The University Record, November 26, 1997

Budgets and Budgeting at the University of Michigan -- A Work in progress*

Nancy E. Cantor
Paul N. Courant

The topic of this paper is the evolving budget system of the University. Or, more precisely, it is what we believe we should do with the current budget system in order to serve the academic purposes of the University. Thus, the paper is as much about those purposes as about the details of our budgeting system. We begin by articulating a set of principles about the academic purposes of the University and the core commitments of its faculty and students. We present a set of reflections on the current budgeting system (Value Centered Management, or VCM) and we conclude with some proposed changes to it (including changes to its name) that we are ready to firmly stand behind, precisely because they articulate so closely with these core purposes and commitments.

At the same time that we propose these changes, we recognize that budgets live in their details, and some of the changes that we suggest will provoke disagreement and thus require some time for digestion and reflection. We believe that we have a good plan, but we are open to the possibility that others will provide information that may change our minds. Therefore, we intend to announce a "waiting period" for reflection, feedback, and discussion, from now until the end of the semester, after which time there will be genuine need to get on with the business of actual budgeting for next year.

Academic Commitments

Having made clear that at least some part of what we say here is only presumptive (and will not take effect at the stroke of midnight), we start with the parts about which we are most sure - that is, those that concern our academic commitments. In our view, a good budget system serves (not defines) these commitments - what the faculty and students most want to do at Michigan is where we must start -- and so our task is to be clear about the commitments, and then to configure the budget so as to make these "desired doings" as affordable as possible.

Before considering the academic commitments that we believe our budgeting system must support, let us recognize explicitly the importance of faculty compensation and the affordability of student tuition to a successful University budget. Compensation and fee containment, as critical as they are, should not keep us from noting that tangible institutional resources are also important because they afford opportunities for pursuing one's passions, for "doing" one's job (trailing a new idea; creating a new course; sitting with a new colleague). To paraphrase the great moral philosopher and economist, Amartya Sen, quality of life depends on more than the tangible goods a person possesses, it depends upon what a person can manage to do or be -- the valued functionings he or she can achieve in leading a life.1 We must worry as much about what faculty and students can manage to do at Michigan as about what they possess by way of rewards for their functionings.

A good budget system will make obvious (perhaps even transparent) how one's self-interest as a faculty member or student is fed by taking part vigorously in the life of the campus and of our broader community of scholarly activity. A bad budget system will narrow the arenas of our participation by encumbering our efforts to take part in such a way as to drain the motivation and fool us into questing after only the tangible rewards, separate of the opportunities for intellectual and social life, that they can buy.

Following this train of thought, the question we must ask is simply put: what do faculty and students want to do and how can we build a budget system that welcomes this activity? It is here, in trying to address this question, that we run the great risk -- one that psychologists call the false consensus bias -- of taking our own experience and attitudes as representative of others. We have tried to limit that risk by discussing these matters with other faculty, most notably members of the VCM Oversight Committee. Nonetheless, what follows is our assessment, as a psychologist who studies personal goals and well-being and an economist who studies the connections between governments and people in producing human welfare, of what participants in the life of the University want to be able to do.

One way to approach the question of academic purposes or "desired doings" is to consider the commitments that are treasured at Michigan and the aspects of University life which make people want to be here. Phil Brickman, a wonderful Michigan social psychologist, once characterized commitments as living in the realm between what one wants to do and what one has to do. He noted that traditional societies make members' commitments highly predictable, whereas our society has made a religion of equating commitments with choices.2 Therefore, we recognize that we tread on sacred grounds of individualism in following this path.

We suggest three general forms of commitments: (1) those that are local in nature and concern one's own work, (2) those that move out of one's local academic neighborhood to involve collaborations with others, and (3) those in which one may participate only indirectly as they are distributed across the campus and constitute the public life of the University. These are readily concretized by considering the benefits to faculty and students of opportunities to participate in the life of a first-class department or a cutting-edge research project; moving then to the excitement generated when a discipline is challenged in collaborations that cross traditions (be they in jointly taught courses, research, performances, or community-service learning projects), and ending with the expansiveness provided by those opportunities that are held in common for the whole University (from the University Library to the Media Union to the football stadium to Hill Auditorium).

The nice thing about commitments is that they can be real or virtual, and the same benefits to well-being can accrue. We are happier at Michigan when we know that a rare language is still taught here or a concert has taken place, even when our happiness comes only from basking in reflected glory, not actual participation. We take pride in the football team, even though few of us have the talent to participate directly. Faculty and students want to be connected to their disciplinary or departmental homes, to the collaborations that extend beyond those local boundaries, and to their common fate as citizens of the University.

We suggest, then, that commitments to actively participate in these three realms -- the local, the collaborative, and the widely dispersed or broadly shared -- undergird the quality of life of faculty and students on campus. Accordingly, when we move to asking questions about budget systems, we are essentially asking how to make these commitments easily possible. That is, a good budget system uses tangible resources to open doors to individual and collective participation, real or virtual. A good budget system: allows a school or college control over resources so as to be able to make hard choices to support the disciplines and curricula that it values, regardless of external funding sources; provides an interdisciplinary research program with enough security to welcome to its table those whose grants will continue to flow elsewhere; promotes the production of goods and activities that are widely available, even though people will generally differ in their use of them. Of course, the tangible resources available for these purposes are not infinite, so in daily practice a good budget system must support those who have to make thoughtful choices about what opportunities to enhance or forego.

Implications for Budgeting and Budgets

The University of Michigan is enormously complicated, and the set of things that we would like to accomplish is large, varied, and often contested. We, therefore, want our planning, and the associated budgeting models and procedures, to be as helpful as possible as we choose among the many things that we think are terrific.

A good budgeting system helps us to see clearly what is possible, and to organize our decisionmaking well; a bad budgeting system gets in the way. But no budgeting system can solve the problem that there isn't enough money to do all that we want to do, and no budgeting system can solve the problem that there are important (even treasured) disagreements about what we want to do in light of the fact that we can't do everything. This is not to say that budgeting systems are unimportant. They affect what gets looked at and how, and what sorts of information are available when decisions are made.

We would like our budgeting system to follow the general logic of the activities that it supports. The information that it requires, and the information that it provides, ideally should be broadly useful to decisionmakers at all relevant levels (from the individual scholar to the Provost and the President). To the extent that it provides incentives for behavior (and it always will, if only because behavior is affected by the kinds of data that are collected) the system should reward behavior at all three of the levels of commitment and participation that we discussed earlier.

The budget system should make it easy for chairs, deans, directors and the provost to see how their decisions affect faculty and students' abilities to do the work on their desktops, to do collaborative work around the campus, and to participate in the University as a whole. The budget system and budgeting practices should allow faculty and students to decide how, where, and with whom to spend their work time and energy.

For as long as we can remember, faculty and students have had an essentially infinite list of exciting endeavors that they would like to engage in. They have proposed elements of the list to chairs, to foundations, to deans, and to others, and a complicated and deliberative process of negotiation between the desired and the possible has determined which of the many things that we might do actually get done. (That we cannot do everything we want is implied by Harold Shapiro's remark that "there has never been enough money to satisfy the legitimate aspirations of a truly enterprising faculty or administration."3) Our systems of budgeting and planning should make clear what is possible without artificially restricting what is possible. (Alas, even the best budgeting system cannot by itself expand what is possible. In the end, time, energy, and available financial support are finite.)

More important than budgeting systems and the incentives they may create our own shared and contested visions of the many things we in the University are about, and the many ways in which individuals and groups in the University articulate their commitments to academic life here. The Provost, both as chief academic officer and chief budget officer, is a locus for promulgating shared and contested visions; budget systems are part and parcel of that promulgation. But the main conversation between the Provost and the University at large ought to be about the visions; the rest, if we do our job right, is administrative detail.

Strengths and Weaknesses of Michigan's Current Budget System, (VCM)


The current budget system at Michigan, VCM, is a variation of a budgeting and resource allocation system generally termed "Responsibility Centered Management," or RCM.4 The idea and the logic are straightforward: give academic units (schools, colleges, organized research units, etc.) credit for both the revenues that they generate and the costs that they incur. At the margin, by allowing them to keep most of the revenues and by charging them for most of the costs, they will have incentives to enhance the resources available to them to contribute to the University's academic and other missions. Because the units themselves are in the best position, generally, to evaluate the consequences of their fiscal behavior, placing responsibility for revenues and costs in the hands of the units will lead to the units making sound choices about what they want to do in light of what is possible.

RCM systems, including VCM, do very well at supporting the first of the three domains of participation and commitment that we identified earlier -- the commitment to one's own work, participating in the life of one's department (or institute, or museum, or school or college). The incentives that VCM provides enable administrators who are at least fairly close to the substantive action of students and faculty to make well-informed choices over a wide range of possibilities. This is VCM's great strength, and we find the case for keeping this aspect of the system intact to be compelling.

In this context, the purpose of RCM systems is not to minimize cost, but to assure well-informed choices and cost-effectiveness. This formulation takes a broad view of cost (emphatically including demands placed on the time and effort of faculty, students, and staff, in addition to dollar cost) and of effectiveness (emphatically including difficult-to-measure but vital intellectual outcomes, as well as such things as credit hours.) Although there has been considerable confusion on the subject, the idea undergirding RCM-systems is not to maximize profit or to behave like a business. The idea is to make apparent to those who are doing the real work the opportunities that they forego or generate by doing that work in one way or another. Where the decisions to be made are primarily local in impact, the decentralized locations where the work is done are generally the best places to evaluate the consequences, including the costs in terms of activities forgone (what economists call opportunity costs), of those activities.5

Where possible, RCM systems also focus unit attention on considering the full cost of their activities, allowing the units to choose among all feasible uses of the resources that they employ. Budgeting and allocation of space is a good example. Until recently, in almost all universities that we know of, space was treated by its residents as a free commodity, and the interest of every dean, department head, and faculty member was to get as much space as possible. Under RCM systems, where units pay for the operating costs (and sometimes also the capital and long-term maintenance costs, as well) of their space, the units can decide, for example, to give up some square feet in order to support more faculty and students. The costs of space and the opportunity cost of its uses become apparent: space uses real resources that can be used for other things. Under RCM systems, the users of the space are in a position to weigh its value against the other things that the same money could buy. The incentive to pay attention to costs is motivated not by the "bottom line" but by the possibility of reallocating real resources to more valuable uses that is inherent in having the locus of decisionmaking coterminous with the locus of budgetary consequence. Those who are making the decisions can see clearly the full consequences of their choices.

Problems with VCM

At the same time, of course, the marked increase in incentives to attend to both revenues and costs lies precisely at the heart of the objections to VCM that have been raised by many faculty and others, including the ourselves and the President, and leads to a variety of untoward behaviors that are encouraged by RCM-type systems. The major objections to VCM floating around campus (sometimes floating very high) are of five kinds. We review them here, and then propose changes in VCM that deal with the merits of the difficulties while preserving what we see as the most positive elements of the system.

  • 1. The Language of the Bottom Line

    Issues of language lie at the heart of many of the difficulties that many faculty have with VCM. While the framers of VCM had different ideas in mind, the emphasis that the system places on revenue and cost have led many faculty members to infer that the University administration wants them to act like members of a firm whose mission it is to maximize dollar profits. Such behavior would be a disaster for the University. Our superb faculty and student body weren't chosen on their ability or inclination to meet a payroll. They were chosen (and they chose to be here) because they have the ability and passion to use remarkable intellectual and creative talents to understand and shape others' understanding of the world around them. That is a business that does not and should not have a bottom line, and should only be thought of as a "business" in that it keeps us all very busy.

    The language of VCM, with its emphasis on identifying and reducing costs, threatens to distort the evaluation of the "business" of a university. The word "cost" is easily and widely interpreted as applying only to the dollar cost of some minimal version of a defined activity. Thus, to minimize the dollar cost of teaching a particular course or curriculum we could hire part-time lecturers, videotape their lectures, and (if not for contractual obligations) employ teaching assistants who are not graduate student instructors. Of course, we wouldn't have the same course that we have now. We would lose the engagement with active research and creative activity that can only be derived by having faculty engaged in teaching and designing the curriculum. We would lose the benefit both to graduate students and to undergraduates of the mutual engagement of the two groups in each other's learning and teaching. We would have cut the dollar expenditures of providing a course or a curriculum, but at the same time we would have reduced the quality, thereby changing what we produced.

    Faculty quite properly find the story in the preceding paragraph to be terrifying; if that's what concern with "cost" implies, they want no part of it. Nor did the original framers of VCM. Nor do we. The cost of any use of our resources, including our time and attention, is the best alternative use that we could make of those resources. This is the economist's notion of opportunity cost, and if faculty, students, and administrators are to make good decisions, it is essential that they be aware of the opportunity cost of their choices, of the consequences of their choices. We have a responsibility to ourselves to accomplish as much as we can with the resources available to us, and we have a responsibility to taxpayers, tuition-payers, donors, and funders to minimize the dollar cost of what we do. Much of what we do, of course, is provide research and teaching that is at the frontier of human capability. The obligation to engage in those activities as economically as possible -- to minimize the cost of the quality of what we produce, broadly defined -- is emphatically not an obligation to provide the cheapest imaginable teaching and research. The purpose of reducing cost is not to reduce quality. It is to expand the set of things that we can accomplish.

    Our budget system does not focus on the "bottom line." It should, however, help us to focus on allowing us maximum possible scope for participation in the lives and activities of a great University.

  • 2. Support for the financially needy

    RCM systems directly raise the question of the fate of units and activities that cannot expect to be self-supporting. From the very beginning, the Michigan implementation of RCM explicitly emphasized that units were not supposed to be self-supporting. The early planning documents and speeches noted that "many of the most important and highest 'value' activities of the University can never be expected to be self-supporting, and that our collective challenge is to work together as effectively as possible to enable those activities to continue and to flourish." Hence RCM became VCM, with "Value Centered" replacing "Responsibility Center."6 The system was designed so that every unit would require central resources, in the form a "Provost's Allocation," giving the Provost the ability to add to the budgets of units (Responsibility Centers) that make "extraordinary contributions to the achievement of the University's mission or to its excellence or reputation, well beyond what could be expected from its resource base." 7 Additionally, within schools and colleges, "the same logic can be employed by deans or unit heads as they make internal resource allocation decisions." 8

    Despite the explicit commitment to deal with the problem, the basic concern remains: under VCM, the ability to be self-supporting is more highly prized than it was under a centralized system, and this has fostered widespread insecurity. These problems also articulate closely with the problems of language and framing that have been engendered by VCM. Many faculty are uncomfortable, and many are appalled by the application of the language of business to the University, even when that language is accompanied by explicit recognition that we are not a business and do not aspire to become one.

    In any case, it is essential to all of the levels of participation and commitment that we have discussed that the budget system allow flexible resources to support valuable activities that cannot be self-supporting.

  • 3. Incentives for Collaboration

    The third set of objections to VCM stems from its tendency to weaken incentives for activities that require collaboration across units.

    Research. In a centralized system, units have little incentive to seek external funds, including Indirect Cost Recovery (ICR) to support the overhead costs (as distinct from direct costs) of research. To correct this problem, VCM assigns ICR and the costs that it is meant to support to the units where the relevant research is done. A problem arises when the allocation of Indirect Cost Recovery (ICR) across units is not straightforward. Deans have an incentive to ask (or to instruct) their faculty to make sure that grants are run through the home unit, and faculty, not surprisingly tend to do what their deans want. Where there are real costs that can be incurred in one place or another, deans and program directors may engage in elaborate treaties to make sure that no one gets taken advantage of financially.

    It is probably true that in most cases these difficulties have fairly natural solutions, with ICR divided according to faculty fractions and use of space on grants. However, many faculty report that they avoid interunit activities because the negotiation costs may be prohibitively high. Directors of interdisciplinary research units, the Vice-President for Research, and a faculty committee appointed by the Vice-President for Research all report numerous cases where interunit cooperation was difficult. They also tell of cases in which collaborative research was avoided in anticipation of difficulties arising from the revenue and cost aspects of VCM, especially having to do with Indirect Cost Recovery from sponsored research. Deans and other unit heads, who are predisposed to cooperate with each other and with faculty who are jointly appointed or appointed elsewhere, have strong budgetary incentives under VCM to make sure that their faculties do their research at home. What was once normal cooperation across units can now be perceived as subsidy.

    Here are two specific examples that are very troubling:

  • A senior scholar who had good intellectual reasons for moving his lab from one school to another found that both deans expected that grants and the resulting ICR would be attributed to their respective units. This problem is being resolved, but the negotiations have usurped much time and energy, leading the scholar in question to opine that many such arrangements simply won't be worth the effort.

  • A small interdisciplinary unit organized a faculty research seminar drawing on faculty from at least four different schools. The seminar is successful -- good scholars are interacting with each other and learning from each other, and the research questions they address and methods they use are being changed by these interactions. But the unit that bears the cost (staff support, space, equipment, etc.) of organizing the seminar does not expect to see any of the resulting ICR, because the faculty will be rewarded for attributing their grants to the units in which they are appointed.

    Especially vulnerable to these kinds of difficulties are units that rely on faculty whose primary appointment is elsewhere. Especially troubling is the certainty that many faculty will decide that it simply isn't worth it to make the effort to do collaborative work across units. This kind of preemptive behavior -- "gaming the system" -- leads to underestimates of the magnitude of the problems: It is always harder to hear the dog that doesn't bark.

    Of course, in many cases these problems might not arise at all; it behooves unit heads to cut their faculty a little slack, or to use simple rules of thumb to divide overhead, especially when the stakes are relatively small. But there is no doubt that faculty expectations that VCM will extend below the RC level have had a chilling effect in interunit collaborations.

    Teaching. Under the budget system that preceded VCM, units had little incentive to generate enrollments, and hence tuition. Prior to VCM, when a unit reduced its enrollments, there would be no immediate fiscal consequences for the unit, but the lost tuition would reduce the University's resources. One of the great improvements in the new budget system is that it makes the connection between teaching and the overall health of the University clear to everyone who has authority over budgets. However, there is no ideal way to attach tuition revenues to teaching activity. At one extreme, all or most revenue could follow the location of the teaching itself. This setup has a number of potential problems: (1) There is the possibility that some courses will be designed primarily as profit centers, based on expected popularity rather than academic merit. (2) There are incentives for schools and colleges to reconfigure their curricula along mercantilist lines, requiring that their students take many courses at home, and even duplicating offerings of other schools and colleges. On the other hand, there can be real benefits if schools and colleges use the credit-hour incentive to attract students from the campus in general to take advantage of opportunities that they are best qualified to provide.

    The other extreme is to have tuition follow registrations. This solves the problems of mercantilism and potential pandering, but it also removes the incentive to provide broadly-pitched courses to students from other locations on campus. Service teaching becomes a burden that has no offsetting benefit. Moreover, having tuition follow registrations provides an untoward incentive to admit students and have them do as much of their work as possible elsewhere. Both extremes have advantages and pose problems.

    The current policy has tuition following registration, so that incentives with regard to curriculum and teaching are analogous to incentives for research and ICR. Units now have no incentive to provide valuable courses that would appeal to students in other units It would be desirable for interested LSA students to take courses in the School of Art, for example, but under the current version of VCM, there are no mechanisms to support the cost of such teaching. The incentive is for deans to admit students and have them taught elsewhere, rather than to provide courses to students who come from other units.9

    Thus, in both teaching and research, the fiscal incentives under VCM are to minimize the extent to which units' activities provide benefits to other units on campus. We are a university whose distinctive strength derives from collaborative activity across the campus, yet we have instituted a budget system that penalizes deans for supporting such activity and rewards deans who penalize it. The problem is not, as has been suggested, that there is something wrong with our deans. It is that deans, like everyone else, respond to the incentives that they face, and in the domain of support for collaborative activities, VCM has created bad incentives.

  • 4. Underproduction of Public Goods.

    A similar set of problems arises with public goods -- activities whose consequences are broadly distributed across campus. Museums and libraries, the system of financial aid, maintenance of walkways, and public safety are but a few of the many examples of public goods. Another public good, more subtle, is the general appreciation that Michigan is a university that is closely engaged with a stunningly large fraction of human knowledge and activity. The key feature of public goods is that their benefits are widely available (although not, in general, uniformly appreciated) reaching well beyond the people and units that produce them.

    Public goods in the University are of three important types. One is basically infrastructure -- the phone system, building and grounds, central auditors required to meet federal rules on grants and contracts, and a host of other services that are widely used. More immediately related to the commitments and functionings of faculty and students are public goods that stem from interactions across campus of research and teaching. Some of these are the specific kinds of collaboration we have discussed earlier, but others are more subtle, including the value of knowing that the University as a whole is first rate. (Note that general excellence is a public good that has direct and obvious value, in that the general quality of the University is a recruiting tool that is used by everyone. It is widely appreciated in academe that it is much harder to sustain a first-rate department when the University as a whole is not strong.) Finally, University policies and programs are themselves public goods; our financial aid policy, our commitment to support museums on campus; our commitment to diversity in the student body are all examples.

    The famous "invisible hand" of market economies does not work especially well in allocating resources that have a substantial impact external to the frame of reference of the relevant decisionmaker. Economists since and including Adam Smith have understood that extra-market mechanisms are necessary to prevent the underproduction of public goods.10 A variety of such mechanisms is generally employed, including political behavior, behavioral norms, and ad hoc agreements of various kinds. The necessity to produce public goods when markets will not do so lies at the heart of the traditional economic argument for government. It also lies at the heart of the argument for the University's "central government," which in this context is the Provost, the President, and the other administrators who provide central services.

    The way in which the budget is presented under VCM invites deans to calculate whether they are getting their school's money's worth from public goods, both in areas where such attention helps focus good decisionmaking about resource allocation (e.g., business services from the central administration) and in areas where it leads to confusing conflation of narrow financial concerns and broad University policy (e.g., what level of support is appropriate for the Clements Library or the Center for Education of Women). Of course, the level of support for CEW or the Clements are issues that merit serious discussion across the campus. But the frame of reference of the discussion should primarily be the University as a whole.

    The Provost can do three sets of things to prevent the underprovision of public goods. Where the public goods are best provided centrally, The Provost can use the resources at her disposal to provide such goods directly, by allocating resources to museums and libraries, for example, as well as to the physical, technical, and administrative infrastructure of the campus. In some cases, the question at hand is one of policy, such as who should be eligible for financial aid or what rules should apply to social interactions between faculty and students. Here the central actions will be partly budgetary and partly regulatory; in the case of financial aid, they are mixed, in that we have a set of rules and a budgetary commitment to implement them. Finally, the Provost can use resources that she controls to reward units that act on behalf of the greater University, and to penalize units that are excessively inward-directed.

    None of these mechanisms solves the problem of how big the central administration should be, and how involved it should be in the activities of decentralized units. The question of "states' rights" has its analogue in the University, and, as in the larger society, the question raises both constitutional and policy questions. We make this point not because we know the right answer, and not because there are any obvious implications for the budget systems, but because it implies a continuing commitment to examine the appropriate mix between central and decentralized sources of authority and resources.

  • 5. Cost Attributions

    A fifth set of problems arises from the way in which VCM treats the costs of central service units (many of which are themselves public goods), including, among many other things, financial operations, the registrar, plant extension, recreational sports, the President's office and the graduate school. Under VCM, the costs of centrally-provided services (and of some central units that are not primarily service units at all, such as the Bentley Library) are allocated to schools, colleges, research and auxiliary units according to some twenty-four different formulas.11 The formulas are generally based on numbers of students, faculty, and staff, and on expenditures of various kinds.

    These allocation formulas are inevitably somewhat arbitrary, and a remarkable amount of decanal attention has been paid to revising the formulas to more accurately (or more favorably) reflect the pattern of use of the relevant central services. More significantly, although the attribution formulas add a market-like aspect to the funding of central services, nominally attributing the charges for those services in general proportion to the use of the services, the formulas are not used in determining the budgets of the service units themselves. These are still determined centrally, and the formulas are used to allocate the centrally-determined total. Thus, if the Provost should decide to increase (reduce) the budget for, say, the Registrar, other units would see increases (decreases) in their budgets according to the formula used to allocate the dollar costs of running the Registrar's office.

    What appears to be a market-like mechanism really is not. Decisions made centrally have automatic budget implications for the teaching, research and auxiliary units to whom costs are attributed. But there is no operational or budgetary control, and the main incentives faced by deans, (and other unit heads) regarding central service costs, are to complain about the attribution formulas and about the level of costs themselves. Beyond complaint, there is little that the deans can do. Moreover, in some cases the University may be constrained by union contracts and other contractual commitments.

    Perhaps even more significantly, there is also little incentive on the part of the central administration either to reduce cost or to enhance quality. Budget reductions are automatically realized by the teaching, research and auxiliary units according to the attribution formulas. They are not available for reallocation as the result of any deliberative process. Similarly, cost increases can be automatically "funded" via the formulas, dispersed across the teaching, research, and auxiliary units. The service units themselves are neither directly responsible to their implicit "customers," who have no control over the service units' budgets, nor to the Provost and CFO, who set the budget totals but who neither incur cost increases nor realize cost savings.

    What looks at first glance to be something like a market, really has none of the characteristics of decentralized authority and responsibility that make markets work well.

    Our Proposed Solutions

    Most, if not all, of the difficulties that we have discussed above have been recognized from the beginning, by the framers of VCM as well as by its critics. Nonetheless, the power of some of the negative incentives in the system, and the difficulty of counteracting the sets of incentives that tend to weaken the ability of the University to define itself as a university community have become increasingly apparent in the implementation of VCM. In our view, the fundamental flaw is that too much in the current budget system is allocated automatically, and too much of the authority to make choices and decisions is driven by incentives to enhance local wellbeing at the actual or potential cost to collaborative activity and to the University as a whole.

    As we seek to revise the budget system to correct these problems, it is important that we not lose sight of the benefits of decentralized budgeting of both revenues and costs. This decentralization is an important aid to faculty and students making well-informed choices in the vital domains of participation in their own units, and in collaboration with others in their own units. Understanding costs -- calculated as what is forgone when all resources used are fully evaluated -- is an essential aid to making good choices. But to say that costs matter doesn't mean that costs are the only thing that matters. (In this, costs are much like almost everything else in human endeavor.)

    Each of the behavioral problems with VCM identified above has implications for our proposed revisions to the budget system, as outlined below. We also want to reemphasize the fundamental goal of improving the language with which we discuss University budgets and University priorities. The budget system should leave the faculty with maximum freedom to articulate its individual and collective passion for teaching, research, creative expression and service, unfettered by artificial political and budgetary concerns. (Fettered, alas, by real ones.)

  • 1. We propose to change the name of the budget model that we use from "VCM" to "the budget model." Budgets are instruments of management, but they are but one of many, and it is impossible for budgets that use automatic allocation mechanisms to articulate fully the purposes of the University. The phrase "Value-Centered-Management" implicitly claims too much for itself, leading both proponents and detractors to see the issues raised by the choice of budget system as being of higher stakes than they are or can be. Budgeting systems are not themselves academic policy. RCM-type systems do not prevent central initiatives and centralized systems do not prevent local initiatives. There are areas in which fairly automatic modes of resource allocation work well, and there are areas that require deliberation and even politics in the service of wise choices and good management.

    Of course, incentives can and do change behavior, and thus the incentive properties of budget systems, as well as of all of our procedures, whether systematic or ad hoc, bear serious examination. But the rhetoric of much of the campus debate claims too much both for and against VCM. A corollary of Lee Bollingers' recent remark that incentive systems do not create values is that they also do not destroy them. We suggest that it is time to talk about the University's budget model as a set of procedures in aid of management, rather than as the implementation, for good or ill, of a guiding public philosophy.

    We see this changing of name as a step towards returning the focus of campus-wide discussion, choice, and behavior to the substance of research and teaching, creative expression and learning. All of us should be engaged in questioning what it is we should do next, in our own work, in collaboration with others, and as members of the larger University. As stated repeatedly in this paper, the budget system ought to serve us as we pose and answer these questions. Giving it a grand name and grand claims get in the way.

  • 2. We will do what needs to be done to assure that the budget system's automatic incentives in opposition to collaborative work are countered, in both research and teaching, and to assure that the location of these collaborations be dictated by the intellectual merits, not by automatic rules regarding who gets financial credit for what. We do not yet know what financial resources will be required to accomplish these goals, and we have no interest in proposing formal programs -- "central initiatives" -- with formal application mechanisms. Rather, we want to assure our colleagues on the faculty and in the deans' offices that we are prepared to reallocate in the service of collaboration, and that we want to hear about their best ideas in both teaching and research.

    Regarding teaching, we will use the Provost's allocation and other mechanisms to help units to support educational initiatives for their own students and for students around campus. A great university offers many ways in which students in one school or college can benefit from activities undertaken in another. When good ideas require changes in the distribution of teaching activity across schools and colleges, we will undertake to move resources around to follow the ideas. We do not intend to do this by revising the tuition attribution formula. Rather, we will be open to proposals for changes in units' curricula, and responsive to the budgetary implications of such proposals. Because the current set of Provost's allocations derive in large part from "holding harmless" a world that contained a great deal of cross-unit instruction, we expect that teaching units will continue to provide approximately the same levels of teaching of others' students as they have in the past. Although the process cannot and should not be automatic, we will be aware of changes in distribution of teaching across schools and colleges, and we will use this information in determining Provost's allocations.

    On the research side, in collaboration with OVPR and with the active help of deans and directors, we will both provide resources and foster modes of operation that will simplify the assignment of ICR when there are disputes. We will also use the Provost's allocation and other resources to reallocate budgets to support units whose mission it is to foster collaborative research but that are likely to be disadvantaged in getting formal credit for ICR. Our goal here is that faculty, students, and others engaged in research see their advantage in finding the best collaborators and venues for their work, not in "gaming" the system so as to maximize ICR for one set of units or another.

    We intend to enlist the VCM Faculty Oversight Committee, other faculty groups, and the deans and directors in helping us to think through the sorts of procedures that will serve us in assuring that faculty and students who want to work and to teach collaboratively will not be inhibited from doing so by the budget system. We also expect to encourage joint appointments of faculty across schools and colleges as a mechanism to encourage fruitful collaborations.

  • 3. We will work with the schools and colleges to develop long-term plans in which the Provost Allocation for each unit is articulated with the units' abilities to generate support from external sources. These plans should be flexible -- although more concrete in the near years than in the out years. They will reflect the high priority we place on supporting valuable activities that cannot be self supporting and the high priority we place on giving due consideration to the consequences for the broader University community of units' activities. These trajectories for the Provosts' allocations, which we will develop and revise annually in consultation with the units as part of the budget process, will add a deliberative component to the units' planning, allowing them to maximize the value that they get from the revenues that they generate through automatic features of the budget model. We reemphasize that one important principle guiding the Provost's allocation is that we have a happy obligation to support many activities well beyond the levels that can be supported by self-generated tuition, gifts and sponsored research.

  • 4. We believe that the budget system could be markedly improved by removing the entire cost attribution part of the budget model, and returning to a system in which costs and functions that are centrally controlled are also budgeted, evaluated, and defended centrally. (Where costs can be measured accurately, as in the case of utilities and a number of other costs of operating space, we will continue to charge the costs directly to users.) Removal of attributed costs will greatly simplify the budget system, making it more transparent to users. More important, it will provide incentives to the Provost, the Chief Financial Officer, and the other Executive Officers to evaluate the performance of centrally budgeted units and to realize cost savings from those units. Any savings will then become available for deliberate reallocation to support University priorities, notably the academic units. More important still, by taking explicit responsibility for activities that are under central control, the central administration is the obvious place for praise or blame (as circumstances warrant) for the performance of those units.

    Lee Bollinger, Robert Kasdin, and Nancy Cantor are all publicly committed to a full evaluation of central service units. The Advisory Committee on University Budgets, a campus-wide committee of faculty, staff, and administrators, has just proposed a set of procedures to guide evaluation of these units. Whatever the results of that evaluation, its consequences are appropriately felt and identified in the central administration, rather than dispersed across campus via a set of arbitrary formulas. In making this change, we will provide extensive data on the costs of the centrally-budgeted units. Our purpose is to increase the accountability of the central administration for the choices that it makes.

    Because the financial costs of centrally controlled units are now budgeted via attribution to the Teaching, Research, and Auxiliary (TRA) units, when we move the central units' costs back to the central administration, we will have to reallocate the revenues that pay those costs as well. There are a variety of ways that this might be done. Our initial thinking favors increasing the University participation tax (currently a 2% tax on "adjusted expenditures") while adjusting Provost's allocations for all units so that these changes in budget model would not themselves cause any immediate net impact on the resources available to any TRA unit.12 It may be that the pattern of units' expenditures is such that more than one revenue source should be used. It is certainly the case that different kinds of units (Teaching, Research, and Auxiliary) should have different tax rates, and it is possible that certain types of activity (e.g., sponsored research) should be taxed at a different rate from other types.

    Whatever structure is used, Provost's allocations will be adjusted in the following way: During the current fiscal year (FY 1997-98) we will compute what 1997-98 budgets would have looked like under the new budget system, and we will compute what the 1997-98 Provost's allocations would have to be such that each TRA unit would have the same 1997-98 sources available for all functions other than covering attributed costs under both budget systems. That level of Provost allocation (new system, no policy changes) will be used as the basis for our budget planning for future years under the changed budget system. Thus, units will be treated in the future as if the changed budget system had been in effect this year, and their net funds available this year would be the same as determined under the current implementation of VCM. In this way, units will be held harmless with respect to the formal changes in the budget system and the budget model. In general, we expect to effect policy changes via further changes in Provost's allocations, but these will be policy decisions, not artifacts of the change in budgeting procedures.

    We intend to continue to provide data regarding (imperfectly) attributed costs. Information on attributed costs will surely be relevant in determining Provost's allocations. However, and much more important, we will also provide the TRA units with information about the total costs and changes in total costs of each of the centrally budget units. By focusing on these, and on the overall tax rate(s) used to fund those units, we hope to enhance the quality of debate on the substance and cost of what central units provide without those debates being obscured by arbitrary formulas. With the CFO and the President, the Provost's office should be the location for much of that debate.

    The quantity, quality, and scope of centrally-provided public goods are appropriately determined via a deliberative process, subject to discussion and political influences throughout the University community. Reclaiming formal responsibility for financing these services and policies should help to make clear to everyone where the responsibility lies. The central units are also the most promising source of funds to support the collaborative activities we are committed to supporting. We expect, and indeed we will welcome, pressure to cut taxes and to improve efficiency. Governments (and University administrations) ought to be able to explain what they are doing and why; if they can't, they should be voted out of office.

    Commitments and Budgets

    Finally, we return to where we began - with the commitments of faculty and students to activities in these local, collaborative, and shared lives. A good budget system should support the opportunity for all of these commitments, while also emphasizing the necessity of deliberative decisions about how and where to spend tangible and nontangible resources. Presidents, provosts, deans, faculty and students should all be committed to supporting these three domains of activity, and to making wise choices within and amongst them. We should always be looking for what is, in our view, really good and worthy of Michigan's attention, and then we should demand a budget system that is flexible enough to attempt to support it.

    Sometimes, there simply won't be the resources to support such activity, much as we would desire to do so. That is a fact of life and we can comprehend it. We all understand that of the many things that enterprising faculty and students want to pursue, only some small fraction can be supported at any given moment. However, it should not be the case, a priori, that any given sphere of commitments, for example, the local as compared with the publicly-shared, receives special weight in the allocation of resources. Rather, what a good faculty and student body deserves from us, we believe, is a budget system that makes a great idea possible, no matter the locus of its instantiation. Accordingly, the changes that we are proposing, and the aspects of the budget model that we are preserving, are intended to provide enough resources to support good local, collaborative, and public activities, and, even more pointedly, to make us all aware that often we need to be creative in reallocating revenues and in offering funding for activities in those spheres that some would fervently like to pursue, though many would be unlikely to support.

    We do, at the end of the day, live in a world rich in the exceptional activities it supports; not a world altogether described by market principles, in which the marketplace would determine which activities were rewarded and which were not. We want to reward those who can garner revenues for their commitments, and yet simultaneously have some room to embrace those whose commitments make all of us, not only those who can generate revenues, better as a University.

    The revisions to VCM that we are suggesting, including the change in name, represent an attempt to keep touch with those commitments in all three domains of University life that faculty and students value; we want, most of all, to have the resources to be light on our collective feet in supporting innovation - wherever it occurs, be it in a school or college, across units, or in some location that is both open to us all and not closely attended to by any one of us. We want to make sure that each of us who allocates substantial resources is held accountable for the opportunities forgone in any choice we make, even as we celebrate the University's ability to support good ideas wherever they reside. Of course, a budget system cannot ensure our good behavior, but we hope that the proposed changes encourage deliberation, and equalize somewhat the attention to different promising domains of activity in the life of the campus.

    * We are grateful to Lou D'Alecy and the Senate Assembly for providing us with the impetus to put these ideas on paper. Many groups and individuals have helped us to develop our thinking on the University's budget: the VCM Faculty Oversight committee, the Advisory Committee on University Budgets (ACUB), the Financial Affairs Advisory Committee, the OVPR Committee on VCM, The Academic Program Group, the Budget Administrators' Group, our colleagues Lester Monts and Pamela Raymond, many other faculty members, and a number of the Executive Officers, particularly Robert Kasdin and Lee Bollinger. We are especially grateful for the wise counsel and thoughtful commentary of Marilyn Knepp. (Cantor is Professor of Psychology, and Provost and Executive Vice-President for Academic Affairs, at the University of Michigan. Courant is Professor of Economics and Public Policy and Associate Provost -- Academic and Budgetary Affairs, at the University of Michigan.)

    1 See, for example, A.K. Sen, "Equality of What?", in S McMurrin, ed., Tanner Lectures on Human Values (pp. 195-220). Cambridge, England, Cambridge University Press, 1980.

    2 P. Brickman, Commitment, Conflict, and Caring, Englewood Cliffs, N.J., Prentice-Hall, 1987.

    3 Harold T. Shapiro, Tradition and Change, Ann Arbor, MI, University of Michigan Press, 1987, p. 139.

    4 See Edward L. Whalen, Responsibility Center Budgeting, Indiana University Press, Bloomington and Indianapolis, 1991.

    5 It is worth noting that while RCM is not aimed at a simple "bottom line" measure, such as profit, its intellectual underpinnings do rely on the basic logic that economists adduce in support of the efficiency of market systems. Decentralized, self-interested agents, in this case mostly deans, produce outcomes that are good (in the sense of welfare maximization, not profit maximization) for the system as a whole. This is the invisible hand idea, which works well when (1) problems of the commons and interunit relations are relatively unimportant, and (2) problems of distribution -- who is rich and who is poor -- are relatively unimportant. When these problems become important, Adam Smith and most economists since recommend that the pure market model be supplemented by governmental structures -- in this case a host of collaborative and centralizing institutions, including the Provost. For further discussion of VCM as economics, see John G. Cross, "A Brief Review of "Responsibility Center Management," processed. . For Smith's discussion of the subject, see Adam Smith, An Inquiry into the Wealth of Nations, 1776, Vol. II, p. 539, ed. E Cunnan , New York, Putnam's Sons, 1904.

    6 "Value Centered Management at The University of Michigan: Report of the Responsibility Center Management Implementation Team," The University of Michigan, March 1995, p.6.

    7 op. cit., p.11. This is but the first of a long list of principles for central allocation that are articulated. The rest of the list, in summary, includes: Special efforts to improve the quality of teaching and research; Special efforts in interdisciplinary collaboration; Unusual efforts improve the quality of students; Special efforts to achieve university-wide objectives such as undergraduate education, multiculturalism, etc.; Extraordinary outreach activities; Financial hardship due to external constraints; High costs relative to revenue potential; Unusual service burdens.

    8 Provost Gilbert Whitaker, University Record, Jan. 9, 1995.

    9 The current Provost's Allocations were constructed to support the immediate pre-VCM level of cross-unit teaching, as well as the immediate pre-VCM levels of other activities. Units that try to save money by reducing their willingness to teach students from other units run the risk, therefore, of reduced Provost's allocations.

    10 Smith, op. cit. For a more modern treatment showing that public goods will be underproduced in the absence of central government, see William H. Oakland, "Theory of Public Goods," in Alan J. Auerbach and Martin Feldstein, eds., Handbook of Public Economics, v. II, North-Holland, Amsterdam, 1987. The basic argument is straightforward: If some of the benefits of an action cannot be captured by the agent who incurs the cost, agents will not consider these external benefits in making their decisions. Public goods are generally available, so private agents, unless motivated by altruism, will consider only their private benefit from producing them, but will consider all of the cost, and hence produce too little relative to the socially optimal amount.

    11 Auxiliary units include such entities as the University Press and the Institute for Continuing Legal Education -- business-like units that are connected to the University. Their total impact is small, so the TRA units, as we call them henceforth, are essentially Teaching and Research.

    12 "Adjusted Expenditures" are total unit expenditures less fellowships and scholarships, equipment purchases, transfers to other units, tuition grants, and subcontracts in excess of $10,000