The gist of 20th century thinking management control is expressed by the mantra: let the managers manage; make the managers manage. Responsibility budgeting is the stock answer given by students of management accounting and control to the question of how to empower managers to manage and, at the same time, motivate them to use their collective intelligence to create value through exchange in product and financial markets and by establishing and sustaining mutually beneficial relationships with customers, suppliers, and especially other members of their organizations.

In this article explain the nature of responsibility budgeting, its intellectual justification, its antecedents, and its present and future use. This is not a straightforward task. We cannot simply explain how responsibility budgeting is used and how it works. Responsibility budgeting makes sense only as a part of a framework of structural, procedural, and monitoring/reporting relationships. We must, therefore, also explain the framework that gives it utility and power. At the same time, responsibilities budgeting and accounting, or their functional equivalents, make an essential contribution to the efficacy of this broader framework of relationships. One cannot arbitrarily mix and match administrative relationships and expect that the outcome will be productive. The efficacy of administrative relationships depends upon their congruity with each other as well as with the purposes and products of the entity in question and the productive and information processing technologies available to it.

Governance Arrangements, Administrative Processes, Contractual Relationships

All governance arrangements and administrative processes are primarily mechanisms for motivating and inspiring people to serve the policies and purposes of the organizations to which they belong. This means that all governance arrangements and all administrative processes can be treated as relationships and that administrative design and implementation can be thought of as negotiating and maintaining those relationships.

One way of describing relationships uses contractual language and talks about principals and agents. This language implies a hierarchical relationship, in which a nominal subordinate (agent) serves the purposes of a superior (principal). On the presumption that behavior is largely self-interested, principal-agent relationships are problematic (give rise to agency costs) only where (a) the efforts of the agent cannot be perfectly observed; (b) the interests of agent and principal diverge; and (c) agents pursue their own interests, i.e., behave opportunistically.

One of the key goals of governance arrangements and administrative processes is the minimization of agency costs. Of course, agency costs also include all resources used to reduce divergences of interest, i.e., identifying collectively beneficial relationships, negotiating contributions, and devising procedures for monitoring performance and sanctioning defectors. Included here are a whole panoply of activities extending from the employment of security guards to the design and implementation of new or reconfigured accounting and reporting systems. Hence, minimizing agency costs means minimizing the sum of costs that result from opportunistic behavior plus the costs of avoiding or controlling that behavior (Zimmerman, 1995). Economic theory tells us that we find this optimum where the marginal costs of controls equal their marginal benefits, as shown in Figure 1.1 (from Breton and Wintrobe, 1975).

Traditional or Weberian bureaucracies rely on rules to govern or prevent opportunistic behavior. In other words, principals specify in detail what agents must do (or must not do), carefully monitor their actions, and sanction all deviations accordingly. The problem with this approach is that agents often have better information about some things than do principals. Principals hire agents because of their superior expertise and to spare themselves the burden of being perfectly informed about every aspect of an organization's operations. In neither case will principals have the knowledge needed to specify in detail what the agent should do without thereby sacrificing performance. This means that rules are not always a wholly satisfactory solution to the principal-agent problem. It is this fact that makes the application of agency theory to the public sector especially important, for it is in the public sector that the opportunity costs arising from detailed rules often seem highest.

Organizational economists do not generally advocate more rules as a way to control opportunistic behavior. Rather, they stress two alternative approaches. One is to improve principals’ abilities to monitor agents. This is often referred to as improving "transparency." The second is to seek ways to align the incentives of agents with principals’ interests. This is the preferred approach of organizational economists and managerial accountants.

A Historical Digression: From Traditional Bureaucracy to the M-Form Organization

Responsibility budgeting became a codified practice beginning with Peter Drucker's exposition in the Concept of the Corporation in the 1940s. Over the past half century, the practice has been elaborated upon in the expansive accounting literature on managerial control and in the literature on strategic management. However, it antecedents are found in the in the organizationl innovations of the second industrial revolution of the late 19th and early 20th centuries.

Once upon a time most large-scale organizations in America were organized like railroads. Operating responsibility was delegated on a geographic or site basis, rather than a line of business basis. Regional chiefs reported to an agency head. Small armies of administrative staff specialists also reported to agency heads. Their job was to gather and process quantities of data for agency heads to use to coordinate activities, allocate resources, and set strategy. Subsequently mprovements in information processing, especially in the realms of accounting and finance, had the effect of increasing the relative efficiency of coordinating organizational activities and the flow of materials through arm's length relationships (as opposed to direct supervision), making it possible to avoid some of the opportunity costs inherent to rule-based governance systems.

The administrative system developed by Alfred Sloan and Donaldson Brown at General Motors in the 1920s demonstrated the maturity of these innovations. Sloan is best known for the multi-product or M-form organizational structure, in which each major operating division serves a distinct product market. Short-run integration under Sloan’s system was achieved via buyer-seller relationships between GM’s five automotive divisions and the divisions making automotive components (e.g., Fisher Body or Delco-Remy). Longer-run integration was achieved via the capital budgeting system devised by in 1923 by Donaldson Brown, GM’s chief financial officer

GM’s operating divisions were managed entirely by the numbers from a tiny corporate headquarters, using the DuPont system of financial control, also devised by Brown. Under this system, each division kept its own books, and managers were evaluated in terms of a return-on-assets target. The operating division managers continued to rely on control by rules and standard operating procedures and detailed resource-requirements plans. Sloan, however, believed that it was inappropriate, as well as unnecessary, for top managers at the headquarters level to know much about the details of division operations. If the numbers showed that performance was poor, it was time to change the division manager. Division managers with consistently good numbers got promoted, ultimately to headquarters. The divisional form of organization is not only a device to resolve a span of be remotely controlled by the numbers from a strategic apex. This is essentially the set of arrangements that Drucker described in The Concept of the Corporation.

Responsibility Budgeting

Responsibility budgeting is now the most common remote control system used by large-scale organizations in the private sector. Within the accounting literature, agency theorists (e.g., Zimmerman 1995) tend to interpret responsibility budgeting as a practice for structuring the contractual relationship between providers of economic resources (principals) and those who apply those resources in economic activity (agents). The broad outline of this relationship is one where substantial decisional authority is decentralized to agents, within the context of well-specified rules determining how agents will be rewarded for their efforts. Rewards are to be based on economic quantities of interest to principals, such as returns on capital employed. According to this perspective, the management process mainly involves acquiring and deploying assets and, to influence this process, principals must establish a consistent set of delegated decisions (grants of authority to acquire assets), performance measures (resulting from the use of assets by agents), and rewards (incentives for the agent to acquire and utilize assets in the principal's interests).

While private businesses were quick to learn bureaucratic control from government, governmental organizations have been slow to adopt remote control systems. The biggest difference between government budgets and responsibility budgets is that government budgets tend to be highly detailed spending or resource acquisition plans, which must be scrupulously executed just as they were approved (Thompson and Jones, 1986). In contrast, operating budgets in the private sector are usually sparing of detail, often consisting of no more than a handful of financial targets. As we noted earlier, Sloan of General Motors, one of the fathers of responsibility budgeting, believed it was inappropriate for corporate managers know the details of responsibility center operations. The notion that responsibility centers should be managed at arm's length, by the numbers, from a small corporate headquarters, reflects the effort to delegate authority and responsibility down into the organization. As the OECD report, Budgeting for Results: Perspectives on Public Expenditure Management (1995), explains, delegation of authority means giving agency managers the maximum feasible authority needed to make their units productive -- or, in the alternative, subjecting them to a minimum of constraints. Hence, delegation of authority requires operating budgets to be stripped to the minimum needed to motivate and inspire subordinates. Under responsibility budgeting the ideal operating budget would contain a single number or performance target (e.g., a production quota, a unit cost standard, or a profit or return on investment target) for each administrative unit/responsibility center.

In responsibility budget formulation, an organization's policies, the results of all past policy (capital budgeting, see Thompson, 1997) decisions, are converted into financial targets that correspond to the domains of administrative units and their managers (Anthony and Young, 1994: 19). In responsibility budget execution, operating performance is monitored and subordinate managers are evaluated and rewarded. Operating performance targets must be expressed in financial terms. This makes it possible to make comparisons across unlike responsibility centers, thereby permitting the relative performance of managers to be evaluated and increasing the motivational efficacy of internal competition. In traditional responsibility budgets this also has the effect of keeping higher levels of administration ignorant of operating details, thereby discouraging them from meddling in the affairs of their responsibility center managers.

Types of Responsibility Centers

The agency theory view lends itself to a description of responsibility centers in terms of the authority of managers to acquire assets and the kinds of financial targets that would align responsibility with authority:

A Strategic Management Perspective on Responsibility Budgeting and Accounting

The practice has also been described in terms of organizational design and strategic management. In these terms, responsibility budgeting and accounting takes place within an organizational configuration known as an M-form, where decisional authority over strategy formulation is reserved to top management, while decisional authority over strategy implementation is decentralized to business units headed by general managers (Mintzberg 1983).

From the management strategy perspective, a responsibility budget is merely an artifact of the management process conducted within such a structural set up. Specifically, the responsibility budget formalizes a performance target for a given business unit over a specified time scale. In the typical case, goals are expressed in terms of economic quantities that reflect the utilization of resources and the financial results obtained as well as other scorecard measures. Responsibility accounting systems are set up to produce these measures so that divisional performance can be compared with targets in a timely manner and, where necessary, adjusted. Because business strategies are usually conceived along product-market lines (single product, differentiated products, multiple products) and because the M-form structures provide a general manager for each product line (rather than for regions or functions), in the management control and strategic management literatures, responsibility budgeting and accounting is broadly endorsed as the mode of organizing and managing large, multiproduct firms whose outputs are by definition heterogeneous.

Diagnostic versus Interactive Control

Under traditional responsibility budgeting and accounting systems, top management exercises control "by the numbers" from a small corporate headquarters, using financial targets, which it sets for the operating divisions. Robert Simons (1995: 102) refers to this kind of control as diagnostic control. Diagnostic control severely restricts the upward flow of operating information within organizations &emdash; making decentralization a necessity as well as an ideal.

This approach, however, is based on the assumption managers know how to improve performance. Hence, as Bob Behn observes," all that is required is to give them either (a) the correct incentives or (b) the necessary flexibility, and they will do it &emdash; they will just know what managerial actions will be most effective in improving performance." Unfortunately, managers don't necessarily know what to do. Consequently, he suggests that "maybe we also need a help-the-managers-manage strategy" (Behn 2003: 2).

An alternative to diagnostic control or control by the numbers is control by debate and dialogue, what Simons calls interactive control. Control by debate and dialogue is a help-the-manager-manage strategy. It is by design a learning process, proceeding from strategic vision through choices and their consequences to better understanding, clearer vision, improved choices, and higher valued consequences. Simons observes that both diagnostic control and interactive control are consistent with the practice of decentralization. However, decentralization is possible under diagnostic control only where top management attends to top management functions and refrains from meddling in the conduct of operations. This takes considerable self-restraint. Nevertheless, Simon argues that the best-managed, decentralized organizations are precisely those where less emphasis is given to meeting financial targets per se than to the effectiveness with which operating managers engage in reflective argumentative exchange and where target setting is a bottom-up process. To meet the burden of argument in this context, operating managers must persuade their superiors that they fully understand every aspect of their businesses &emdash; costs, trends, operating efficiency, marketing strategy, competitive position &emdash; and that their action plans and programs will realize the larger organizational purpose or interest.

The other Axis of Responsibility Center Design

Responsibility centers are usually classified according to a second axis or dimension:

On this axis, a responsibility center can be either a mission center or a support center. The output of a mission center contributes directly to an organization's objectives or purpose. The output of a support center is an input to another responsibility center in the organization, either another support center or a mission center.

Formerly, in most large complex organizations in the private sector, individual production units were typically standard cost centers; staff units were typically discretionary expense centers. Indeed, only mission centers were allowed to be investment centers. The reasons for this are complex, but they go to difficulties associated with expensing intermediate and joint products. Mission centers in private sector organizations produce final products that are easily priced and that are expensed following generally accepted accounting practice. In contrast, support centers produce intermediate products and these were, until recently, hard to cost, let alone price, with accuracy. Attempts to do so were often either excessively arbitrary or prohibitively costly.

Nowadays, however, advances in information technology, managerial accounting, and organizational design have made it possible and, in some cases, beneficial to treat every responsibility center in an organization as an investment center (see Thompson, 1998).

Paradoxically, public sector organizations are a mirror image of large complex organizations in the private sector. We know now how to treat support centers in most organization as quasi-profit or even investment centers (Lapsley, 1994; for additional public sector examples, see Anthony & Young, 1994: 371-374; Kaplan & Cooper, 1998: 245-251). But, because the final products of government's core mission centers are public goods that are passively enjoyed (Vining & Weimer, 1998), pricing final outputs remains for the time being and for the foreseeable future either excessively arbitrary or prohibitively costly. This means, for example, that, while it might make sense to treat services with direct commercial counterparts such as military depot maintenance, spare parts management, or facilities support centers as investment centers, it will continue to be necessary to treat the armed forces' combatant commands as discretionary expense centers. Fortunately, as far as exhaustive expenditures are concerned, about 75 percent of the activities performed by the US federal government fall into the support category and, for the most part, state and local governments are not in the business of supplying pure public goods (see Goldin, 1977).

Transfer Pricing

Under responsibility budgeting, support centers provide services or intermediate goods to other responsibility centers in return for a notational transfer price, organizations are structured to take advantage of specialized knowledge and local conditions, center managers make decisions and are held responsible for the overall financial performance of their centers. Sound transfer pricing is, therefore, the key to aligning the incentives of responsibility center managers with organizational interests.

Transfer pricing is also important to transparency within organizations. It helps to determine the costs of services provided by one unit to another, which is central to measuring performance relative to a financial target, and therefore plays a major role in establishing, as well as manipulating, the incentives facing responsibility center managers. Transfer pricing also reveals the internal costs of service decentralization where costs are incurred in transferring decision rights to others within an organization. When one sub-unit transfers tangible assets, knowledge, skills, etc., to another, both units calculate the cost as a means of revealing their liquid and tangible asset use internally and in external provision of service.

There are two common approaches to transfer pricing:

(A third method is based upon fully distributed average cost of the service or product.)

However, the circumstances that justify large complex organizations -- economies of scale and scope -- render these simple transfer-pricing mechanisms problematic. Scale economies are usually the result of large, lumpy investments in specialized resources -- technological knowledge, product specific research and development, or equipment. These investments tend to give rise to bilateral monopoly, a circumstance that provides an ideal environment for opportunistic behavior on the part of suppliers and customers. For example, once an intermediate product producer has acquired a specialized asset, customers may be able to extract discounts by threatening to switch suppliers. In that case, the supplier may find it necessary to write off a large part of the specialized investment. Or, if demand for the final good increases greatly, the intermediate product supplier may be able to extort exorbitant prices from customers. Hence, where the relationship between intermediate product supplier and customer is at arm's length, opportunistic behavior may eliminate the payoff to what would otherwise be cost effective investments. For example, the Report of the Commission on Roles and Missions of the Armed Forces (Commission, 1995; see also Thompson & Jones, 1994) suggested that budget authority should flow through the combatant commands to the military departments. Were that the case, lacking a long-term credible commitment on the part of the Joint Chiefs and the combatant commanders, the navy's investment in specialized assets like aircraft carriers would permit it to be exploited in peacetime. In wartime, of course, the tables would be turned.

The new economics of organizations tells us that vertical integration occurs because it can mitigate this problem, in part through the substitution of direct supervision for remote control (see Williamson, 1985). For example, in a study of military procurement, Scott Masten (1984) demonstrated that specialized investments are critical to vertical integration. Where intermediate products were both complex and highly specialized (used only by the buyer), there was a 92 percent probability that they would be produced internally; even 31 percent of all simple, specialized components were produced internally. The probability dropped to less than 2 percent if the component was unspecialized, regardless of its complexity.

Unfortunately, the problems that arise in arm's length transactions where there are few alternative suppliers/customers also arise where one attempts to replicate free market forces within the organization, allowing buying and selling responsibility centers complete freedom to negotiate prices (laissez-faire transfer pricing). Traditionally, economists have argued that services should be transferred at marginal or incremental cost to the buying responsibility center. But this can seriously distort the evaluation of support center performance and tend to eliminate incentives to improvement.

As a result, organizations face a serious dilemma. They can maximize short run performance by using marginal cost in internal transactions, thereby seriously distorting performance measurement and incentives, which will cause shortfalls in long-run performance. Or they can sacrifice short-term performance by relying on laissez-faire transfer pricing, thereby obtaining superior measures of the support center's contributions to organizational performance, and improve the chances of maximizing performance in the long term. Organizations can, promote short-run performance by using incremental cost pricing or they can promote long-term performance by using laissez-faire pricing, but they cannot do both simultaneously using either of these simple transfer pricing mechanisms.

In theory, bilateral monopoly can be governed quite satisfactorily by unbalanced transfer prices, multi-part transfer prices, or quasi-vertical integration. Under unbalanced transfer prices, the selling responsibility center is credited with the full cost of the transacted item (often standard cost), plus an agreed upon markup, the buying center is charged its marginal cost, and the organization’s accounts are adjusted to reflect the difference between the two. Unbalanced transfer prices are rarely used, however, where market prices are available. Under, multi-part transfer prices, the service delivered is decomposed to reflect underlying cost drivers and priced accordingly (your home phone bill is an excellent example of a multi-part tariff). Under quasi-vertical integration, the buyer invests in specialized resources and loans, leases, or rents them to their suppliers. Quasi-vertical integration is common in both the automobile and the aerospace industries, and, of course, it is standard procedure for the Department of Defense to provide and own the equipment, dies, and designs that defense firms use to supply it with weapons systems and the like (See Monteverde & Teece, 1982). Other organizations that rely on a small number of suppliers or a small number of distributors write contracts that constrain the opportunistic behavior of those with whom they deal.

In still other cases, desired outcomes can be realized through alliances based on the exchange of hostages (e.g., surety bonds, exchange of debt or equity positions) or just plain old-fashioned trust based on long-term mutual dependence. Toyota, for example, relies on a few suppliers that it nurtures and supports (Womack, Jones, Roos, 1990). They have substantial cross-holdings in each other and Toyota often acts as its suppliers' banker. Toyota maintains tight working links between its manufacturing and engineering departments and its suppliers, intimately involving them in all aspects of product design and manufacture. Indeed, it often lends them personnel to deal with production surges and its suppliers accept Toyota people into their personnel systems.

Toyota's suppliers are not completely independent companies with only a marketplace relationship to each other. In a very real sense, they all share a common purpose and destiny. Yet Toyota has not integrated its suppliers into a single, large bureaucracy. It wanted its suppliers to remain independent companies with completely separate books -- real profit/investment centers, rather than merely notational ones -- selling to others whenever possible. Toyota's solution to the bilateral monopoly problem appears to work just fine (Womack, Jones, Roos, 1990). In fact, with the exception of unbalanced transfer prices, none of the solutions to the bilateral monopoly problem noted here presumes vertical integration. All that is required is full access to cost and production information (Milgrom and Roberts, 1992). Of course, all of these solutions to the transfer pricing/organizational design are potentially available to government organizations. Indeed, many of them were pioneered by federal acquisitions personnel or imposed by public utility commissions. They are not, however, widely understood or appreciated by public administrators and financial managers.

Responsibility Budgeting in Government

The origins of responsibility budgeting and accounting in government can be traced to the Planning, Programming, and Budgeting System (PPBS) era in the US Department of Defense (1961-1967). Responsibility budgeting and accounting was the centerpiece of Project Prime, perhaps the most promising of the organizational design and development efforts initiated under Secretary of Defense Robert McNamara. Project Prime was the brainchild by Robert N. Anthony (Juola, 1993: 43-44), who succeeded Charles Hitch as defense controller in September 1965. Anthony saw the need for clarification of the purpose of each of the administrative units that comprised the Department of Defense, their boundaries, and their relationships to each other, and for an account structure that would tie the entire organization together. Anthony (1962) proposed that the Department of Defense:

The principal formal device by which a measure of intra-organizational decentralization was and is accomplished within the US Department of Defense is the revolving fund. These funds involve buyer-seller arrangements internal to the Department of Defense. They have actually been in use for some time. The navy had a revolving fund as early as 1878. Modern-day revolving funds date to the 1947 National Security Act, which authorized the defense secretary to use them to manage support activities within the Department of Defense. Two kinds of funds have been established under this authority: stock and industrial funds. Stock funds are used to purchase supplies in bulk from commercial sources and hold them in inventory until they are supplied to the customer -- usually a military unit or facility. Industrial funds are used to purchase industrial or commercial services (e.g., depot maintenance, transportation, etc.) from production units within the Department of Defense. Both kinds of funds are supposed to be financed by reimbursements from customers' appropriations (Juola, 1993: 43).

Anthony's proposal would have expanded the scope of this device and enhanced its effectiveness by establishing rules for setting transfer prices prospectively rather than retrospectively and by making support center managers responsible for meeting explicit financial targets. Internal buyer-seller arrangements encourage efficient choice on the part of support centers, as well as the units that use their services, only if prices are set ahead of time and support centers charge all of their costs against revenues earned delivering services. Furthermore, their managers must be fully authorized to incur expenses to deliver services, and held responsible for meeting the stated financial goals of their centers (Bailey, 1967: 343).

Project Prime failed. One reason for its failure is that the federal government of the US accounts for purchases, outlays, and obligations, but it still does not account for consumption. Full value from the application of responsibility budgeting can be obtained only where government adopts a meaningful form of consumption or accrual accounting (measuring the cost of the assets actually consumed producing goods or services). Because the US government does not account for resource consumption, its cost figures are necessarily statistical in nature (i.e., they are not tied to its basic debit and credit bookkeeping/accounting records). Without the discipline that debit and credit provides, these figures are likely to be satisfactory only for illustrative purposes or where a decision maker must make a specific decision and a cost model has been tailored to the decision maker's needs. Another reason for the failure of Project Prime is that US appropriations process does not perform the capital budgeting function satisfactorily, a problem that PPBS did not really address and certainly didn't fix. Besides which, the existing process procrusteanizes every operating cycle to fit the fiscal year.

Responsibility budgeting next surfaced in the United Kingdom, as part of the Thatcher government's Financial Management Initiative, announced May 17, 1982 (Pollitt, 1993; Lapsley, 1994). The Financial Management Initiative called for a radical change in the internal structure and operations of government agencies. Objectives were to be assigned to responsibility centers. Costs were to be systematically identified. They were to be measured on an accrual basis (i.e., matching resources consumed to services delivered) and include not only the direct costs of service delivery but overheads as well. This identification enabled those responsible for meeting particular objectives to be held accountable for the cost of the resources they were consuming

The scope of responsibility accounting and budgeting in the UK was further extended in 1988 by the Thatcher government's Next Steps Initiative. In the last eight years, much of the British civil service has been reorganized into a set of executive agencies that have been given considerable administrative and fiscal flexibility and expected to meet annual financial performance targets. The heads of these executive agencies are no longer career civil servants. They are recruited from either the private sector (about 25 percent) or public sector, hired on short term contracts, with pay and tenure contingent on their success in meeting annual performance targets. By April 1996, there were 125 executive agencies in the UK, with 37 more candidates under consideration, covering about 75 percent of the British civil service (Roberts, 1997).

Following the launch of the Financial Management Initiative in Great Britain, other governments -- Australia, Canada, Denmark, Finland, and Sweden -- have adopted responsibility budgeting and accounting. None, however, has moved as far or as fast as New Zealand. Moreover, New Zealand's reformers explicitly recognized their debt to agency theory (Boston et al., 1996; Lewis et al., 1996).

New Zealand

Most of the external attention given to New Zealand's public management reforms has focused on its efforts to improve transparency: the adoption of accrual accounting and reporting on performance. New Zealand was the first country to publish a full set of standard financial statements, including a balance sheet of assets and liabilities and an accrual based operating statement of income and expenses. However, the changes made in the structure of the government of New Zealand designed to promote effective resource use and investment are even more significant than are the changes in financial reporting. First of all, New Zealand's Parliament privatized everything that was not part of the core public sector. The residual core public sector now includes a mix of policy and regulatory and operational functions and the military services, policing and justice services, social services such as health, education, and the administration of benefit payments, research and development, property assessment, and some other financial services.

Second, Parliament redefined the relationship between it and the heads of government agencies. Agency heads lost their permanent tenure and are now known generically as 'chief executives.' They are appointed for fixed terms of up to five years, with the possibility of re-appointment. Each works to a specific contract, the conditions of which are negotiated with the State Services Commission and approved by the Prime Minister. The State Services Commission also monitors and assesses executive performance. Remuneration levels are directly tied to performance assessment.

Third, Parliament changed the way it appropriates funds for use by the remaining government agencies to link appropriations to performance, allowing Parliament fiscal control, but, at the same time, providing greater fiscal flexibility for agency heads. The basis of appropriation depends on the agency's ability to supply adequate information about its performance. Three modes of appropriation are possible, recognizing that some agencies provide goods and services that are more commercial or contestable than do others.

All agencies started out in Mode A, but most have progressed either to Mode B or C. Under Mode A, agencies were discretionary expense centers and Parliament appropriated funds for the purchase of resources. Indeed, the only change from the budget process in effect before 1989 (or, for that matter, the budgets used by most governments throughout the world) is that separate appropriations were provided for expenditures for plant and equipment. This mode remained in force until the agency developed a satisfactory accrual accounting system and identified its outputs, both of which are needed for performance assessment.

Under Mode B, most agencies are quasi-profit centers. This mode is designed for agencies that supply traditional, noncontestable, governmental services: the central control agencies, including the State Services Commission, most regulatory and police functions, and some justice services, i.e., policy agencies and activities that include an element of compulsion for the buyer. Under this mode, Parliament appropriates funds retrospectively to reimburse agencies for expenses incurred in producing outputs during the period covered by the contract, whether for the government or third parties. Costs are measured on an accrual basis; they include depreciation, but exclude taxes and the return on funds employed. Changes in an agency's net asset holdings are also explicitly appropriated.

Under Mode C, agencies are investment centers. Appropriations pay for the outputs produced by the agency and for any changes in the agency's net assets. Agencies in Mode C are required to pay interest, taxes, and dividends and must establish a capital structure. Mode C agencies are set up in a competitively neutral manner so that their EVAs can be assessed by comparison with firms in the private sector. The prices paid for the outputs supplied by Mode C agencies are supposed to approximate fair market prices. In general, this means that agencies must show that they are receiving no more than the next best alternative supplier would receive for providing the outputs. Mode C agencies are not permitted to borrow on their own behalf or to invest outside their own areas of operation. Each month, each agency reports on its financial position and cash flow and resource usage and revenue by output. Variances are calculated and explanations provided. Under Mode B, managers are free to make some decisions about investments in plant and equipment; they may make even more decisions under C. The fact that their financial performance is one of the main bases upon which managerial performance is assessed helps insure that those decisions will be sound (see Schick, 1996; for criticisms of the New Zealand reforms, see Mascarenhas, 1996 and Quiggen, 1998).

Government's key decisions remain firmly in the hands of Parliament. The decisions that have the most significant future consequences for the government of New Zealand's stakeholders are clearly those which have to do with the kind, quantity, and quality of service provided by the citizenry. Under the existing system of appropriations and financial reporting, those issues must be explicitly confronted when cabinet enters into long term contracts with agencies, state owned enterprises, and firms to deliver service outputs and its consequent liabilities must be stated in present value terms.

The United States

Responsibility budgeting and accounting was adumbrated in the United States and influenced the now defunct Defense Management Report Initiatives of the Bush/Cheney era in the Department of Defense, and arguably the content of both the Chief Financial Officers Act and the National Performance Review's calls for mission-driven, results-oriented budgets and, more recently, performance based organizations (OECD, 1995: 230). Still, it has had little or no practical effect in this country.

There are two explanations for this fact. The first is that many students of the expenditure process reject the notion that remote control can be reconciled with the American legislative budgetary process. Some people even assert that it can be practiced only by responsible unitary governments on the Westminster model, although that claim is belied by the Swiss and Swedish examples (Schedler, 1995; Arwidi and Samuelson, 1993) and various state (Barzelay, 1994) and local governments here in the US (Kaplan & Cooper, 1998: 245-251). Of course, it would not be easy to reconcile responsibility budgeting with the American legislative process, but we do not believe that they are necessarily incompatible (see Thompson, 1994; Harr, 1989; Harr and Godfrey, 1991 & 1992). A second possible explanation for its failure to leave its mark on government accounting and budget practices in the United States is that, unlike most other countries, America has large, well-organized corps of government accountants, auditors, budgeters, program analysts, and teachers of government accounting and budgeting. All of these groups have vested interests in differentiating public from private practice, because that difference gives value to their expertise. A third reason seems to be that many people, in and outside of government, evidently believe that "public" necessarily implies Prussian-style bureaucracy. Where the purpose of the organization in question or the technology available to it make Prussian-style bureaucracy inappropriate, they will hear of no alternative short of full-scale privatization.

The Clinton Administration's second-term government reform efforts have centered on pushing the concept of performance-based organizations (PBOs) modeled after Britain's Financial Management Initiative (Green, Jones, & Thompson, 1999; see, however, Roberts, 1997). The main theme of this reform effort is the use of contracts to hold PBOs accountable for financial performance.

The progress of this effort has been glacial. When legislation for the first PBO candidate, the Patent and Trademark Office, was sent to Congress, it aroused an intense debate between the administration and the chair of the House Judiciary subcommittee on courts and intellectual property regarding the relative merits of the PBO model versus a corporate model. This debate has been reproduced in various venues for successive PBO candidates.

What Goes Around, Comes Around

It is somewhat ironical that governments are beginning to embrace remote control at the same time many well-managed businesses are abandoning it (Bruggeman, 1995; Otley, 1994; Bunce, Fraser, Woodcock, 1995). These businesses have abandoned remote control because they are no longer compartmentalized the way they once were and it simply doesn't reflect the way they are now put together (Bruggeman, 1995; Otley, 1994; Bunce, Fraser, Woodcock, 1995). Arguably, decompartmentalization is being driven by the information revolution, which is breaking down economies of scale and scope built upon functional specialization (Reschenthaler and Thompson, 1996). According to Michael Hammer, modern data bases, expert systems, and telecommunications networks provide many, if not all, of the benefits that once made internal specialization of administrative functions like personnel, finance, accounting, etc. attractive (Hammer, 1990: 108-112). To the extent that the provision of these services requires specialized skills, they are increasingly contracted out to specialist firms. The people in the organization who actually do its real work perform the rest.

Hammer claims that jobs should be designed around an objective or outcome instead of a single function; that functional specialization and sequential execution are inherently inimical to expeditious processing; that those who use the output of activity should perform the activity and the people who produce information should process it, since they have the greatest need for information and the greatest interest in its accuracy; that information should be captured once and at the source; that parallel activities should be coordinated during their performance, not after they are completed; and last, that the people who do the work should be responsible for decision making and control built into job designs (Hammer, 1990).

Decompartmentalization has led to smaller, flatter organizations, organized around a set of generic value-creating processes and specific competencies. Some single-mission organizations are now organized as virtual networks, some multi-mission organizations as alliances of networks. Philip Evans and Thomas Wurster refer to both of these kinds of organizational arrangements as hyperarchies, after the hyperlinks of the World Wide Web (Evans and Wurster, 1997: 75). Evans and Wurster assert that these kinds of organizations, like the Internet itself, the architectures of object-oriented software programming, and packet switching in telecommunications, have eliminated the need to channel information, thereby eliminating the tradeoff between information bandwidth (richness) and connectivity (reach). Evans and Wurster describe virtual networks (structures designed around fluid, team-based collaboration within the organization) as deconstructed value chains, and alliances of networks (the pattern of "amorphous and permeable corporate boundaries characteristic of companies in the Silicon Valley") as deconstructed supply chains, in which "everyone communicates richly with everyone else on the basis of shared standards."

The system used by IBM at its plant in Dallas, Texas, is an example of an existing virtual network. It has been designed to mimic a market-like, self-organizing system. Everyone in the organization plays the part of customer or provider, depending on the transaction, and the entire plant has been transformed into a network of dyads and exchanges. Each exchange is a closed loop involving four distinct steps: request from a customer and offer from a provider, negotiation of the task to be performed and the definition of success, performance, and customer acceptance. Until this last step is completed, the task remains unfinished. Each closed loop of workflow is further broken down into subloops. Under this system, even simple tasks give rise to dozens of loops and interconnecting lines; more complex tasks, such as modifying a major product, to hundreds; and managing the entire Austin plant to thousands. IBM uses powerful computers to keep track of all of these loops and lines, to chart all activities and operational flows within the plant, to keep track of progress being made at each stage of each transaction, and to prod tardy participants into action -- this is control built into job design with a vengeance.

The effect of this system has been to break down departmental boundaries, eliminate bottlenecks, and to empower employees to take initiatives and coordinate themselves. As a by-product, the computer systems that keep track of all these loops and lines also identify the resources going into a particular job, almost entirely eliminating the need for cost allocation. Moreover, this information is available both prospectively and retrospectively to anyone in the organization.

Some well-managed multi-mission organizations such as Johnson & Johnson, 3M, and Rubbermaid have already organized themselves into loose alliances of networks, sharing only their top management, a set of core competencies, and a common culture (Quinn, 1992). The control systems used by these organizations are like those of centralized bureaucracies in that they collect a lot of real-time information on every aspect of operations, including nonfinancial information (see Table 1), but unlike the control systems of stovepiped centralized bureaucracies, which were erected on the premise that the exercise of judgment should be passed up the managerial ranks, this information is used to push the exercise of judgment down into the organization, to wherever it is needed, at the point of sale, at delivery, or in production (Simons, 1995). From top management's perspective, the primary purpose of this information is to provide them with insight into the integrity, competence, and morale of their network managers and employees so that they can allocate their best people to the most important jobs.

















How far hyperarchy will go is an open question. Evans and Wuster (1997) claim that it will destroy all hierarchies, whether of logic or of power, "with the possibility (or the threat) of random access and information symmetry." If hyperarchy is where we are all heading, responsibility budgeting and accounting is at best an intermediate stage (Otley, Broadbent, & Berry, 1995). It is now apparent, as it really was not before, that responsibility budgeting restricts the upward flow of operating information within organizations -- making decentralization a necessity as well as an ideal. In contrast, networks and alliances are information rich environments. For the most part, access to information is symmetrical in fully networked organizations -- equally available to all the people in the organization.

Why not skip the intermediate stages and go directly to networked organizations? One of the referees for this article suggested that some governments might be moving in that direction. He notes the trend toward performance efforts and accomplishments reporting and the wide-spread acquisition of so-called enterprise resource planning (ERP) systems built around common data structures and centralized information warehouses, which permit data to be entered and accessed from anywhere in the organization (SAP, PeopleSoft, Oracle, etc.) Robert Kaplan and Robin Cooper, 1998: 25) argue, however, that organizations that try to move directly to a system where everyone communicates richly with everyone else on the basis of shared standards, without passing through a recommended period of experimentation with operational-feedback and cost-measurement systems, will almost surely fail.

We are less certain that this is the case. Nevertheless, we would point out that decentralization can work in an information-rich environment only where top management attends to top management functions -- strategic planning, organizing, staffing, the intellectual and cultural development of the organization -- and refrains from meddling in the conduct of operations. This takes self-restraint, and self-restraint must be learned. For that reason, it may make sense for governments to experiment with responsibility budgeting rather than going directly to new modes of organization and control. Few have had much experience with decentralization and almost none with self-restraint (see Johansen, Jones, and Thompson, 1997).


Anthony, R.N., "New Frontiers in Defense Financial Management," The Federal Accountant, XI, June 1962: 13-32.

Anthony, Robert N., and David W. Young, Management Control in Nonprofit Organizations, fifth ed., Homewood, IL: Richard D. Irwin, 1994.

Arwidi, Olof, and Lars A. Samuelson, The Development of Budgetary Control in Sweden: A Research Note, Management Accounting Research, 4/2, June 1993: 93-107.

Bailey, Martin J., "Defense Decentralization through Internal Prices," in Defense Management, S. Enke (ed.), Englewood Cliffs, NJ: Prentice-Hall, 1967.

Barzelay, M., with B.J. Armajani, Breaking through Bureaucracy: A New Vision for Managing in Government, Berkeley: University of California Press, 1992.

Boston, Jonathan, John Martin, June Pallot, and Pat Walsh, Public Management: The New Zealand Model, London: Oxford University Press, 1996.

Breton, A. and R. Wintrobe, "The Equilibrium Size of a Budget-Maximizing Bureau." Journal of Political Economy, 83, 1975: 195-207.

Bruggeman, Werner, "The Impact of Technological Change on Management Accounting," Management Accounting Research, 6/3, Sept. 1995: 241-252.

Bunce, Peter, Robin Fraser, and Lionel Woodcock, "Advanced Budgeting: A Journey to Advanced Management Systems," Management Accounting Research, 6/3, Sept. 1995: 253-265.

Chandler, Alfred, Strategy and Structure: Chapters in the History of Industrial Enterprise, Cambridge, MA: MIT Press, 1962.

Commission on Roles and Missions of the Armed Forces, Directions for Defense, Washington, DC: US Government Printing Office, 1995.

Evans, Lewis, Arthur Grimes, and Bryce Wilkinson, with David Teece, "Economic Reform in New Zealand 1984-95: The Pursuit of Efficiency," Journal of Economic Literature, 34, December 1996: 1856-1902.

Evans, P.B., and T.S. Wurster "Strategy and the New Economics of Information," Harvard Business Review, September-October, 1997: 71-82.

Green, Mark, L.R. Jones, and Fred Thompson, "Local Heroes? Reinvention Labs in the Department of Defense," in Brudney, O'Toole, & Rainey Advancing Public Management, Washington DC: Georgetown University Press, 2000: 153-172.

Goldin, K.D., "Equal Access vs. Selective Access: A Critique of Public Goods Theory," Public Choice, 29, spring 1977: 53-71.

Hammer, M., "Reengineering Work: Don't Automate, Obliterate," Harvard Business Review, July-August, 1990: 104-112.

Harr, David J., "How Activity Accounting Works in Government," Management Accounting, 72, September 1990: 36-40.

Harr, David J., "Productive Unit Resourcing: A Business Perspective on Government Financial Management," Government Accountants Journal, Summer 1989: 51-57

Harr, David J., and James T. Godfrey, "The Total Unit Cost Approach to Government Financial Management," Government Accountants Journal, Winter 1992: 15-24.

Harr, David J., and James T. Godfrey, Private Sector Financial Performance Measures and Their Applicability to Government Operations, Montvale, NJ: National Association of Accountants, 1991.

Johansen, Carole, L.R. Jones, and Fred Thompson, "Management and Control of Budget Execution," Chapter 29 in Public Budgeting and Finance, fourth ed., R. Golembiewski & J. Rabin (eds.), New York: Marcel Dekker, 1997: 577-584.

Jones, L.R., and Jerry McCaffrey, "Federal Financial Management Reform," Handbook of Public Budgeting, R. Meyers (ed.) San Francisco: Jossey-Bass, 1998.

Juola, Paul, "Unit Cost Resourcing: A Conceptual Framework for Financial Management," Navy Comptroller, 3/3, April 1993: 42-48.

Kaplan, Robert S., and Robin Cooper, Cost & Effect: Using Integrated Cost Systems to Drive Profitability and Performance, Boston: Harvard Business School Press, 1998.

Lapsley, Irvine, "Responsibility Accounting Revived? Market Reforms and Budgetary Control," Management Accounting Research, 5/3,4, Sep.-Dec. 1994: 337-352.

Mascarenhas, R.C., "Searching for Efficiency in the Public Sector: Interim Evaluation of Performance Budgeting in New Zealand," Public Budgeting & Finance, 16/3, Fall 1996: 13-27.

Masten, Scott E., "The Organization of Production," The Journal of Law and Economics, 27/4, 1984: 403-417.

Masten, Scott E., J.W. Meehan, and E.A. Snyder, "The Costs of Organization," The Journal of Law, Economics, and Organization, 7/1, 1991: 1-25.

Milgrom, Paul, and John Roberts, Economics, Organization, and Management, Englewood Cliffs, NJ: Prentice Hall, 1992.

Monteverde, Kirk, and David J. Teece, "Appropriable Rents and Quasi-Vertical Integration," The Journal of Law and Economics, 25/3, 1982: 403-418.

[New Zealand] Treasury, Government Management: Brief to the Incoming Government, Vol. I. Wellington, NZ: The Treasury, 1987.

OECD, Budgeting for Results: Perspectives on Public Expenditure Management, Paris: Organisation for Economic Co-operation and Development, 1995.

Otley, David, "Management Control in Contemporary Organizations: Towards a Wider Framework," Management Accounting Research, 5/3,4, Sep.-Dec. 1994: 289-299.

Otley, David, Jane Broadbent, and Anthony Berry, "Research in Management Control: An Overview of its Development," British Journal of Management, 6 (Special Issue), Dec. 1995: 31-44.

Pitsvada, B.T., "Flexibility in Federal Budget Execution," Public Budgeting and Finance, 3, Spring 1983: 17-26.

Pollitt, Christopher, Managerialism and the Public Services: Cuts or Cultural Change in the 1990s? second ed., Cambridge MA: Basil Blackwell, 1993.

Quiggin, John, "Social Democracy and Market Reform in Australia and New Zealand," Oxford Review of Economic Policy, 14/1, spring 1998: 76-95.

Quinn, James Brian, Intelligent Enterprise: A Knowledge and Service Based Paradigm for Industry, New York: Free Press, 1992.

Reschenthaler, G. B., and Thompson, Fred, "The Information Revolution and the New Public Management," Journal of Public Administration Research and Theory, 6/1, 1996: 125-144.

Roberts, Alasdair, "Performance Based Organizations: Assessing the Gore Plan," Public Administration Review. 57/6, Nov.-Dec. 1997: 465-479.

Rodriquez, Justine Farr, "Budgetary Accounting and Accrual Accounting: Reconciling the Differences," presentation, Association for Budgeting and Financial Management Meeting, ANA Hotel, Washington DC, October 29, 1996.

Rosenberg, Nathan, and L.E. Birdsall, How the West Grew Rich: The Economic Transformation of the Industrial World, New York: Basic Books, 1986.

Schedler, K., Ansatze einer Wirkungsorientierten Verwaltungsfuhrung: Von der Idee des New Public Managements (NPM) zum konkreten Gestaltungsmodell, Bern: Paul Haupt, 1995.

Schick, Allen, The Spirit of Reform, [New Zealand] State Services Commission, 1996.

Scott, G., P. Bushnell, and N. Sallee, "Reform of the Core Public Sector: The New Zealand Experience," Public Sector, 13/3, 1990: 11-24.

Simons, Robert, Levers of Control: How Managers Use Innovative Control Systems to Drive Strategic Renewal, Boston: Harvard Business School Press, 1995.

Thompson, Fred, "Cost Measurement and Analysis," Handbook of Public Budgeting, R. Meyers (ed.) San Francisco: Jossey-Bass, 1998.

Thompson, Fred, "Capital Budgeting," International Encyclopedia of Public Policy and Administration, Boulder CO: Westview Press, 1997: 330-337.

Thompson, Fred, "Business Strategy and The Boyd Cycle," Journal of Contingencies and Crisis Management, 3/2, June 1995: 81-90.

Thompson, Fred, "Mission-Driven, Results-Oriented Budgeting: Financial Administration and the New Public Management," Public Budgeting & Finance, 14/3, Fall 1994: 90-105.

Thompson, Fred, "Matching Responsibilities with Tactics: Administrative Controls and Modern Government," Public Administration Review, 53/4, 1993: 303-318.

Thompson, Fred, and L.R. Jones, Reinventing the Pentagon: How the New Public Management Can Promote Institutional Renewal, San Francisco: Jossey-Bass Publishers, 1994.

Thompson, Fred, and L.R. Jones, "Controllership in the Public Sector," Journal of Policy Analysis and Management, 5/3, 1986: 547-571.

Vining, Aidan, and D.L. Weimer, "Passive Use Benefits: Existence, Option, and Quasi-Option Value," Handbook of Public Finance, F. Thompson and M. Green (eds.) New York: Dekker, 1998: 319-346.

Williamson, O.E., The Economic Institutions of Capitalism, New York: The Free Press, 1985.

Womack, James P., Daniel T. Jones, and Daniel Roos, The Machine that Changed the World: Based on the Massachusetts Institute of Technology 5-Million Dollar 5-Year Study on the Future of the Automobile, New York: Rawson Associates, 1990.

World Bank, Bureaucrats in Business: The Economics and Politics of Government Ownership, New York: Oxford University Press, 1995.

Zimmerman, Jerold L., Accounting for Decision Making and Control, Chicago: Irwin, 1995.