IV. Public Expenditure Productivity

Ke-young Chu, and Richard Hemming
Published Date:
September 1991
  • ShareShare
Show Summary Details

G.A. Mackenzie

How can productive expenditure be distinguished from unproductive expenditure? What are the relevant concepts of efficiency?

To what extent can efficiency be increased by transferring public sector activities to the private sector?

What criteria should be used to judge whether the balance of expenditure between and within public programs is appropriate?

The motivation for restructuring the expenditure side of the budget is often found in macroeconomic imbalances that are judged unsustainable. In the absence of opportunities to secure increases in revenue, an expenditure adjustment is unavoidable. However, even when a clear case can be made for a cut in public expenditure, formidable difficulties remain. While it is uncontroversial to recommend the elimination of the least productive programs, reaching agreement on which programs are the least productive is normally quite difficult. The purpose of this note is to try and identify the characteristics and categories of expenditure that should be the priority targets in the event of either a need for expenditure cuts or a desire to accommodate new expenditure demands within the context of fixed resource availability.

Defining Unproductive Expenditure

The concept of economic inefficiency in public expenditure is central to any definition of unproductive expenditure. In what follows, attention is focused on four aspects of inefficiency. The first type of inefficiency arises from an inappropriate assignment of activities to the public sector as opposed to the private sector. Two other types of inefficiency arise from specific characteristics of public provision. Allocative inefficiency manifests itself as a mix of public programs that does not best meet the objectives that justify government intervention. Productive inefficiency describes a situation where goods and services are not provided by the public sector at minimum cost. It should be noted that allocative inefficiency is sometimes termed product mix inefficiency, in recognition of the fact that cost minimization involves decisions about the allocation of inputs. There is also a macroeconomic dimension to unproductive expenditure, related to the efficiency with which public expenditure serves broad macroeconomic objectives.

Public versus private provision

The note on Public Expenditure and Resource Allocation describes the contribution of welfare theory to the assessment of the appropriate role of government in a modern mixed economy. In essence, there is a clear case for intervention in the presence of market failure or when the distributional consequences of the market are considered unfair. Otherwise, competitive markets provide the incentive to seek economic efficiency, that is both allocative and productive efficiency. The product market guides prices and output—producers who do not sell what the public wants at a price they are willing to pay will not make a profit. The capital market constrains costs—producers who do not make adequate profit will go bankrupt or be taken over. If the distributional outcome is acceptable, there is also social efficiency.

To the extent that government intervention extends beyond those activities characterized by market failure or unacceptable distributional outcomes, the additional public expenditure that results could be regarded as unproductive. Private provision of goods and services that do not exhibit these characteristics will, in principle, be at least as efficient as public provision. Notwithstanding its theoretical attraction, the welfare-based minimalist approach to government intervention—which places the burden of proof on advocates of more extensive intervention—has to be applied with caution.

Clearly, government intervention varies widely in degree. In part, this reflects differences in the extent of market failure and the acceptability of distributional outcomes. In addition, political, social, and economic factors, largely unrelated to a welfare-based approach to intervention, are important. Moreover, not all of the arguments used to justify more extensive intervention are necessarily bad ones. For example, if governments choose to finance programs that benefit particular groups to preserve political stability, to support a declining industry so as to prevent widespread unemployment in regions dependent on that industry, and to maintain control of strategically important sectors (e.g., petroleum and air transportation), this is not necessarily misguided. But rather than debating the merits of each and every instance of intervention, the search for unproductive expenditure is better directed toward those activities where public provision is obviously wasteful and/or private provision is likely to be clearly superior. However, it is impossible to lose sight of the fact that every public program benefits some interest group, and proposals to cut even the most wasteful programs will meet vigorous opposition from the current beneficiaries.

The note on Privatization discusses the relative merits of private and public provision in more detail, and in particular the scope for devolving public sector activities—through asset sales, joint ventures, contracting out, and franchising—to the private sector. In summary, increased exposure to competition is likely to yield the largest gains in economic efficiency, and therefore the link between privatization and competition generally determines the strength of the case for privatization. Where a private market already exists, there is probably a case for privatization, although not necessarily an associated prospect of large efficiency increases. However, where the private sector is excluded from a particular market, and there is no reason to believe that public provision offers compelling political, social, or economic advantages, privatization accompanied by liberalization should increase efficiency substantially. There may also be scope, while retaining certain activities predominantly in the public sector, for allowing private sector involvement in some of the less central aspects of these activities.

Allocative inefficiency

It is difficult to comment on the extent of allocative inefficiency resulting from an inappropriate mix of programs in the absence of information about the social welfare function (SWF). A SWF can be viewed as a representation of social preferences across the full range of public sector goods and services, and other public sector activities such as regulation. Society would prefer that mix of expenditure which maximizes social welfare as reflected in its SWF. It should also be the objective of government policy to achieve such a mix. However, the SWF the government uses in seeking to maximize social welfare may differ from that of society. This will arise when there are preference revelation problems, and society has no effective and/or efficient means of indicating its SWF to the government. Despite the best efforts of the government to reflect social preferences in its decision-making, there will be a loss of allocative efficiency in such circumstances.

The government may also override social preferences. On the basis of paternalism, the government could decide that individuals make inappropriate choices as far as merit goods are concerned—individuals choose to invest in too little education and consume too much alcohol and tobacco, for example—and the government will therefore seek to alter individual behavior. Whether there is a loss of allocative efficiency with a paternal government is the subject of debate. Much rests on the possibility of there being other justifications for intervention that is undertaken for ostensibly paternal reasons. For instance, persuading people to invest more in education and to limit their consumption of alcohol and tobacco can be defended by reference to externalities, in which case government intervention is intended to increase allocative efficiency. Only to the extent that paternal intervention cannot be justified by reference to market failure—and with some imagination most public programs can be so defended—could it reasonably be claimed that there is an associated loss of allocative efficiency.

The most serious allocative inefficiency reflects the failure of the government to provide a mix of goods and services consistent with any reasonable SWF (see below for further discussion). This would arise where other objectives compromise the ability to maximize social welfare. These objectives typically reflect the personal preferences of politicians, public sector managers and organized labor, who take advantage of their positions and the administrative complexity of public agencies to pursue activities that serve their own interests. The resulting loss of allocative efficiency is often said to reflect state or government failure.

Productive inefficiency

The efficiency of production has both a technical aspect—Is output maximized with given inputs?—and an allocative aspect—Does the input mix reflect opportunity costs? Numerous factors can contribute to high production costs in the public sector. Some of these need not in the strictest sense point to productive inefficiency, since they are a function of the objectives of government. For example, overmanning in the public sector may reflect a deliberate decision about employment generation and does not necessarily imply productive inefficiency. There is, however, a general presumption that the public sector does not use the least-cost production methods.

Productive inefficiency is often explained in terms of technical considerations (inadequate investment appraisal, low levels of research and development, etc.). However, while these can play a role, they are symptomatic of more general shortcomings in the internal organization of the public sector. These shortcomings give rise to what is often referred to as X-inefficiency. The root causes of X-inefficiency overlap with those that give rise to allocative inefficiency. These include the political and bureaucratic considerations that generate ever-increasing budgets; salary and wage-setting mechanisms that neither reward efficiency nor penalize inefficiency; and a fiscal illusion that leads the public to accept both increases in the number of public programs and the rising costs of running them. It will be recalled that these are essentially the basis of the bad arguments used to explain rapidly growing public expenditure that are discussed in the note on Public Expenditure and Resource Allocation. They provide the scope for inappropriate political manipulation of public agencies and lead to ineffective monitoring and control, poorly motivated management, and inflexible public sector labor markets. The resulting loss in productive efficiency is another aspect of state failure. That most other sources of high costs in the public sector would probably be eliminated if the organizational structure was improved is reflected in the fact that productive inefficiency is often equated with X-inefficiency.

Consistency between macroeconomic and microeconomic objectives

The note on Public Expenditure and Sustainable Fiscal Policy describes how the appropriate level of public expenditure can be judged by reference to its consistency with broad stabilization objectives. The note on Public Expenditure, Stabilization, and Structural Adjustment emphasizes the role of the composition of public expenditure in raising the sustainable growth rate. Expenditure switching that allows either (or some combination of) higher growth, lower inflation, and a greater debt servicing capacity without conceding another objective can be seen to imply an unambiguous reduction in unproductive expenditure. However, once objectives have to be traded off against one another, that is, growth vs. inflation and/or balance payments difficulties, normative judgments come into play in the same way as when the microeconomic aspects of efficiency are considered.

Matters become even more complicated when the macroeconomic and microeconomic aspects of efficiency are combined. Consider, for example, a SWF function that has growth, inflation, the balance of payments, poverty alleviation, and environmental protection as arguments. Again, if through expenditure switching one can do a little better with respect to one objective without conceding on others, then some unproductive expenditure can be said to be eliminated. But with an expanded SWF the possibility of conflict is increased, and the need to trade off objectives is a more likely occurrence. For example, growth vs. poverty alleviation and/or environmental protection may now be an issue. Normative judgments will have to play an even larger role. In general, the more carefully one attempts to characterize the SWF, the greater the possibility of a conflict between objectives and the larger will be the element of judgment in deciding whether an expenditure program is unproductive or not.

Widening the set of objectives also has consequences for the notion of sustainability. When discussing sustainable fiscal policy and sustainable growth, these have been judged so far by reference to stabilization objectives. However, it is clear that stability is not enough. If the pursuit of distributional objectives provides a compelling case for government intervention, sustainability should also be judged by reference to the consequences of policies and outcomes for equity. Indeed, when referring to sustainable growth, it is generally taken for granted that this refers to stable and equitable growth. It is also increasingly being argued that the rate of depletion of environmental resources be reflected in the concept of sustainability.

Identifying Unproductive Expenditure

Notwithstanding the formidable difficulties that are, in principle, involved in identifying unproductive expenditure, the aim of this section is to try to provide some practical guidance as to how such an assessment might, in certain instances, be made. However, it should first be noted that even if unproductive expenditure can be isolated, attempts to reduce it will tend to meet political resistance. As indicated earlier, such expenditure often serves the interests of political and other pressure groups who will be vocal in defense of their favored programs and projects. Across-the-board cuts, or those that concentrate on programs for which the beneficiaries are less powerful or well-organized, meet least resistance. It is often the case, therefore, that the programs which bear a disproportionately large share of cuts, especially in developing countries, are infrastructure investment projects and other programs with more diffuse benefits but relatively high economic and social rates of return.

Based upon the earlier discussion, the key questions to which the policy analyst seeks answers are: (i) Should a particular activity be undertaken in the public sector? (ii) Is the mix of public sector activities appropriate? and (iii) Are these activities being undertaken at minimum cost? As pointed out earlier, the considerations that bear upon the question as to the scope for efficiency gains from increased private sector involvement in activities currently undertaken in the public sector is discussed in the note on Privatization. Attention is therefore focused on allocative and productive inefficiency in the public sector.

The most obvious sources of allocative inefficiency are prestige projects that serve no apparent economic or social purpose. Examples are numerous: second airports where an existing facility is underutilized, multi-lane highways with capacity that far exceeds projected traffic flows; and public buildings which provide facilities that are never going to be in demand. The worst excesses are reflected in so-called white elephants, usually major civil engineering projects that serve to increase the prestige of those who order their undertaking. But even if unambiguously wasteful projects can be identified, reducing their number is quite another thing. The fact that ostentatious expenditure is often found to be largest in some countries that can least afford it, and where the opportunity cost in terms of forgone expenditure on essential infrastructure, social services, and anti-poverty programs is therefore higher, points to the entrenched nature of such expenditure.

Once white elephants and other obviously wasteful projects have been identified, further improvements in allocative efficiency require that expenditures be prioritized. This is difficult. It is necessary to take into account: the relevant characteristics of the SWF; the complementary nature of different activities such as the need for operations and maintenance expenditure to maximize the return from a particular investment; and substitution possibilities, given that any particular sector objective can be met in a variety of ways. It is likely that only sector experts will be able to rank expenditure programs in their particular sector.

Assuming that a reasonable ranking can be provided, the next step involves choosing between expenditure programs in different sectors. In principle, one is attempting to restructure the mix of expenditure until either cuts are equally distasteful in all sectors or a reallocation of expenditure can yield no further increases in social welfare. In practice, one ends up looking for proximate indicators of misallocation. Again, sector experts may be able to provide the necessary indicators. For example, low literacy rates combined with low mortality rates might point to too little being spent on education and too much on health care; an examination of teacher/pupil and doctor/patient ratios could provide support for this conclusion. Empty roads and crowded ports could indicate a similar misallocation of resources. Numerous other examples of expenditure imbalances can be constructed. Such indicators are frequently used in World Bank public expenditure reviews as the basis of arguments for a reorientation of spending to increase overall expenditure productivity.

Pricing policy may also provide an indication of misallocation. For example, theoretical arguments suggest that prices should be set equal to marginal cost for efficiency and therefore only public goods, which are nonrival in consumption, should be provided free and financed out of general revenue. To the extent that other goods are provided free—or, indeed, even if they receive any unjustifiable subsidy—the resulting overconsumption will imply allocative inefficiency. It would be more efficient to charge an appropriate price for such goods and services and either reduce distorting taxes and borrowing, or redirect resources to more productive expenditure programs. For a further discussion of charging for public sector goods and services, see the note on Pricing and Cost Recovery.

In the case of productive inefficiency, sector experts may also be able to provide guidance as to whether goods and services can be delivered at lower cost, say by using more efficient technology or combining capital and labor in different ways. Expenditure imbalances can also signal the presence of productive inefficiency. To this end, sector experts may be able to indicate whether there is an unnecessarily high doctor/nurse ratio in public health clinics, whether there is a high teacher/pupil ratio when basic teaching supplies are not available, or some other imbalance exists.

The above discussion has focused on exhaustive expenditure, reflecting the public sector’s claim on the economy’s real resources. In the case of subsidies and transfers, the distinction between allocative and productive inefficiency is blurred. For instance, consider a general food subsidy or a universal income transfer. While such programs are mainly intended to benefit the poor, they help everybody. Payments to the nonpoor give rise to both allocative inefficiency—it would be better to spend this money in other ways—and productive inefficiency—in that the poor could be aided at lower cost. If the programs do not in fact reach all the poor, this exacerbates inefficiency. Generally speaking, any measure designed to help a specific target group that also benefits others, except insofar as the spillover benefits can be justified by reference to this being necessary to ensure that the poor are indeed helped, should be judged inefficient. These issues are discussed further in the note on Poverty and Social Security.

Attempting to distinguish between good and bad expenditures by reference to their relative productivity—even if it is difficult to do—illustrates rather clearly that the normal distinction between current and capital expenditure is not especially helpful in this regard. There is just as likely to be wasteful capital expenditure as wasteful current expenditure. The fact that expenditure adjustment has been borne primarily by capital spending is not necessarily undesirable; there is only a problem if the affected projects are relatively productive. Similarly, calls to contain current expenditure may be counterproductive if it is the more valuable programs that are affected. An optimal expenditure structure will comprise a mixture of capital and current spending that reflects, among other things, the government’s objectives (i.e., its SWF), the relative productivity of different public expenditures, and the expenditure activities of the private sector. Reasonable objectives can be met by an expenditure structure that is justifiably biased toward current, but highly productive, expenditure.

Unproductive expenditure is an elusive concept. It is not easily defined, and the definition offered here, while deriving from some notion of the success with which the objectives of government intervention are met, does not readily lend itself to direct practical application. Relying instead on proximate indicators of inefficiency requires detailed information, and specialized skills to interpret it. The most obvious source of both information and the necessary sectoral expertise is the World Bank.


    Heald David Public Expenditure (Oxford: Martin Robertson1983).

    TaitAlan A. and Peter S. Heller International Comparisons of Government Expenditure, Occasional Paper No. 10 (Washington: International Monetary Fund1982).

    • Search Google Scholar
    • Export Citation

    Other Resources Citing This Publication