9 Improving Allocative Efficiency

Marc Robinson
Published Date:
October 2007
  • ShareShare
Show Summary Details
Jim Brumby1

The traditional view of public finance attributes three major goals to the management of public finance: stabilization, reallocation, and redistribution.2 It is the interrelationships between these that give rise to the major issues in public finance. For instance, provision of public educational services through the levying of taxation affects real income distribution, while distributional activities inevitably affect relative prices and consumer demands, and therefore the allocation of resources to the production of various goods. Of these interrelationships, there is perhaps no more crucial question for public finance than the following: How can society collectively determine the optimal amount of public goods, and how shall the costs of these goods be apportioned among individuals (Due and Friedlaender, 1973, p. 48)? The scope of this chapter is, first, to acknowledge the contribution that the answering of this question can have for the goal of stabilization through management of the aggregate fiscal position, but mainly to examine the specific mechanisms that can assist improving allocative efficiency through public budgeting.

The chapter begins by looking at the notion of allocative efficiency and reviewing the root causes of changes in allocative efficiency. It then sets up a tentative framework of modern budgeting mechanisms, prior to describing each of these mechanisms, focusing in particular on mechanisms to give effect to changes in government priorities and in relative prices of government intervention.

Defining allocative efficiency

A reasonable definition of allocative efficiency is the production of the best or optimal combination of outputs by means of the most efficient combination of inputs. Allocative efficiency involves the output combination which would be chosen by individual consumers responding in perfect markets to prices which reflect the true costs of production. Pareto optimal output involves the point at which the marginal social value equates with marginal social cost; accordingly, it involves efficiency in the product mix and efficiency in consumption.

Adapting the concept of allocative efficiency to the public sector is far from straightforward: resource allocations are not made by individuals; often, neither tastes nor outputs are well-articulated; the relationship between the output and some measure of welfare, such as outcome, may be difficult to gauge; input prices are often distorted and not transparent; and data on true costs of production are often not available. In addition, there is the problem of the aggregation of preferences—individuals with different preferences have different marginal rates of substitution at the efficient scale of output for publicly supplied goods. It is well appreciated that there is no market mechanism that will determine the efficient scale of output of publicly provided goods or allocate the costs of this production among the members of society. Political and bureaucratic institutions, in all their flawed glory, are left to do this.

Over the past decade, an accepted way of looking at budgetary institutions has been to see three levels of efficiency-related objectives associated with budgeting. At the highest level (Level 1), there is a concern about the aggregate expenditure control and the burden placed by the state on the productive sector of the economy. Nicholas Barr would call this macro efficiency, but many budget writers may just refer to this as being concerned with aggregate control (see Barr, 1998, p. 9; World Bank, 1998, p. 2).3 At the second level (Level 2), the objective is to pursue interventions which are consistent with the priorities of society, as represented by decision makers. On occasion, this level is referred to as being concerned with allocative efficiency. The third level (Level 3) is concerned with the government’s role as a supplier of services, seeking production of such services in a technically efficient way. Improving allocative efficiency would see reallocation from lower to higher priorities of government and from less effective to more effective programs.

Allocative efficiency and fiscal stabilization

It is axiomatic to the tasks of public budgeting that there are tensions between the aggregate and its parts. The three levels of efficiency-related objectives are interrelated. Perhaps it is trivial to make the point that without some form of constraint at the aggregate, there would be no scarcity, and there would be no public budgeting problem as such. But, alas, this would lead headlong into a dispute with what, from a stabilization point of view, the economy could sustain as an achievable aggregate fiscal position. The reality of budgeting is that the sum of the parts will hold as a constraint only if portfolio pressures are constrained. The realization of the need to manage this common pool problem has led to many reforms to strengthen this role of fiscal institutions, especially the central budget office, over the past 20 years or so.4 Evidence of this is seen in the increased use of budget limits being placed on budget formulation by central budget offices, rather than allowing an aggregation of bids.5

The potential contribution of achieving productive efficiency (Level 3) objectives to assist with the achievement of (aggregate control) Level 1 objectives appears straightforward. Improved productive efficiency provides a choice for resource allocates (that is, those with decision rights over resource allocation), along the following lines:

  • the efficiency gain can be reapplied to the same productive process to yield higher output;

  • it can be reapplied to another output;

  • it can flow through to the bottom line and result in fiscal consolidation (lowering the estimated cost of taxes through time); or

  • it can be returned to taxpayers via a tax cut now, or in the future through lower debt.

Similarly, there is a potential contribution from improved expenditure prioritization (allocative efficiency) to Level 1 control objectives, along the following lines:

  • a reduction in production of one output (or the level of resources to one intervention) may yield a fiscal dividend which can be reapplied; or, alternatively,

  • an increase in an output (or resources devoted to an intervention) will need to be funded from some source (either through a transfer from another output, taxation, or debt).

The better the ability of government to identify its priorities, and as important, its non-priorities, the easier the task of managing the aggregate constraint and the more credible that constraint. Further, provided there is limited dissonance between government priorities and those of the people, there may be less pain associated with fiscal consolidation, even that involving fiscal retrenchment.

Linking allocative efficiency and performance budgeting

A public finance construct that sought to describe government, well-informed and non-distorted, as a central point for making decisions on the marginal welfare gains of alternative spending options or an additional dollar of spending versus a reduced dollar of taxation, could be considered somewhat fanciful. With apologies to Alchian and Demsetz (1972), if the firm is a nexus of contracts, then government organization is a labyrinth of (hard-to-measure) formal and informal contracts, executed in space protected from the self-correcting powers of the market.6

There are a number of factors associated with the state which introduce complexity on the design and oversight of resource allocation procedures (Gorringe, 2001, p. 126). These factors include:

  • the power of the state (associated with tyranny, commitment, and capture);

  • centralization (associated with making decisions and coordinating action, limited competition);

  • organization (agency and other contracting issues); and

  • collective choice (free-riders, rational ignorance of voters, limitations of voting systems, and so on).

Modern budget management attempts to address these factors. For instance, the twin initiatives of applying increased managerial flexibility alongside more accountability is a clear case where political decision-makers have assigned some of their rights to others. In this one area, there is a specialization of role and function, which deals with centralization; there is a control placed on government power, which deals with problems associated with credible commitment; and there are also issues associated with organizational management. If increased managerial flexibility is provided as a way of delivering on a harder and more credible budget constraint, consistent with the political objectives, then it would also be reasonable to expect such flexibility to be positively related to improved fiscal outturns.7 Notwithstanding these sorts of initiatives, the characteristics of many of the services being supplied by government (such as policy advice and diplomatic representation) suggest that they rely at least to some degree on the organizational form of a bureaucracy, in which the costs of making decisions and implementing them are high (Ouchi, 1980).

As Schelling has pointed out (in a different context), a budget manager can take lessons from traffic signals—although planning is often associated with control, the crucial element is often coordination (Schelling, 1978, p. 121).

In this context, the role played by the provision of performance-based information can be considered. Such information moderates the problems associated with information asymmetry, allowing at least some informed consideration as to whether government interventions are working.

Changes in allocative efficiency

There are likely to be three main factors driving a reallocation of government resources reflecting changed priorities to increase allocative efficiency: preference changes; relative cost changes; and affordability changes caused by movements in national income.

Preference changes

There are several sources of change associated with preferences. By its nature, government is concerned with collective action and collective will. Those charged with power to allocate resources do so based on some concept of the will of the governed. Therefore, if either the will of the governed was seen to change or the way that that will was relayed politically, then that would amount to a preference change. At times, events may lead to rapid changes in preferences, such as an increased preference for domestic protection in the US following the terrorist attacks of September 11, 2001. Politicians are, of course, active players in the creation of, and interpretation of, public preferences.

In the language of performance budgeting, preferences are generally concerned with outcomes, with the concern being, for example, to improve the health of the population, its education achievement, or, perhaps, perceived safety. Although empirically, we can observe that at times the public or a political party may express their preferences in terms of an output (such as improved taxation inquiry services) or even inputs (such as having a preference for the public sector making some services, such as higher education), for the most part, it is reasonable to argue that changes in political preferences are concerned with desired outcomes. In this way, it is possible to see a corollary between changing preferences based on outcomes, and consumer theory, which is concerned with maximizing an individual’s utility.

There are two important asides here. First, just with an individual, a lack of observed outcome success does not necessarily translate into more (or less) funding. It may lead to the redesign of the intervention. For instance, lack of success in drug enforcement programs may or may not yield an increase or a reduction in funding. Second, political agency problems and poor budgeting procedures can combine to make a mess of expenditure prioritization. Quite simply, governments have often allocated resources to poor uses. In poor budgeting systems, preferences are generally not well articulated—the clearest statement of preferences may come through analyzing actual budget allocations.

Relative cost changes

There are numerous ways in which the cost of an achieved public intervention may change. First, and most simply, changes in input prices may feed through to changes in output costs, ceteris paribus.8 Second, while the output characteristics may not change, the input technologies may change, involving a different configuration of inputs, yielding a different output cost. Third, the nature of the output may change through a redesign of the intervention itself, such as by choosing a different policy instrument. (For instance, measures to reduce demand through preemptive interventions have been used in the judicial systems (pre-trial conferences), and have been recently adapted in some health systems, where telephone consultations between health providers and patients are used to sort patients into those who should access outpatient services and those who should not.) The task of the budgeter is to manage these potential sources of change in such a way as to achieve the agreed fiscal and budgetary objectives.

In some cases, there may appear to be a mixture of the two effects. For instance, the overwhelming evidence of an emerging aging problem—involving a smaller share of the population engaged in the workforce—produced a change in the perceived costs associated with universal pension schemes, and a relatively abrupt change in the tastes of some countries for universal, tax-supported pension and health systems. Arguably, the preference for financially sound older people did not change, but the realization of the non-affordability of achieving that through traditional policy means has led to a substitution of new forms of intervention, replacing the old forms.

Income changes

As the aggregate budget constraint may become relatively more relaxed as income increases, there are likely to be different mixes of consumption that will satisfy allocative efficiency Decreasing marginal demand for certain goods or services is often associated with necessities, while increasing demand is associated with luxuries. In government, the (intended and revealed) consumption patterns of poor countries are discernibly different from those in advanced countries.

The particular budget management issues associated with sudden changes in affordability have given rise to a subset of budget management literature. In particular, additional strains can be placed on the quality of budget allocation in economies which are exposed to wide income variations perhaps related to high dependence on trade, especially commodities.9

From sources of change into plans of action

These sources of change to resource allocation are not always self-evident and frequently require considerable finesse to see them converted into reallocations. As Schick has commented, to achieve positive public expenditure outcomes, it is necessary that information, incentives, and other institutional arrangements be properly aligned (Schick, 1999, p. 2). Contributing to these institutions and mechanisms of budgeting is a skilled and informed Ministry of Finance (MoF) focused on these matters. The limited capacity of Ministries of Finance in developing countries is one factor constraining the efficiency of resource allocation in such countries—Caiden and Wildavsky refer to this as a lack of redundancy (1974).

The transformation of traditional command-and-control MoFs into today’s modern MoFs is a necessary aspect of seizing and converting allocative change through preference, price, and affordability changes. This transformation has been commented on elsewhere (Schick, 1997, and more recently, Wanna et al., 2003). Modern MoFs are responsible for designing systems of coordination which bring together information at appropriate times, in digestible forms, to be acted upon by decision-makers who have their incentives aligned to the allocative efficiency goal. Their task is to design a budget allocation system which, through the behaviors of the participants, augment rather than detract from the goal of allocative efficiency. This requires information about the production and cost of government-funded outputs, and the degree to which these outputs contribute to the achievement of government outcomes.

Reallocative mechanisms

The desire to reallocate often stems from factors outside the budgeting system. For example, political preferences may abruptly change due to a change in government, or as a result of a natural or economic disaster. However, even in such circumstances, the use of performance information in the budget process assists as an instrument for translating preferences (including those of a new government) more effectively into expenditure allocations—that is, as an instrument for better prioritizing expenditure.

Practices across countries in the past two decades or so indicate that there have been changes associated basic mechanisms to assist improving allocative efficiency, in particular:10

  • the medium-term orientation to the budget has been strengthened to identify current and future costs of policies;

  • efforts have been made to align better budget resources with the policy priorities of government, and to reduce fragmentation across government.

In support of this, a high proportion of advanced countries and a growing proportion of emerging and developing economies introduced government-wide initiatives to improve output information at least five to ten years ago. And ongoing adaptation is continuing, with Organization for Economic Cooperation and Development (OECD) countries recording major ongoing initiatives in the past five years as well (Curristine, 2005, p. 6). Nowadays, fully three-quarters of OECD countries include performance information with their budgets (Joumard et al., 2003, p. 129).

Mechanisms to deal with changes in fiscal fortunes

An increase in available resources gives the government an opportunity to allocate these resources to their best use. The choice basically comes down to deciding whether to strengthen the financial position of government or not (to be more precise, whether to use money to increase net financial worth, or whether to spend it on increased services or lower taxes). In the first category are decisions that would hold the increased resources as financial assets (whether in a stabilization reserve or other), or use the windfall to pay down debt; in the second category may be decisions to lower the tax rates, or to increase spending in a particular way. The opposite occurs during a downturn, with the particular way that such a downturn will flow through to the fiscal position being influenced by whether the overarching control is a deficit limit control—thereby forcing either an increase in taxation or a reduction in spending—or an appropriation control, in turn forcing an increase in taxation or an increase in debt in order to meet appropriation requirements.11

When countries face high fiscal stress, then prioritization may be most important and crucial. A failure to prioritize may see an unaffordable increase in aggregate spending, with ensuing flow-ons to the tax rate, and associated deadweight costs. If the fiscal scoring system relies on cash rather than accrual accounting, then capital spending is particularly at risk. This was the case in Europe during the 1990s (Brumby, 2000, p. 3). But history suggests that governments have often stumbled in allocative efficiency in good times as well. When the marginal value of each tax dollar seems lower, then the decision-maker may be prone to allocating it particularly poorly.

In the same way that the free cashflow problem has been documented as a bete noire for firms seeking high-returning investments (Jensen and Meckling, 1976), fiscal abundance can induce growing allocative inefficiency. In both cases, there is a lack of credible constraints on the power of the decision-makers. Bigger, more diversified government does not necessarily satisfy voters, just as bigger, more diversified firms do not necessarily satisfy shareholders. Indeed, one recent empirical study showed that governments interested in maximizing the life satisfaction of their voters should, regardless of their ideology, limit their direct interventions in the economy to allow voters a high degree of personal freedom (Bjornskov et al., 2005, p. 23). To the end of constraining the available discretion, there have been two major initiatives in recent years: the development of fiscal rules, and the creation of fiscal boards.12

Mechanisms to reflect political preferences through strategy management

Allocative efficiency can be enhanced if, assuming a robust political process, resources flow to where they are most valued by the ruling government, as agents of the people. An important source of articulation of these preferences may be through the strategic phase of the budget process. The strategic phase involves agreeing to the following:

  • an aggregate constraint to apply to all spending, in the context of the outlook for the economy over the medium term;

  • an indicative scope for the pursuit of major new initiatives;

  • identification of sectors/policies and programs for special treatment, either for cuts or additional spending; and

  • identification of the general rule to apply in transmitting limits to government departments.

Box 9.1.Budget formulation in developing countries

The benefits from the strategic phase of the budget may be more tenuous in countries where the annual budget is not implemented as presented and fiscal blight occurs regularly. In such circumstances, many of the decisions which actually allocate resources are not made during formulation at all but are, in reality, made during execution, often via cash rationing. In such circumstances, the immediate need is to put in place a budget which could be considered realistic and which has some hope of implementation. Sadly, the chronic disconnect between budgeting and planning in many poor countries was identified by Caiden and Wildavsky in Planning and Budgeting in Poor Countries (1974) some 30 years ago. At the very least, a credible strategic phase of the budget will require effective budget implementation.

In addition to the problems of perpetual fiscal blight, some developing countries may have suffered from resource misallocation associated with corruption. One mechanism designed to exert pressure on improving the quality of the internal control systems in order to ensure that resources go to where they are meant to go is the Public Expenditure Tracking Survey (PETS). The initial set of PETSs carried out in countries such as Uganda and Mozambique showed there to be a high proportion of leakage from the intended purpose stated in the budget.

One explanation for the lack of credibility associated with the development of robust budgeting systems in such countries is that they are ineffective in constraining or dealing with the types of uncertainties that such countries face. In particular, high levels of uncertainty may induce short-sighted behavior and a preference for systems which can be manipulated to meet short-term needs. An extreme form of this dynamic incentive (and misallocative equilibrium) could be associated with high adult mortality rates. In such circumstances, adults are unlikely to invest in actions that generate long-term benefits and short-term costs, preferring instead to take actions generating short-term benefits at long-term costs.a

a See a recent study by Lorentzen et al. (2005).

Aggregate constraint

The advice to government on the aggregate expenditure limit is generally the core work of the minister of finance. This is where Musgrave’s concern for stabilization and efficient allocation between the public and private sectors come to the fore (Musgrave and Musgrave, 1959). The minister needs to convince his colleagues of the desirability of a limited spending envelope. To do this, he would be expected to enlist the leader’s assistance. The argument for the aggregate limit may need to focus on the advisable tax take, the appropriate debt setting, and the pressures from the community to spend—the sorts of issues that Schick has raised in terms of sustain-ability. By focusing on the limit and the collective aspirations of the government as a whole, it may be possible for the leaders to gain collective agreement.

Managing the baseline

Decision-making behaviors suggest that unless there is real growth in the expenditure aggregate to allow new allocations, decision-makers will be faced in an unsolvable dispute where no one spender will be prepared to give up resources to fund the initiatives of another spender.13 The Minister of Finance can try to break this impasse, without adding to real spending, by imposing an across-the-board rebasing of all agencies. The determination of the amount of and the reasoning behind this device may differ through time and between jurisdictions, but its purpose is clear: to meet the need to constrain the aggregate and provide some funds for new initiatives or higher priority spending (or reductions in taxation). In some cases, this may be subject to a very simple rule—that ministries’ budgets will not increase in nominal terms—while on other occasions, there may be a set of quite specific deflators (or adjustors) attached to different input or output types. Though it could be argued from an allocative efficiency point of view that such across-the-board cuts are likely to induce misallocations, this can be disputed if the cuts are kept reasonably modest, in the 1-2 percentage point range of administrative spending. Notwithstanding the very large productivity increases witnessed in knowledge-intensive industries in advanced economies over the past 15 years or so, there appears to be reluctance for government to impose a standard efficiency dividend of more than 1.5 percent of administrative costs.14

The degree of flexibility on the spending side can also be assisted by the specific identification of spending programs that will be disbanded. But in many cases, there are no incentives for ministers to bring forward such plans unless they believe that they will receive a comparable level of resources returned to them. There are several ways that this can be dealt with—one is to seek indicative proposals for, say, 2-5 percent cuts from each ministry, with the Ministry of Finance reviewing the veracity of the claims, and the ministers determining collectively which they would like to pursue. There are two important ingredients to ensuring a contribution from this sort of exercise: (1), to ensure that incentives are working to have agencies’ reveal the right information; and (2), to ensure that the MoF’s officials are sufficiently knowledgeable in the areas of the agencies’ businesses that they can interrogate the agency staff. Another way is to use some fomalized program review, either collectively or bilaterally—this is discussed later.

Identification of specific initiatives

During the strategic phase, individual ministers or spending agencies can be expected to make arguments that their initiatives which may require additional funding are supportive of the government’s key priorities. In so doing, they would be expected to align their requests with perceived shifts in public opinion and the policies of the government. For a government already in office, the underlying argument is that the particular initiatives will yield better results than those lost from reductions in spending in other areas. For a newly elected government, the argument is that some of the results that have been achieved are not those that are valued by the new government—which has a set of different preferred results—which have been endorsed by the people, and are thus legitimate. In both cases—newly elected and continuing government—choices concerning policy priorities will draw on some form of citizen legitimization.

This is the core strategic management task. Strategic decision-making requires acceptance of a framework for considering trade-offs. In the 1960s, this was seen as part of the initiative to use Planning, Programming, and Budgeting Systems (PPBSs), which focused mainly on the planning task aimed at more systematic strategic decision-making, but failed to achieve this for a range of reasons including the excess bureaucratic paperwork and failed to deal adequately with the incentive problems faced by operational management. There is a general consensus that the PPBS exercises bore little fruit for a variety of reasons (Boston et al., 1996, p. 384) including their creation of an excess of bureaucratic paperwork.

Determination of the bounds of strategic versus operational management lies at the heart of the decentralization dilemma (see the appendix to this chapter). Clearly, cabinet could not run a process which considered all manner of operational decisions, as it would become overburdened, and the information needs would grind effective decision-making to a halt. Accordingly, central government agencies need to frame the agenda and the preparation of materials for the strategic phase. Different approaches to this task have been adopted; with a majority of OECD countries saying that they have in recent years introduced mechanisms to increase the focus on overall priorities (Joumard et al., 2003, p. 121).

The budget process can usefully differentiate the management of the resource base to meet the costs of pre-existing policies, and the management of new initiatives. Depending on the size and complexity of agencies, it may also prove useful to allow some discretion for reallocation to new priorities at the level of portfolio ministers. In such circumstances, not all funding requests for new initiatives would be referred to the minister of finance or the cabinet. Instead, a rule would be created to determine the characteristics of new priorities that would need to be considered by central decision-makers, versus those that could be treated at agency level. Generally, the more detailed the appropriation law, the less discretion can be provided at the agency level. Arguably, all new proposals involving substantial additional funding should be considered at the center; and presumably all initiatives which may invoke public debate should be referred to the center, whether they involve additional funds from the center, or can be funded from reallocation at the agency. Different jurisdictions, and different ruling parties within the same jurisdictions, may take a range of approaches to this issue of determining the degree of discretion over new initiatives at the agency level.

One important safeguard to ensuring that the center is not overrun by work is to assign a gatekeeper role. In this role, perhaps through an officials’ committee, all new expenditure initiatives are first assessed to ensure that they are bona fide, and that the cost implications are correctly estimated. This ensures that only validated initiatives are considered by central decision-makers, and that all baselines remain firm. Simply put, it is likely that the top-level central decision-makers (in a parliamentary system, the cabinet) will find things more manageable if such a committee is set up.

In the case of a newly formed government, the strategic phase may run somewhat differently. There may have already been a government formation process—to which officials may or may not have been a part—which has specified the most important policies and objectives for the government. These may be identified in a coalition agreement or similar other. As such, an agreement or existence of a government policy platform would be taken as the starting position for the strategic phase.

Planning period

An extended planning period—beyond the annual focus of the budget—provides a means to assess sustainability of policy settings, but is complicated with issues associated with uncertainty. Most governments have sought to use the outyear estimates to assess current settings rather than formalize this information as formal budgets. Such updated information, through, say, a forward estimates system, can assist in interpreting the dynamic position of the budget and the key drivers of change.

Alignment and fragmentation

Much of the literature about medium-term expenditure frameworks suggests that there is a role for the representation of government preferences in terms of sectoral or functional spending limits (see World Bank, 1998; Le Houerou and Taliercio, 2002, p. 10). However, there can be difficulties with converting such spending limits onto the business plans and organizational allocations that agencies require for their operations. If there is no specific responsibility aligned to the functional or sectoral area, as is often the case, then the allocation of resources within that functional area may be a fraught process involving conflicting priorities and political power, with perceptions of winners and losers. In such cases, the estimates provided are little more than numbers on a page. In an assessment of public expenditure management systems carried out by the World Bank and the IMF, this lack of integration between planning and medium-term budgeting was found to be common, with 18 of the 25 countries being assessed as having medium-term projections poorly integrated into their annual budget cycle (IDA and IMF, 2005, p. 26).

There are two main responses to address this issue: either reorganize to better align accountabilities and functional classification, or change the definitions of sectors and functions to move them away from their pre-existing definitions to better approximate the distribution of accountabilities (this is the case, for example, in Sweden, with the use of expenditure area limits). As the distribution of decision rights within government addresses many issues, many of which may be directly political, then the second approach may prove more expedient.15

Box 9.2.Managing the fragmentation threat to allocative efficiency

Government administrations have introduced many devices which have the potential to fragment the budget and protect particular interests. These can counteract the goal of allocative efficiency by lessening the degree of scrutiny on these arrangements and by introducing multiple sets of priorities, serving different interests.

Form of fragmentationCause of potential misallocationRemedial action to improve allocative efficiency
External loans and grantsImposition of donors’ priorities; draining of matching resourcesGeneral budget support
Dual budgetingFailure to link capital and operating may lead to new non-productive assetsIntegration between capital planning and operations planning; business plans
Extrabudgetary funds, autonomous agenciesOwn priorities dominate those of governmentClearing house for consideration of all expenditures (virtual funds); contractual mechanisms
Tax earmarking/user chargesOver-resourcing to one area at the expense of othersClearing house for consideration of all expenditures; contractual mechanisms
Tax expendituresBenefits mistargeted and poorly measuredEstimation and reporting of the cost; consideration with all expenditures
Uncontrolled virement between programsSubstitution of different purpose expenditureSpecifying clear rules about purpose and virement
Quasi-fiscal activities (QFAs)Subsidies being captured by special interestsTransparency or allocation of community service obligations and other QFAs
Single-year considerationLack of awareness on emerging fiscal trends leading to abrupt changes in allocationsMulti-year budget frameworks

Alignment between sectoral limits and ministerial or bureaucratic authority may be not sufficient to deal with some forms of fragmentation, a number of which are described in Box 9.2. Many of the institutional arrangements described are designed to remove resource allocation discretion from the center, in some cases, hard-wiring it through specific devices and in others, subjecting it to other forces, with the effect of biasing the allocation of resources.

Treatment of capital

A standout feature of much of the traditional public budget in many countries has been the homogeneous treatment of capital and operating expenditures. Given the different economic characteristics of capital and operating expenditures, the failure to measure and treat them differently would likely lead to misallocation.16

At the agency level, capital budgeting needs to mirror the strategic planning processes of governments as a whole; in particular, business cases need to be made in relation to the following: expanding operations; reducing costs through substitution of labor or other inputs by capital; and replacement or refurbishment of old capital stock. A capital budgeting system needs to be able to cope with all these different demands for capital funds. Consideration of the different characteristics of capital spending does not in itself increase fragmentation—indeed, examination of the linkage between asset creation and operating activities is a crucial component to strategic planning. An aspect worthy of consideration is whether there may be opportunities to make use of private sector involvement in public investment. A key long-term allocative consideration to be taken into account is the nature of and exposure to financial risks—this is a large topic to be dealt with elsewhere.

Mechanisms to consider changes in the form of intervention

The substitution of one form of intervention over another to achieve the same outcome is a means of improving allocative efficiency. In such cases, the redistribution of resources to the new intervention will be carried out if the new intervention dominates the old intervention in terms of effectiveness and, most probably, value for money.17 There have been a number of ways that budget systems have been reformed to attempt to bring consideration of the type of intervention to the fore in considering resource allocation. Some of these measures have been particularly high profile, such as the redirection of resources to consumers through vouchers, rather than the consumption of resources as suppliers of government services. To varying degrees, all these mechanisms seek to deal with the issue of informing the central decision makers, so that they can compare the effectiveness of spending programs across functional and organizational areas. A critical issue remains concerning the quantum of information to be sent to the center and the span of the center’s decision control, versus the nature of decisions left in spending agencies. Accordingly, some mechanisms which tend to focus more on routine or operational management issues tend to transmit information to the local level only, whereas more strategic information (or more material) information needs to find its way to the central decision-makers.

The following mechanisms have been used by administrations in recent years:

  • formal program review processes—also known as fundamental reviews—zero-base reviews, and output price reviews, sometimes resulting in change of policy instrument or type of intervention;

  • ongoing policy advice and review, sometimes associated with a decoupling of policy, funder, and provider; and

  • requirements to produce performance information, which in turn may be considered by auditors either in attest auditing, or via performance audits.18

Box 9.3.Integrating government resource allocation practices: Victoria’s Integrated Management Cycle

The State of Victoria, Australia, is one of many jurisdictions that implemented formally integrated planning cycles to government. The Integrated Management Cycle (IMC), which was introduced more than ten years ago, seeks to improve government sector performance by:

  • enhancing responsiveness to government’s desired outcomes and priorities;

  • strengthening resource allocation and management processes;

  • improving the quality of service provision; and

  • facilitating corporate and individual performance assessment.

The components include:

  • whole-of-government budget sector planning;

  • departmental and agency planning;

  • annual budget and resource allocation;

  • legislative planning;

  • internal government evaluation, review, and reporting; and

  • external information, reporting, and accountability requirements.

The first stage—known as Setting the Budget Strategy—is the process by which the Expenditure Review Committee of cabinet (ERC) reviews the government’s short- to medium-term overall policy and financial strategy, and sets the government’s strategic direction and framework for the upcoming budget, including: aggregate budget strategy, targets and objectives; strategic output and asset investment priorities; and department-specific initiatives that may be considered for funding at ERC Stage 2.

In setting the budget strategic output and asset investment priorities, the cabinet committee’s decisions are based on a range of information, including:

  • the government’s broad long-term social, economic, and environmental goals and measures;

  • a presentation by the Treasurer containing information on the fiscal and economic outlook including actual and projected performance against budget targets and objectives;

  • presentations from portfolio ministers on departmental strategic priorities, including proposed cross-portfolio initiatives; and

  • information and/or advice from other stakeholders.

Subject to capacity constraints, specific initiatives that are consistent with the agreed budget priorities are approved by ERC for consideration for funding at ERC Stage 2. Decisions made in ERC Stage 1 provide direction to ministers and departments in developing detailed new initiatives in the annual budget process by indicating the short- to medium-term strategic priorities of government and those initiatives that ERC may consider for funding in ERC Stage 2, in which detailed proposals from ministers for new output and asset initiatives that were agreed upon in ERC Stage 1 are considered for funding.

Source: Department of Treasury and Finance (Australia) (2005).

Program review processes

These processes can be distinguished by the fact that they are managed by the center of government—most usually the cabinet office or budget office (Kraan, 2005, p. 39). Typically, such processes involve the targeted review of certain programs, with respect to certain fundamental questions:

  • whether the program is justified for government intervention at all;

  • whether the program should be administered by a different level of government;

  • whether the program uses the appropriate policy instrument for its purposes; and

  • whether the program is sufficiently targeted to ensure that its benefits flow to the intended recipients.

Not surprisingly, program review represents a struggle between the center and line agencies. Portfolio ministers can attempt to bypass such review mechanisms. The UK and the Netherlands currently have formal program review mechanisms, while a particularly celebrated case in terms of reallocation was carried out as part of the fiscal consolidation efforts in Canada during the 1990s. In the Canadian case, the central focus was provided for this task by the cabinet office (known there as the Privy Council Office).19 In coalition governments, it is important to have broad representation on the program review’s overseeing body so that program review does not turn into a witch hunt of one party or another. It is fundamental that support occurs at the highest level, and, if necessary, the Minister of Finance’s own programs and those of the premier should be included early in the round.

Arguably, institutionalized program review is nothing more than what a modern, analytically informed budget office should be doing routinely and should have been doing for some time. Through its assessment of policy and the use of intervention analysis, the budget office acts as the agent of the center, making recommendations to government decision-makers on improving the allocation of resources. This role as a contester of agency programs and priorities is fundamental to the modern role of the budget office, seeing it move away from command and control; instead, it would restructure institutional arrangements to alter the incentives on key players, in particular ensuring that the traditional problems of information asymmetry associated with government activity are mitigated to some extent. The need to perform program review means that the budget office is forever moving closer to the role of an office of strategic management, with capability in many of the major policy lines of government (such as education) and a need for more deeply developed analytical skills, rather than the clerical skills of yesteryear. Quite simply, budget offices no longer do the degree of transaction-oriented work they used to do.

The logical frame of reference for program review can be adjusted as needs be. For instance, the affordability concerns with aspects of the welfare state in the context of changing demographics appear to have triggered very high-profile individual cases of program review on such issues as the affect of the aging population, or universal access to some welfare benefits, on the budget. These cases may directly spin out of the concern for strategic management at the center and the need to position the fiscal aggregates into a long-term sustainable position. By far the majority of OECD countries have carried out reviews in the past decade or so concerning the affects of aging.

Program review may result in the determination that the service should not be funded through general taxation, but instead should rely on user charges. A number of goals can be addressed through the appropriate use of user charging, with many having an allocative aspect. For example, the choice of who to charge may give rise to the question of whether those being charged (or their representatives) can effectively monitor standards, exert countervailing pressure on costs, and find alternatives to public provision. The additional inflow of income to the government and the possible reduction in production volume can provide some flexibility for the line agency or the government decision-makers to consider alternative uses for the funds. Guides exist as to the economic circumstances when user charging is beneficial, and the appropriate design associated with effective implementation.20

Ongoing policy evaluation and review

Program review may be supported by an ongoing cycle of program evaluation. Some jurisdictions have set goals of reviewing all government spending programs over a period of time and, consistent with that goal, have introduced mechanisms such as agency-based evaluations to ensure that critical information is flowing continually within agencies and—whenever possible—to the center about the effect of government interventions. However, this form of review is chiefly distinguished from program review by its requirement for a lower level of involvement from the center. Notwithstanding, the center sets the rules for this form of review and draws selectively on the information uncovered. A problem remains with information congestion, as the amount of data created by an evaluation each year of, say, 20 percent of all government programs can be overwhelming. Leverage can be gathered by the judicious allocation of budget office staff, or other independent analysts with similar orientations, to the evaluation teams involved in such work. Further advantage can be gained by requiring the transmission, and perhaps even publication, of all such reports to the center.

While different nomenclature may be used in different jurisdictions, the basic forms of evaluation can be about improvements in program design, improvements in process, or whether the program has contributed in a valid way to outcomes.

Role of performance information

The assessment as to whether policy and administration are addressing the priorities of government requires information about the performance of the policies. This is a topic worthy of separate consideration; suffice to say that information about the use of inputs, the production of outputs and the affects on outcomes is central to assessing whether interventions are achieving the desired goals in a cost-effective manner. This issue is extensively covered in Part Three of this book.

Summing-up: ten rules of thumb to assist with allocating resources in accordance with priorities

The foregoing discussion leads to the view that the way in which the budget is managed can lead to resource allocation which more closely reflects priorities. In other words, the design of budgetary mechanisms matters. Budgeting in accordance with priorities is not purely a matter of political will or some exogenous factor applied to decision-making. The preparedness to make decisions and the quality of those decisions can be influenced by the mechanisms at hand. In drawing the chapter to a conclusion, ten rules of thumb are recommended; of these, five relate to institutional design, and five relate to information requirements.

Institutional design factors

Item 1: A committee of the cabinet should be established to represent the collective interest in overseeing the budget. This committee should in turn be supported by a technical committee, which acts to ensure that only justifiable claims can be made on the agenda of the cabinet committee.

Item 2: The budget entity should be defined broadly to ensure that all public programs and public financial exposures are appropriately captured. There should be an absence of unconstrained fragmentation mechanisms.

Item 3: The performance and accountability system should be integrated with budget management concerns; agency performance in its use of resources is an important aspect of the assessment of the performance of government chief executives.

Item 4: An efficiency or productivity factor should be applied to the baselines of all agencies. In the absence of any empirical evidence to the contrary, the factor should be applied at the rate of 1.5 percent per year.

Item 5: A system of fundamental program reviews should be instituted, with the reports being communicated through the officers’ technical committee through to the cabinet committee. The trigger for such reviews should be agreed at a political level, and while can include criteria such as when a department seeks additional resources to produce the same level of outputs, they should also be cast wider to ensure that low-profile programs are also subjected to such reviews.

Information factors

Item 1: Intended outcomes are required during budget formulation and are also reported to the legislature, alongside the budget proposal. Proposals to change budgetary allocations are informed by consideration of the changes in intended outcomes. Ex-post reporting on these is required.

Item 2: The outputs to be produced in support of achieving the outcomes are also reported to the decision-makers at the agency, in the executive, and the legislature.

Item 3: Forward estimates (for at least three years) of the cost of unchanged policy are provided to the legislature on a classification basis which is well aligned to the political and bureaucratic distribution of authority.

Item 4: Performance reporting, in terms of the outputs produced, is subject to ex-post attestation through audit.

Item 5: All cost information should be full (accrual-based) cost.

Appendix: public budgeting and decentralization

Public budgeting involves the distribution of decision rights via the institutions available to government. As such, it presents as a case of the decentralized dilemma.

The knowledge of decision-makers is limited by both current scientific progress and also by their capacity to absorb, comprehend, and process the available knowledge (Zoghi, 2002, p. 6). This introduces trade-offs in the optimal point of allocation for decisions. Jensen and Meckling (1990, p. 4) point out that when knowledge is valuable in decision-making, there are benefits to collocating decision authority with the knowledge that is valuable to those decisions. There are few approaches to achieving this: moving knowledge to the individual with the decision rights; alternatively to shift the decision rights to the individual who possesses the knowledge; or a combination of the two.

The first approach leaves the organization faced with the costs of transporting knowledge. Radner (1992, p. 1393), drawing from the literature of computer science, defines four aspects of the processing of information that involve costs: observation of the data about the environment, capabilities and resources of the processors, the communication network that stores and moves the data, and the delay between the observation of the environment and the implementation of the decision. The magnitude of these costs will be influenced by the type of knowledge being transferred, current technology, and the organizational structure. As a result, the nature of these costs will be uniquely determined for any given organization, or sets of organizations, such as a government. By example, the movement of too many decisions to the center will overload it, as decision-makers are human with only a finite amount of time available to allocate to each decision—for instance, under the then Czechoslovakian state planning system, planners were reputed to have had only six minutes available to spend on planning each industry (Gorringe, 2001, p. 142). The second approach of shifting decisions to those with knowledge leaves the organization faced with agency costs that result from transferring decision rights—a problem of coordinating the decentralized decisions. These costs include the costs of developing and enforcing a set of contracts among agents with conflicting interests.

In making the determination of the best location for decision rights, and hence seeking to achieve allocative efficiency through these decision-makers, an organization is faced with trading off between agency costs and knowledge transfer costs. Figure 9.1 shows the intuition behind this.21 On the left-hand side, there is complete centralization—one single planner makes all decisions and all relevant knowledge is transferred to him or her. This results in zero agency costs—the principal and the agent are one. A movement toward the right side of the figure indicates decentralization, which increases agency costs and decreases the cost of transmitting information. When all decision rights have been transferred to the individuals in possession of specific knowledge, the organization is completely decentralized, as indicated by the vertical line on the far right of Figure 9.1. At this point, agency costs are at their highest, and information costs at their lowest. Optimally, a location of decision rights that minimizes the sum of these two costs would be chosen.

Figure 9.1.Trade-offs between inconsistent objectives and poor information

Source: Adapted from Jensen and Meckling (1990).

The relevance for reallocation of resources between competing priorities is clear. The more a resourcing decision relates to the choice over the type and purpose of intervention, and thus influenced by the political preferences, the higher the costs of moving critical information to the decision-makers will be. Conversely, the more a decision relates to dealing with local conditions, the better placed the local agent is to respond to this information, but the more there is a need to make government policy clear so that the action can be consistent with that policy. As Schick has suggested, the center requires strategic capacity, and the organizations of government require operational information (Schick, 1997, p. 17).

Different administrative structures deal differently with this trade-off in the allocation of discretion. Some countries, such as Sweden, New Zealand, the Netherlands, and the UK, are characterized by strong devolution, whereas some others such as Italy, France, and Germany are characterized both by more legalistic traditions and more central control (Kraan, 2005, p. 25). The introduction of the Next Step agencies in the UK could be considered a case where the value accorded to local operating information was given precedence over the value given to closeness to political control. While there may be differences between jurisdictions in terms of the relative value put on decentralization, it is also clear that there are differences within jurisdictions as well. For instance, the relatively new budget control or appropriation taxonomies in New Zealand and the UK, which have different controls for administrative costs, transfers, and capital expenditures, demonstrate differentiated discretion according to the underlying economic nature of the expenditures.


The author would like to acknowledge comments on this chapter from colleagues including Holger van Eden, Marc Robinson, and Theo Thomas, and from Professor Philip Joyce. Any errors remain the author’s.

Developed by Musgrave in the classic work The Theory of Public Finance (Musgrave and Musgrave, 1959).

Literature exists which assesses governments that are too small and thus unable to deliver the merit and public goods necessary for efficient operation of the economy. On the other hand, burgeoning government relies on increasing rates of taxation, which encompass increasing marginal rates of deadweight loss. A consequence of this is that not only must the positive welfare consequences of public intervention increase to offset the growing deadweight costs of taxation, but also that variability in tax rates will, ceteris paribus, lead to higher deadweight cost than raising the same revenue through a constant rate of taxation over the same period. This, accordingly, relates aggregate management of expenditure very directly to broader allocative efficiency.

This is the focus of writing associated with Von Hagen, among others. The idea that a common pool problem is intrinsically rooted in the typical public budget process can be traced back to Weingast et al. (1981).

Whereas 60 percent of US states used no ceilings in 1970, this number had fallen to 3 percent by 2000 (Burns and Lee, 2004, p. 3).

Political agency problems can induce distortions in resource allocation; accordingly, the degree of the political agency problems is likely related to the observed rate of distortion. For instance, as Alesina and Tabellini recently argued, the pro-cyclical nature of fiscal policy in developing countries is evidence of voters—poorly informed and victims of an agency problem which will not allow rents to be pushed to zero—seeking a second-best solution to constraining politicians through pro-cyclical fiscal policy. This has the effect of starving the Leviathan, but with associated welfare losses (Alesina and Tabellini, 2005, p. 3).

The argument is that the more responsive movement of resources at the line management level is, the less pressure is placed on spending caps, the more likely fiscal consolidation exercises that rely on reductions in spending are to be successful. There has been a lack of empirical research on this question.

Some jurisdictions will refer to this as an output price, as there is an arm’s-length contract between the funder and the supplier, even where the supplier is a government agency.

For instance, see Davis et al. (2003).

One could also argue that fiscal rules that have variously set limits on expenditures, deficit levels, taxes, or debt, have also contributed substantially to the incentives for reallocation. Without such constraints, the desire to meet new priorities can be added on to an expenditure base, rather than involve a reallocation from lower-priority areas to higher-priority areas. But fiscal rules are the subject of other analyses.

In some cases, even where the overarching control is an appropriation control, the executive retains the right to vary downward spending from these appropriation limits. In other cases, the appropriation control may be seen as a floor and a ceiling.

Fiscal rules should be anchored in a medium-term perspective, be defined over the cycle, and be unambiguous in their coverage. Fiscal rules should distinguish the balance sheet effects of different forms of expenditure, so as to counteract the bias to cut capital spending first. Independent fiscal boards are a relatively new innovation, which has been touted as an aggregate control protector in certain circumstances (see Bonato et al., 2005, p. 4).

While neither of these conditions may look profound, it is worth noting that Fund staff recently argued that the fiscal framework rule applying in Sweden had been interpreted as a yearly structural balance and as an average balance—nominal or structural—over the cycle (Bonato et al., 2005, p. 12).

The Fund has identified two types of fiscal agencies: Independent Fiscal Agencies (IFAs), which mimic independent central banks on the fiscal side, and which could be mandated to enforce fiscal balance targets; and fiscal councils, which would not receive any authority over policy, but rather would provide independent analysis and assessment of fiscal developments (IMF, 2007, p. 2). The rationale for the agencies depends on the extent to which they could be expected to diminish the risks of an injudicious use of discretion. The Fund points out that to date, there is no IFA in operation in any country.

This behavior is called loss aversion, in which the marginal utility from a gain is strictly less than the marginal utility from an equivalent loss. Reallocations will not naturally occur as decision makers (spenders) will fight extremely hard to protect themselves from the loss associated with reductions in resources (Kahnemann and Tversky, 1979).

Introduced first into Sweden and Australia, the idea of the efficiency dividend was to take from baselines that amount of spending which otherwise would have been converted into increased outputs or higher-than-necessary input prices.

In this context, it is perhaps worth recalling that the internationally-accepted functional definitions of government are primarily to assist the comparative analysis of government. Accordingly, the functional depiction has clear analytical uses, but perhaps less clear decision-making use. Rather than being an instrument for allocation, such functional data represent the allocations made.

A case in point would be a fiscal scoring system that focused solely on a measure such as the Public Sector Borrowing Requirement.

In many cases, this may involve substituting one output with another, but on some occasions it may principally involve substituting regulation or law for an output.

The use of performance information in attest auditing requires there to have been an accounting standard issued by the jurisdiction’s accounting standards setter.

The Privy Council Office, the Ministry of Finance and the Treasury Board operated together to pursue this change agenda; however, the Privy Council Office, which reports directly to the Prime Minister, was the lead.

See, for example, New Zealand Treasury (2002) and OECD (1998).

The figure is used with the permission of Professor Jensen.


    AlchianA. and H.Demsetz1972Production, Information Costs and Economic Organization,American Economic Review Vol. 62(5) pp. 77795.

    AlesinaA. and G.Tabellini2005Why Do Politicians Delegate? NBERWorking Paper WP11531 (Cambridge, Mass.: NBER).

    BarrN.1998The Economics of the Welfare State (Oxford: Oxford University Press).

    BjornskovC.A.Dreher and J.Fischer2005The Bigger the Better? Evidence on the effect of Government Size on Life Satisfaction Around the World,Institute of Economic Research Economic Working Paper Series (Zurich: Swiss Federal Institute of Technology).

    BonatoL.E.Tsounta and F.Balassone2005Sweden: Article IV Selected Issues (Washington: International Monetary Fund).

    BostonJ.J.MartinJ.Pallot and P.Walsh1996Public Management. The New Zealand Model (Oxford: Oxford University Press).

    BrumbyJ.2000Budgetary Devices That DeliverOxford Policy Institute Policy Brief No. 3 (Oxford: Oxford Policy Institute).

    BurnsJ.C. and R.D.Lee2004The Ups and Downs of State Budget Process Reform: Experience of Three Decades,Public Budgeting and FinanceFall pp. 119.

    CaidenN. and A.Wildavsky1974Planning and Budgeting in Poor Countries (New York: Wiley).

    CurristineT.2005Performance Information in the Budget Process: Results of OECD 2005 QuestionnaireGOV/PGC/SBO(2005)6 (Paris: OECD).

    DavisJ.R.Ossowski and A.Fedelino2003Fiscal Policy Formulation and Implementation in Oil Producing Countries (Washington: International Monetary Fund).

    Department of Treasury and Finance (Australia) 2005The Integrated Management Cycle (Melbourne: Treasury of Victoria).

    DueJ.F. and A.F.Friedlaender1973Government Finance. Economics of the Public Sector (Irwin: Homewood).

    GorringeP.2001Economics for Policy (Wellington: Institute of Policy Studies).

    IDA and IMF2005Update on the Assessments and Implementation of Action Plans to Strengthen Capacity of HIPCs to Track Poverty Reducing Public Spending (Washington: International Monetary Fund).

    IMF2007Promoting Fiscal Discipline—Is There a Role for Fiscal Agencies?forthcoming (Washington: International Monetary Fund).

    JensenM.C. and W.H.Meckling1976Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure (Cambridge, Mass.: Harvard University Press).

    JensenM.C. and W.H.Meckling1990Specific and General Knowledge and Organizational Structure,” in Contract Economicsed. L.Werin and H.Wijkander(Oxford: Blackwell) pp. 25174. Available at SSRN: < <>.

    JoumardI.P. M.KongsrudY.S.Nam and R.Price2003/2Enhancing the Cost Effectiveness of Public Spending: Experience in OECD Countries,OECD Economic StudiesIssue 2No. 37 pp. 10961 (Paris: OECD).

    KahnemanD. and A.Tversky1979Prospect Theory: An Analysis of Decision under Risk,Econometrica pp. 26391.

    KraanD-J.2005Reallocation: The Role of Budgetary Institutions (Paris: OECD).

    Le HouerouP. and R.Taliercio2002Medium Term Expenditure Frameworks: From Concept to Practice. Preliminary Lessons from AfricaAfrica Region Working Paper 28 (Washington: World Bank).

    LorentzenP.J.McMillan and R.Wacziarg2005Death and DevelopmentCDDRL Working Paper No. 48 (Stanford: Stanford University Press).

    MusgraveR.A. and P.B.Musgrave1959The Theory of Public Finance (New York: McGraw-Hill).

    New Zealand Treasury2002Guidelines for Setting Charges in the Public Sector (Wellington: New Zealand Treasury). Available at: <>.

    OECD1998Best Practice Guidelines for User Charging for Government Services (Paris: OECD). Available at: <>.

    OuchiW.1980Bureaucracy, Markets and Clans,Administrative Science Quarterly Vol. 25 pp. 12945.

    RadnerR.1992Hierarchy: The Economics of Managing,Journal of Economic Literature Vol. 30 pp. 1382415.

    SchellingT.1978Micromotives and Macrobehavior (New York: Norton).

    SchickA.1997The Changing Role of the Central Budget OfficeOCDE/GD/(97)109 (Paris: OECD).

    SchickA.1999A Contemporary Approach to Public Expenditure Management (Washington: World Bank Institute).

    WannaJ.L.Jensen and Vries2003Controlling Public Expenditure (Northampton: Edward Elgar).

    WannaJ.J.Kelly and J.Forster2000Managing Public Expenditure in Australia (Crows Nest: Allen & Unwin).

    WeingastB.K.Shepsle and C.Johnsen1981The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics/’Journal of Political Economy Vol. 89(4) pp. 64264.

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

    ZoghiC.2002The Distribution of Decision Rights within the Workplace: Evidence from CanadianAustralian and UK Establishments BLS Working Paper No. 363 (Washington: Bureau of Labor Statistics).

    Other Resources Citing This Publication