together. Board meetings are not for reviewing the past, being entertained by staff, helping staff do its work, or performing ritual approvals of staff plans. As a result, many board meetings may not look like traditional board meetings at all, but learning and studying sessions or joint meetings with other boards, particularly in communities where boards rarely talk with each other.
The CEO is always present, but is not the central figure. Other staff might be present when they have valuable input on matters the board is to decide. For community boards, with rare exceptions meetings would be open—not to please the law, but because a board commitment to transparency. The board is not merely a body to confirm committee decisions, but the body that makes the decisions. Board committees might be used to increase the board's understanding of factors and options, but never to assume board prerogatives or remove difficult choices from the board table. In contrast to the old bromide that "the real works takes place in committees," in Policy Governance the real work takes place in the board meeting.
Board meetings should thus be more about the long term future than the present or short term future . . . more about ends than means . . . more about a few thoroughly considered large decisions than many small ones. And by their very character, meetings should demonstrate that the board's primary relationship is with owners, not with staff.
The Policy Governance model recognizes that any governing board is obligated to fulfill a crucial link in the "chain of command" between owners—whether legal or moral in nature—and operators. The board does not exist to help staff, but to give the ownership the controlling voice. The board's owner-representative authority is best employed by operating as an undivided unit, prescribing organizational ends, but only limiting staff means, making all its decisions using the principle of policies descending in size. The model enables extensive empowerment to staff while preserving controls necessary for accountability. It provides a values-based foundation for discipline, a framework for precision delegation, and a long term focus on what the organization is for more than what it does.
The Policy Governance model provides an alternative for boards unhappy with reactivity, trivia, and hollow ritual—boards seeking to be truly accountable. But attaining this level of excellence requires the board to break with a long tradition of disastrous governance habits. And it offers a challenge for visionary groups determined to make a real difference in tomorrow's world.
Argenti, John. Your Organization: What Is It For? London: McGraw-Hill Europe, 1993.
Carver, John. "Redefining the Board's Role in Fiscal Planning." Nonprofit Management and Leadership, 1991, 2 (2), 177-192.
———. "The Founding Parent Syndrome: Governing in the CEO's Shadow."Nonprofit World, 1992, 10 (5), 14-16.
———. "Shaping Up Your Bylaws." Board Leadership, July-Aug 1995, No. 20, 4-6.
———. "Policy Governance Model Views Citizens as Owners." Nation's Cities Weekly, January 29, 1996a, 5.
———. Three Steps to Fiduciary Responsibility. The CarverGuide Series on Effective Board Governance, No. 3. San Francisco: Jossey-Bass, 1996b.
———. Board Assessment of the CEO. The CarverGuide Series on Effective Board Governance, No. 7. San Francisco: Jossey-Bass, 1997a.
———. Board Self-Assessment. The CarverGuide Series on Effective Board Governance. No. 8. San Francisco: Jossey-Bass, 1997b.
———. Boards That Make a Difference, 2nd edition. San Francisco: Jossey-Bass, 1997c.
———. "Reinventing the Governance in Government: The Next Frontier for City Councils." Nation's Cities Weekly, January 27, 1997d, 10.
———. The Unique Double Servant-Leadership Role of the Board Chairperson. Voices of Servant-Leadership Series, No. 2. Indianapolis: The Robert K. Greenleaf Center for Servant Leadership, 1999.
———. "The Opportunity for Re-inventing Corporate Governance in Joint Venture Companies," Corporate Governance � An International Review, 8 (1), January 2000a, 75-80.
———. "Remaking Governance: The Creator of 'Policy Governance' Challenges School Boards to Change," American School Board Journal, 187 (3), March 2000b, 26-30.
———. "Un nouveau paradigme de gouvernance: un nouvel équilibre entre le conseil d'administration et le chef de la direction", Gouvernance: Revue internationale, 1 (1), Printemps 2000c, 100-108.
———. "A Theory of Governing the Public's Business", Public Management, 3 (1), March 2001, 53-72.
Carver, John, and Miriam Carver. Basic Principles of Policy Governance. The CarverGuide Series on Effective Board Governance, No. 1. San Francisco: Jossey-Bass, 1996.
———. Reinventing Your Board. San Francisco: Jossey-Bass, 1997.
Carver, John, with Caroline Oliver. Corporate Boards That Create Value: Governing Company Performance from the Boardroom. San Francisco: Jossey-Bass, 2002.
Chait, Richard P., Holland, Thomas P., and Taylor, Barbara E. Improving the Performance of Governing Boards. Phoenix: American Council on Education and The Oryx Press, 1996.
Drucker, Peter F. Management: Tasks, Responsibilities, Practices. New York: HarperCollins, 1974.
Geneen, Harold S. "Why Directors Can't Protect the Shareholders." Fortune, 1984, 110, 28-29.
Gillies, James. Boardroom Renaissance. Toronto: McGraw-Hill Ryerson and The National Centre for Management Research and Development, 1992.
Greenleaf, Robert K. Servant Leadership. New York: Paulist Press, 1977.
———. The Servant as Leader. Indianapolis: The Robert K. Greenleaf Center, 1991.
Herman, Robert D., and Heimovics, Richard D. Executive Leadership in Nonprofit Organizations. San Francisco: Jossey-Bass, 1991.
Murray, Vic. "Is Carver's Model Really the One Best Way?" Front & Centre, Sept. 1994, 11.