Book
Share
Chapter

7 Making Performance Budgeting Work

Author(s):
Marc Robinson
Published Date:
October 2007
Share
  • ShareShare
Show Summary Details
Author(s)
Marc Robinson

Performance budgeting practice is not monolithic. There are different forms of performance budgeting which seek to link results and funding in different ways. Performance budgeting systems also vary in respect to their objectives—in particular, some place more emphasis upon allocative efficiency than upon technical efficiency, or vice versa. They also have, as discussed in Part One, differing information requirements. In Part Two, we turn our attention from information requirements to the type of link between results and funding which performance budgeting systems strive to build. The primary objective here is to identify and assess the mechanisms by which each form of performance budgeting aims to achieve its intended impact upon allocative and/or technical efficiency, with the aim of deepening our understanding of what approaches to performance budgeting work and in what context.

In surveying the key issues for Part Two, it is convenient to start by considering those forms of performance budgeting which aim primarily to improve allocative efficiency, particularly through better expenditure prioritization in the central budget. This is followed by a discussion of the key issues which arise in respect to newer forms of performance budgeting which seek to develop rather tighter links between funding and results, often with the primary objective of inducing better agency-level performance.

Ensuring performance information is used for expenditure prioritization

As discussed in Chapter 1, expenditure classification is a basic tool of all forms of performance budgeting which aim to improve expenditure prioritization. Good expenditure prioritization requires the classification of public expenditure in the budget process into categories which facilitate allocative decision-making. The most fundamental and enduring instrument for this purpose has been the program classification of expenditure, the methodology of which is discussed in Chapter 5. Some forms of performance budgeting, with less enduring influence, have sought to take the principle of expenditure classification to facilitate allocative decision-making purposes even further. Zero-base budgeting, for example, breaks down expenditure into a larger number of “decision packages”—usually within programs—in order to systematically assess options for increasing or reducing expenditure directed to particular objectives. The intention is, in either case, the same—that information about the costs and benefits of programs or decision packages will be used by decision-makers, together with information about alternative expenditure options, to make better decisions about where limited public funds can be used to best effect. To do this, information about program costs must, of course, be accompanied by information on program results (see Chapter 4).

Producing the information to permit better expenditure prioritization does not, however, ensure that this information will be used. There are many cases even in recent years of countries where considerable effort has been made to introduce program classification, and to develop measures of program performance, without any significant use having apparently been made of that information by budget decision-makers. It is for this reason that, as Schick (2002, p. 23) notes, “there is consensus these days that changing budget classifications does not itself change budget allocations.” Even within OECD countries, as Teresa Curristine reports in Chapter 8, a large increase in the availability of performance information has not been matched by a corresponding increase in its use in the budget process.

There are, of course, a number of possible explanations for such a state of affairs. The problem may lie in part with the nature or adequacy of the performance information itself. It may reflect inflexibilities built into public expenditure—such as revenue earmarking and rigid civil service employment practices which make it hard to reallocate public employment from low priority to higher-priority purposes. It may be, more fundamentally, that political systems are so dysfunctional that there is negligible interest in efficiency and effectiveness as criteria for expenditure prioritization decisions. But even if none of these types of problems exist, and the ground is ripe for information on the costs and benefits of programs to have a substantial impact on the expenditure allocation, this will not happen unless the budget process facilitates the use of this information. Expressed differently, to the extent that “traditional” budget processes tend to have an inbuilt bias towards allocative inertia and so-called “incrementalism,”1 performance budgeting is unlikely to succeed in improving expenditure prioritization unless the budget process is reformed by the inclusion of mechanisms and procedures which aim to facilitate allocative flexibility and which, more specifically, facilitate the effective use of performance information in allocative decisions.

What form should such expenditure prioritization mechanisms and procedures take if they are to work well? It cannot be expected that there should be a single answer to this question. Different mechanisms and procedures may be appropriate for different political and administrative systems and cultures. Nevertheless, international experience can shed significant light on what can work and what is unlikely to work. It is for this reason that a number of the chapters in this part of the volume—both case studies and thematic pieces—focus upon expenditure prioritization processes and their relation to performance budgeting. By way of background to these studies, a number of overarching points may be identified. Many of the issues are considered in greater depth by Jim Brumby in Chapter 9 on performance budgeting and allocative efficency.

The first point is that expenditure prioritization needs to be conducted as a systematic routine integrated with the budget process. Thus, for example, the British spending review process (see Chapter 12 by Peter Smith) involves a routine of biennial reviews which lead directly to the setting of three-year ministry spending envelopes. Another successful prioritization mechanism has been the Australian Expenditure Review Committee (ERC), a cabinet committee supported by the finance ministry which was established at the national government level in the early 1980s with the express task of reviewing both existing programs and new spending proposals in such a way as to feed into annual budget decisions (Xavier, 1998; Campos and Pradhan, 1999). The case study in this part of the volume by John Pierce and Michael Di Francesco (Chapter 14) outlines how a version of this system works in Australia’s largest and most economically important State, New South Wales.

As both of these examples illustrate, expenditure prioritization tends to work best when the budget is put on a multi-annual basis, through either multi-year allocations or a system of rolling forward expenditure estimates. Expenditure prioritization is about choices between alternative policies, and the fiscal impact of specific policy choices can often not be measured by their cost in the prospective budget year. This point has long been recognized by performance budgeting advocates, well before the recent resurgence of interest in medium-term budgeting. The original US program budgeting system (PPBS), for example, aimed to plan expenditure over at least a five-year time horizon (Anshen, 1967, p. 356).

Expenditure prioritization-setting is most likely to be successful in terms of changing the actual allocation of resources if it is managed by the Ministry of Finance. Past experience with ad hoc reviews carried out by individuals appointed from outside government2 was generally disappointing. So also has been experience with the model, common in many Latin American and African countries, where separate planning/development ministries develop medium/long-term development plans which set strategic priorities which are supposed to guide annual budget formulation. Because such planning is conducted away from the budget process and by people who are not budget decision-makers, it all too easily becomes a paper exercise with little real impact on the allocation of resources. This problem was also experienced in the implementation of Medium-Term Expenditure Frameworks (MTEFs) in countries where the sectoral expenditure prioritization component of the MTEF process was carried out in isolation from the budget process, either by consultants or by ministries other than the MoF (Le Houerou and Taliercio, 2002).

As the British and Australian examples make clear, the expenditure prioritization process must include systematic routines for the review of existing programs—that is, of ongoing “base” public expenditure. All too often, it is only new spending initiatives which receive any analytic attention in the budget preparation process. Many countries have established processes for the appraisal of new expenditure proposals, while requiring nothing from line agencies by way of justification of expenditure on established programs. Even in countries where there is an explicit strategic phase in the budget process—for example, a cabinet meeting early in the budget process to set overall spending priorities (see Chapter 9) or a development planning process of the type described above—“priority-setting” is all too often more about identifying new priorities for expenditure than about identifying low-priority areas which can be cut to make room for these new priorities. The absence of systematic processes for the review of base expenditure virtually guarantees that hopes about the impact of performance budgeting will be disappointed, since it is precisely in the review of existing programs that performance information can be expected to make its greatest contribution.

It is crucial that expenditure review should not be confined to so-called “discretionary” expenditure, but should encompass the full gamut of public expenditure. All too often, expenditure on items such as civil service salaries and “mandatory” items such as social security benefits is treated as untouchable and exempt from review. Where this happens, expectations of what performance budgeting can deliver are bound to be disappointed, as the Spanish case study (Chapter 17) indicates. Successful expenditure prioritization processes set up in the UK, Australia, and elsewhere make no distinction between discretionary and non-discretionary expenditure, and do not exclude standing appropriations from the expenditure review spotlight.3 They recognize that all expenditure is flexible in the medium to long term, and that reducing the inflexibilities which inhibit the reallocation of funds is an important prerequisite of successful performance budgeting.

Important lessons about how to organize expenditure prioritization were learned from experience with the earliest form of program budgeting (especially US PPBS—see Chapter 1) and, subsequently, zero-base budgeting (ZBB). These systems strove to establish a central expenditure prioritization mechanism which was comprehensive—in the sense of encompassing the whole of the general government sector each year—and quite detailed. They were highly centralist, assuming “control in detail from the centre, as opposed to delegated authority, incentive structures and local initiative” (Spackman, 2002, p. 8), and viewed expenditure prioritization as essentially an exercise in central planning.4

Experience with these systems demonstrated that such comprehensive and detailed central expenditure prioritization mechanisms are too much for the budget process to handle. “Bounded rationality”—practical limits to how much information can be absorbed and how much complex rational calculation can be performed (see Chapter 2)—limits the scope for central planning of expenditure allocation in government just as it limits the scope for central planning of the economy as a whole. As a recent review of the history of performance budgeting by the Florida government put it:

One important lesson to be learned from past experience is that ambitions of budget reforms have often outstripped the analytic and information management capacity of government agencies. The labor-intensive nature of these systems led to a recognition that they were not feasible. (Office of Program Policy Analysis and Government Accountability, 1997)

The failure of detailed comprehensive planning was particularly striking in the case of ZBB. The attempt to review every program from bottom up every fiscal year—that is, to systematically consider all options ranging from cutting out programs to expanding them—proved to be impossibly ambitious. Even a scaled-down version of ZBB known as “alternative budgeting” proved so cumbersome that it was abandoned quickly in all jurisdictions which tried it out.5

Central expenditure prioritization mechanisms therefore need to be designed in such a way that they facilitate the use of relevant information for good allocative decision-making while recognizing the limits on the capacity of central decision-makers to analyze and plan. A number of alternative approaches have been taken to this problem. One is to conduct periodic rather than annual reviews of priorities. The British Spending Review process is a good example—not only are the Spending Reviews biennial rather than annual, but the more in-depth “Comprehensive” Spending Reviews are conducted only infrequently (see Chapter 12). Another alternative is to aim for comprehensive review of expenditure priorities on a staggered timetable over a number of years. Thus the US Program Assessment Rating Tool (PART), discussed by Denise Fantone in Chapter 10, has been applied in a cycle in which 20 percent of government programs are reviewed, so that all programs are reviewed over a five-year cycle. Some other countries (for example, Chile, Australia) have taken a less deliberately structured approach, reviewing selected areas of public expenditure each year on the basis, not of a predetermined review plan, but of top-level judgments about which programs are most ripe for review.

There is widespread contemporary recognition that, in part because of the limits of central planning, central budget decision-makers should in general focus their attention on broad expenditure allocations—particularly at the sectoral level—and that it is both necessary and desirable that there should be substantial decentralized allocative decision-making at the line agency level (World Bank, 1998). This raises many important questions about the interface between central and decentralized expenditure allocation processes. A particularly important issue here is that of how the central budget process may be designed to create incentives for, and pressures on, line agencies to keep their own expenditure priorities under active review, and to make maximum use of performance information for this purpose. The more expenditure prioritization is carried out at the decentralized level, the more well-designed performance budgeting systems need to concern themselves with how to improve resource allocation within line agencies as well as in the central budget process. The starting point of encouraging agencies themselves to take good expenditure prioritization seriously is, however, to create good, performance-informed expenditure prioritization at the center. If line agencies know that central decision-makers will look hard at the performance of their programs when determining their budget funding, they will be motivated not only to prioritize their own budgets much more effectively, but also to improve program design and management. This, as Fantone makes clear, has been one of the aims of the US Office of Management and Budget in its implementation of the PART system.

The effectiveness of expenditure prioritization processes depends in part on the overall quality of the budget preparation process. It is particularly important that the initial macro-fiscal steps of the budget process be well-managed. Aggregate fiscal discipline and good expenditure prioritization have a mutually reinforcing effect. It is only in the context of a hard budget constraint that priority-setting can be effective—otherwise the temptation is to respond to new expenditure priorities by continually increasing aggregate expenditure, rather than reallocating expenditure from lower-priority or less effective programs. As Brumby points out in Chapter 9, there are strong grounds to believe that the converse is also true: namely, that good expenditure prioritization facilitates the enforcement of fiscal discipline. Better expenditure prioritization processes might be expected to deliver the capacity to achieve better-quality fiscal consolidation when and if necessary—because, if aggregate expenditure needs to be cut, an improved capacity to target cuts at lower-priority areas should in principle both make fiscal consolidation less difficult than it might otherwise be, and also make it more enduring. Moreover, a capacity to allocate expenditure more effectively should contribute to restraint in the longer-run rate of growth of public expenditure.

It is therefore not surprising that fiscal stress has not infrequently provided the impetus for the introduction of performance budgeting systems aimed at improving expenditure prioritization. A good recent example of this is Brazil, where severe fiscal crisis in 1995-96 (subsequently compounded by broader macroeconomic crisis) led directly to the introduction of a new priority-setting process—initially in the form of the Brasil em Ação mechanism (1996-98) aimed at prioritizing 42 major programs, and subsequently the development of improved systems of multi-annual performance budgeting. These budgeting reforms were closely integrated with reforms of aggregate fiscal policy-making (IMF, 2001). In other countries, less acute—but nevertheless very real—aggregate fiscal stress has played a similarly important role in promoting performance budgeting developments. Thus the case study of Spain (Chapter 17) makes it clear that the fiscal pressure associated with the Maastricht fiscal rules has been a crucial driving force behind performance budgeting reforms in that country.

Turning from the nature of the expenditure prioritization process to the actors in that process, the case studies in this part of the volume highlight the crucial role of the Ministry of Finance. If there is to be good expenditure prioritization, MoF officials must be able to advise political leaders on reallocation options—and, in particular, on where cuts may be made to fund new priorities and imperatives. It is therefore essential that the MoF concern itself closely with the type of performance information with which it is provided by line agencies, in order to assure itself that this information is what it and the political leadership require. The type of involvement which this requires is well-illustrated in the guiding role which, as Brice Lannaud makes clear in Chapter 11, the French finance ministry has played in the development of both the program classification and performance measurement system which underpins the new French performance budgeting system. Similarly, in Chile (see Chapter 13 by Marcela Guzmán) the budget directorate of the MOF not only specifies the key performance indicators to be provided by line agencies in the budget process, but also manages the system of program evaluation which is the distinctive feature of Chilean performance budgeting. In the US, it is Office of Management and Budget examiners—not outside evaluators, or other civil servants—who use PART to systematically gather from line agencies the information and assign summary performance ratings to programs for use in the budget process. As this example makes clear, it is critical that MoF officials carry out their own assessments of program performance, using the information available to them. They must, in short, become program analysts as well as financial analysts. This point is all too often overlooked by countries introducing performance budgeting, as is illustrated in the Spanish case study (Chapter 17), in which Ruiz-Herta et al. identify as a key problem the “imbalance” which exists between the volume of performance information which the MoF requests of line ministries, and its capacity to process and analyze that information.

The central role of the MoF makes the bureaucratic “muscle” of the MoF an important determinant of the success of performance budgeting in improving expenditure prioritization. This is a point which is well-illustrated by the examples of Chile and the UK, in both of which the finance ministry and finance minister have great authority vis-à-vis line agencies and ministers.

As crucial as the role of the MoF is, expenditure prioritization cannot be regarded as a purely technical task carried out by civil servants based solely upon technical criteria. As Curristine emphasizes in her overview of performance budgeting in OECD countries (Chapter 8), expenditure prioritization is necessarily political, and electoral calculus is inevitably important in the budgetary decisions taken by political leaders. To set up an expenditure prioritization mechanism which explicitly or implicitly excludes political input is almost certainly to doom it to irrelevance. In general, therefore, expenditure prioritization procedures need to be designed so as to feed useful information on the costs and results of existing programs, and expenditure alternatives, to the top political budget decision-makers, who can then weigh this information together with other relevant considerations including electoral pressures. This information, as Pierce and Di Francesco emphasize in Chapter 14, must be provided in a usable summary form which recognizes the severe time constraints facing political leaders.

The corollary of this, as Curristine again emphasizes, is that expenditure prioritization mechanisms will not work without strong commitment from the political executive. Not even the best-designed performance budgeting system can be expected to have any positive impact on expenditure prioritization if the political leadership of the executive branch is indifferent to the efficiency and effectiveness of public expenditure. One of the best examples of the positive impact of strong political commitment is the UK, where the Chancellor of the Exchequer (finance minister) actively leads the Spending Reviews and both the reviews and the associated Public Service Agreements have the strong support of the Prime Minister.6 The best way of organizing political leadership of the expenditure prioritization process will vary from country to country, depending in part on the political system. In countries with a strong executive presidency—Chile being a case in point—the involvement of the President is of paramount importance. On the other hand, the more powerful the cabinet (council of ministers), the more important it is that expenditure prioritization be a systematically-organized cabinet process. Thus, an important feature of the Australian ERC mechanism referred to above is the way in which it secures cabinet consensus on key spending priorities, and strengthens the hand of the finance minister vis-à-vis spending ministers.7 This is also one of the objectives which has been pursued in Spain through the establishment of the cabinet-level Spending Policy Commission which, as Ruiz-Huerta et al. make clear in Chapter 17, has been intended as an instrument to transform the spending prioritization into a multilateral negotiation process rather than a purely bilateral one between the finance ministry and each spending minister.

It is often asserted that it is crucial to the success of performance budgeting that legislators learn to make extensive use of performance information when voting on the expenditure priorities established in the annual budget. However, such hopes have in practice usually been disappointed. Thus Curristine reports survey results indicating that only 19 percent of OECD legislatures, and only 8 percent of members of legislative budget committees, make use of performance information in decision-making. Even in the UK, where the government has nailed its colors to the mast through explicit commitment to performance targets for public services, Smith (Chapter 12) reports that parliamentary scrutiny of performance against these targets has been rather limited. Perhaps a little more realism may be required in respect to parliamentary use of performance information. The time constraints, limited support services, and political incentives facing legislators are such that it may not be appropriate to expect too much from them by way of review of program performance. Moreover, in many political systems the actual power of the legislature over expenditure prioritization is quite limited. This is true, to a greater or lesser degree, in most parliamentary systems, where the fact that the political executive is drawn from the party or parties which control the parliament usually leaves real (as opposed to formal) power over expenditure prioritization in the hands of executive government. It is also true in a number of presidential systems where the budgetary power of the legislature is constitutionally constrained, often as a consequence of historical experience with legislative fiscal irresponsibility. Chile falls into this category, as also did France prior to the new budgeting reforms. Prior to the passage of the Loi Organique relative aux Lois de Finances (LOLF), the French parliament was unable to change in any way the allocation of resources in the budget proposed by the government. As Lannaud describes, a key breakthrough embodied in the LOLF is that the parliament has acquired the power to shift funds between “programs” within the same “mission” (although not between missions). France is unusual in that the prime impetus for the introduction of performance budgeting came from the legislature rather than executive government, and it will therefore be particularly interesting to see what use parliament makes use of its new powers, and of the extensive performance information with which it will be henceforth provided.

Linking results and funding more tightly to promote better agency performance

Improved expenditure prioritization is not the only objective of performance budgeting. Although it is the primary objective of program budgeting, there are other forms of performance budgeting which differ in the way in which they seek to link funding and results, and in the objectives which they aim to achieve by building that linkage. Chapter 1 identified three other performance budgeting mechanisms which have been an important part of recent performance budgeting developments, namely:

  • funding-linked performance targets

  • agency-level budgetary performance incentives

  • formula funding.

None of these mechanisms are concerned with improving expenditure prioritization in the central budget. Rather, their aim is to put pressure on agencies to improve the effectiveness and/or efficiency with which they spend public money. To achieve this, each seeks to tighten further the link between funding and results. Program budgeting aims to deliver a significant, but nonetheless very loose, funding/ results linkage—by ensuring that information on the effectiveness and efficiency of programs is systematically taken into account in making expenditure prioritization decisions. By contrast, the more contemporary performance budgeting mechanisms seek to build either or both tighter ex-ante and ex-post linkages between funding and results.

Tightening the ex-ante links between funding and results means relating the level of funding provided to each agency more closely to the results which that agency is expected to deliver with that funding. The tightest form of such linkage is formula funding, in which funding is a mathematical function of expected results. The idea of funding-linked performance targets, on the other hand, aims to ensure that targets are appropriate to the level of funding, so as to increase their credibility and motivational impact.

Performance-based financial incentives to agencies aim, by contrast, to tighten the ex-post linkage of funding to results. The linkage, in other words, flows from actual (rather than expected) performance to funding levels, and the expectation is that performance-contingent financial rewards or sanction will motivate agencies to perform better. The strongest form of performance-based financial incentive is that which is created by “purchaser-provider” models. As outlined in Chapter 1, in purchaser-provider systems funding is linked by formula to actual results (“payment-for-results”) and, as a consequence, government agencies make profits or losses depending on their performance. “Performance bonuses” are a weaker version of the performance-based financial incentives mechanism, in which agencies receive top-up funding based on selected performance measures, as a supplement to core funding which is unrelated to actual performance.

One of the aims of the material in this part of the volume is to shed light on how well these mechanisms work, and under what circumstances. In this respect, certain key points emerge. The first is that the effort to build stronger links between funding and results has been most successful in sectoral performance budgeting systems. As Smith outlines in Chapter 12, the formula funding principle has been widely applied in the public sector in areas such as education and social services. Perhaps the most striking example of the success of sectoral performance budgeting systems, however, is the “diagnosis related group” purchaser-provider system of hospital funding. There is strong evidence that this system has delivered significant improvements in technical efficiency and that fears that it would induce erosion of the quality and other perverse effects were largely unjustified.

It has not, on the other hand, proven possible to create such tight linkages at the level of the government budget as a whole. The “accrual output budgeting” experiment in New Zealand and Australia represented the most ambitious attempt ever to link funding and results tightly right across the budget sector, inspired by the purchaser-provider principle. As discussed in Chapter 16, it was not a success. A key reason for this is that the relationship between output and cost is, for many government services, too uncertain to permit a tight linkage of funding to the quantity of outputs delivered. The lesson to be learnt is that both the purchaser-provider mechanism, and the formula funding mechanism more generally, are impractical in the case of many public services.

At the level of the government budget as a whole, budget-linked performance targets seem to have worked better. The British Public Service Agreement (PSA) system, which Smith discusses in Chapter 12, appears to have been very successful in boosting public sector performance, although there is debate about perverse effects (see below). Two factors seem to have been particularly crucial to the success of the PSA system. The first is that the pressure on agencies to meet the targets is considerable, reflecting strong commitment from the political leadership. The second is that the targets are set as part of the budget process, under the leadership of the finance ministry and finance minister.

This contrasts markedly with experience elsewhere. Although the setting of agency-level performance targets has become quite common internationally—under the influence of the broader current of “managing-for-results” thinking—such target-setting is often beset by the problems that, first, the pressure on agencies to meet the targets is weak, and, second, that target-setting is carried out quite separately from, and often with no regard to, funding decisions. Underlying this, as Pollitt (1999, p. 5) notes, is the fact that “financial management and performance management systems tend to develop separately as parallel systems that may or may not (or only to varying degrees) be harmonious or even compatible.” In many countries, government-wide performance management systems—including the processes for setting performance targets—are managed by institutions separate from the finance ministry (President’s offices. Prime Minister’s departments, or civil service commissions), without any links to the budget process.

By contrast, experience in countries like the UK or Chile suggests that assigning institutional responsibility for target-setting to a powerful finance ministry and finance minister underlines the political commitment to the targets and helps to ensure that they will be taken seriously by agencies. It does this partly by strengthening the perceived ex-post linkage between results and funding. In other words, if the finance minister and finance ministry are at the center of the target-setting process, agencies are far more likely to believe that their performance against those targets will be considered in determining future budget allocations.

A striking feature of the PSA system is that it is primarily outcome-focused; that is, the great majority of PSA targets focus on outcomes such as student knowledge, and reductions in rates of cancer and heart disease. This outcome-orientation is representative of a broader international trend in performance budgeting at the level of the government budget as a whole. The international vogue in the 1990s for output-focused performance budgeting—influenced by the New Zealand experiment—has now faded. The dominant contemporary view is now that, at the level of the government budget, the focus should be on outcomes. Outcomes are, after all, the true goals of public policy, whereas outputs are a means to that end.

The outcome focus puts a spotlight on the key question of how tightly one can link funding to results. As discussed in Chapter 4, the relationship between outcomes and funding is often particularly uncertain, because of the impact of “external factors” on both the magnitude and the timing of outcomes. For example, the results which can be achieved by public programs designed to reduce cancer are highly uncertain for a range of reasons, including the partially uncontrollable nature of the risk factors (dietary and smoking habits, environmental deterioration, and so on) which impact on the incidence of the disease. So if targets are to be set for the outcomes expected over, say, a three-year time horizon, and the level of funding provided is to be linked to these targets, how does one deal with this uncertainty? Exactly what type of results/funding linkage can be built under such circumstances?

The answer, as Smith makes clear, is that there can indeed be no tight calibration between outcome targets and the level of budget funding. Often, the linkage will be no tighter than an expectation that if funding is increased, tougher targets will be set. Even such a loose linkage is, however, arguably important and beneficial, particularly by contrast with a system in which targets are set completely independently of the budget process, and without regard to the funding levels provided.

This implies that the ex-post linkage between outcomes and funding must also be a loose one. As a number of studies in this part of the volume make clear, it is unrealistic to expect any mechanical linkage between past performance and future funding. Thus Fantone stresses that there is no automatic relationship between a program’s PART performance rating and the level of funding proposed for that program in the President’s budget. Curristine reports that survey results suggest the same to be true of OECD countries as a whole. This is hardly surprising. It is common sense that if a program is performing badly, there can be no assumption that a funding cut is the appropriate response. It may be that program redesign or management change is what is required, or even that additional funding is needed.

This suggests that it is not useful to think of the performance incentives which performance budgeting can create at the level of the government-wide budget as “rewards” and “sanctions” analogous to those which operate in market transactions and which are mimicked by sectoral purchaser-provider systems. It is notable in this context that the one attempt to create a government-wide system of explicit agency-level rewards for good performance—in the Canadian state of Ontario—endured only a couple of years. Nevertheless, even if agencies are well aware that there is no automatic connection between performance and future funding, the knowledge that past program performance will be an important consideration in expenditure prioritization decisions can be expected to help motivate them to perform better. So rather than treating the government-wide budget process as a means of creating financial rewards and sanctions for agency performance, it is probably better to focus on creating good, performance-informed expenditure prioritization in the knowledge that one of its by-products will be to encourage agencies to improve program design and management so as to reduce the risk of having their budgets cut in the prioritisation process.

Changing behavior through performance budgeting and management

There is a further set of important issues concerning the behavioral impact of those forms of contemporary forms of performance budgeting—and “managing-for-results” (MFR) more generally—which aim to encourage and pressure agencies to perform better through the specification of expected levels of performance, whether through performance targets or other means. One concerns the potential for behavioral distortion. It is well-established in both theory and practice that performance measures and targets which are incomplete or capable of manipulation can induce behavioral distortions (see, for example, Ridgway, 1956; Smith, 1990, 1995). Given that the “existence of such perverse effects is uncontroversial” (Pidd, 2005, p. 483), the real question at stake is whether their magnitude is such as to throw into doubt the very efficacy of MFR and MFR-related performance budgeting systems. This question has been the subject of vigorous public controversy in some countries. As Smith outlines, in the UK, critics of the Public Service Agreements regime have complained about the alleged perverse affects of PSA performance targets in a range of areas. Although there is considerable general literature relevant to the issue of perverse effects—such as literature in the field of organizational economics—there is relatively little analytic literature on the question in the specific context of the public sector. To the extent that the public sector has different characteristics to the for-profit sector, this presents difficulties. It has been important, therefore, to provide some treatment of this issue in this part—both through discussions in a number of the case studies (in particular, that of the UK), and also in a thematic chapter which examines questions of behavioral impact more generally.

The question of behavioral impacts is important not only because of the issue of behavioral distortions, but also because of the issue of the source of motivation to perform more generally. Although many contemporary forms of performance budgeting and MFR aim to induce agencies to perform better, it is not always completely clear how it is intended that this impact be achieved. In the case of those forms of performance budgeting which seek to create financial rewards for agencies with good performance, the mechanism is in a sense explicit. However, whether agency-level financial rewards are part of the performance budgeting mechanism or not, a more fundamental issue arises—how can the individuals who staff the agency be induced to perform better? Through stronger incentives—financial and other explicit rewards and sanctions such as performance pay—linked to the achievement of organizational goals and targets? If so, this reinforces the importance of considering the question of behavioral distortions, because it is once again a familiar theme in the literature that the more “high-powered” (stronger and more direct) performance incentives are, the greater the potential for behavioral distortions if the measures upon which those incentives are based are imperfect. Another concern raised by critics of MFR is that sharper extrinsic incentives may undermine more “altruistic” or public spirited motivation to perform, which it is argued is particularly important in many areas of the public sector. How serious are these concerns and what are their implications for performance budgeting? To the extent that public employees are primarily motivated by altruistic commitment to the public good, is giving them the freedom to perform—by reducing controls which limit their freedom of action—perhaps more important than giving them incentives to perform and monitoring their performance?

Despite the importance of these behavioral issues, they have not been well-researched. To deal with them thoroughly would require deep foundations in the theory and empirical literature of industrial motivation and behavior, which is something well beyond the scope of this volume. However, in recognition of their importance, Chapter 18 reviews relevant literature, attempting to suggest links between performance budgeting and sources of individual motivation, and to draw out implications for the design of incentive systems.

In conclusion, the objective of the thematic and case studies in this part is to shed light on what types of performance budgeting system can potentially work, and whether they are potentially effective at the central budget level and/or at the decentralized level. Knowing what can potentially work is clearly of great importance to those designing performance budgeting systems in deciding, from the menu of choices available, what type of system they should be attempting to build. It does not, however, provide all that is required to make such decisions. The other crucial consideration is whether relevant preconditions have been met for the successful adoption of certain forms of performance budgeting. Preconditions, as well as sequencing and implementation strategy, are considered in Part Three.

Notes

Incrementalism has been defined as “inattentiveness to the (budgetary) base”—in other words, that budgetary decision-makers take the budgetary base more or less for granted as the starting point in budget formulation, and focus their attention on the size of the increment (or, occasionally, decrement) in agency or program budgets (Berry, 1990).

Such as the Grace Commission appointed by President Reagan in the US or the Rayner scrutiny reports commissioned by the then Prime Minister, Margaret Thatcher, in the UK.

The need for expenditure prioritization to focus on the totality of government expenditure is closely related to the widely-accepted case for budget comprehensiveness (including by limiting the use of extrabudgetary funds) (see, for example, Allen and Tomasi, 2001, p. 49).

In the 1960s and 1970s, advocacy of the early program budgeting models was often linked with sympathy for so-called “indicative” planning of the economy as a whole.

Alternative budgeting, as noted in Chapter 1, involved requiring line ministries each year to present in their budget submissions options for, and analyses of the impact of, marginal changes of predefined magnitudes in each program (for example, 10 percent cut, 5 percent cut, 5 percent increase).

Experience in the UK and a number of other countries makes clear that it is far too superficial to regard politics and budgetary rationality as mutually exclusive, in the way that many US academic commentators, in the tradition of Aaron Wildavsky, continue to do (see, for example, the assertion of Lu (1998, p. 167) that “[the] politics of public budgeting makes it almost impossible for decision-makers to use rational data to allocate resources”). Insofar as either (1) the effectiveness and efficiency of programs has electoral significance or (2) the political leadership is committed as a matter of principle to improved public sector performance, performance information may have a significant impact in the allocation of resources notwithstanding the fundamental importance of political considerations. The tendency of US analysts to treat politics and budgetary rationality as antitheses may owe much to the specific context of the US political system, with its fragmentation of budgetary power between Congress and the executive, and the incentives which individual Congressmen have to use their powers over the allocation of expenditure to win local and interest group support.

Particularly important is the involvement of the Prime Minister and inclusion of two or three powerful ministers with big spending portfolios. By assigning these spending ministers a key role in the enforcement of overall government spending priorities, one reduces the danger that they will use their power in the cabinet to defend the interests of spending ministries against the finance ministry.

References

    AllenR. and D.Tomasi (eds.) 2001Managing Public Expenditure: A Handbook for Transition Countries (Paris: OECD).

    AnshenM.1967The Program Budget in Operation,” in Program Budgeting: Program Analysis and the Federal Budgeted.D.Novick (Cambridge, Mass.: Harvard University Press).

    BerryW.D.1990The Confusing Case of Budgetary Incrementalism: Too Many Meanings for a Single Concept,Journal of Politics Vol. 52(1) pp. 16796.

    CamposE.J. and S.Pradhan1999Budgetary Institutions and the Levels of Expenditure Outcomes in Australia and New Zealand,” in Fiscal Institutions and Fiscal Performanceed. James M.Poterba and Jürgenvon Hagen(Chicago: University of Chicago Press).

    IMF2001Brazil: Report on Observance of Standards and Codes (ROSC)—Fiscal Transparency Module (WashingtonDC: IMF).

    Le HouerouP. and R.Taliercio2002Medium Term Expenditure Frameworks: From Concept to Practice: Preliminary Lessons from AfricaAfrica Region Working Paper Series No. 28 (WashingtonDC: World Bank).

    LuH.1998Performance Budgeting Resuscitated: Why is it Still Inviable?,Journal of Public Budgeting Accounting and Financial Management Vol. 10(2) pp. 15172.

    Office of Program Policy Analysis and Government Accountability1997Performance-Based Program Budgeting in Context: History and ComparisonReport 96-77A. Florida Legislature.

    OsborneD. and T.Gaebler1993Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector (New York: Plume).

    PiddM.2005Perversity in Public Service Performance Measurement,International Journal of Productivity and Performance Management Vol. 54(5/6) pp. 48293.

    PollittC.1999Performance Management and Financial Management—How to Integrate Them? (Paris: OECD).

    RidgwayV.F.1956Dysfunctional Consequences of Performance Measurement,Administrative Sciences Quarterly Vol. 1(2) pp. 2407.

    SchickA.2002Does Budgeting Have a Future?OECD Journal on Budgeting Vol. 2(2) pp. 748.

    SmithP.C.1990The Use of Performance Indicators in the Public Sector,Journal of the Royal Statistical Society A Vol. 153 (1) pp. 5372.

    SmithP.C.1995On the Unintended Consequences of Publishing Performance Data in the Public Sector,International Journal of Public Administration Vol. 18(2/3) pp. 277310.

    SpackmanM.2002Multi-Year Perspective in Budgeting and Public Investment Planning,Draft background paper for discussion at session III. 1 of the OECD Global Forum on Sustainable Development: Conference on Financing Environmental Dimension of Sustainable DevelopmentOECDParis.

    World Bank1998Public Expenditure Management Handbook (Washington, DC: World Bank).

    XavierJ.A.1998Budget Reform in Australia and Malaysia Compared,Public Budgeting & Finance Vol. 18(1) pp. 99118.

    Other Resources Citing This Publication