III. Public Expenditure and Resource Allocation
- Ke-young Chu, and Richard Hemming
- Published Date:
- September 1991
What economic arguments can be used to justify government intervention? Which activities should be undertaken by the public sector?
How can the government influence distributional outcomes?
Is the observed growth of public expenditure with income inevitable? What arguments have been offered to explain this growth?
Government intervention in the economy is motivated by a wide range of economic, social, and political objectives. The fact that the share of public expenditure in GDP, and the composition of expenditure, varies so much across countries reflects how much importance is or has in the past been attached to different objectives. This note focuses primarily on the economic justification for intervention, and discusses in particular the imperfections that typically characterize private markets and which give rise to a compensatory role for government. While many public programs can be defended by reference to market failure, others have to be justified using different arguments. In this connection, distributional objectives are important. Conclusions are also drawn about the link between microeconomic objectives and the size of the public sector.
The Optimality of Competitive Markets
The starting point for any discussion of government intervention is the voluntary exchange model of a competitive economy. According to this model, a competitive economy will result in an allocation of resources such that (i) inputs cannot be reallocated to yield a higher output of one good without reducing the output of another good, and (ii) total output cannot be reallocated to generate a higher level of welfare for one consumer without reducing the welfare of another consumer. This is referred to as Pareto optimality. A Pareto optimum is defined in terms of a set of marginal conditions, requiring equality between marginal rates of transformation in production and marginal rates of substitution in consumption. A corollary of these conditions is the marginal cost pricing rule (see the note on Pricing and Cost Recovery). However, these marginal conditions only support a Pareto optimal competitive equilibrium if a number of underlying assumptions are satisfied.
There are many different ways of characterizing these assumptions. For simplicity, they can be conveniently summarized as the following: (i) markets must exist for all goods—if market prices cannot be used to control supply and demand, or if economic agents do not share the same information, some goods and services may not be provided at all; (ii) there must be no externalities—prices must reflect all costs and benefits associated with production and consumption; and (iii) there must be decreasing returns to scale—unbounded scale economies lead to a noncompetitive economy dominated by a few large firms. To the extent that these assumptions are not satisfied, then there is market failure. Moreover, even if the marginal conditions do result in a Pareto optimal competitive equilibrium, this may not be socially optimal. There are many Pareto optima, each associated with a different distribution of factor ownership. For social optimality it must also be assumed that (iv) society judges the initial distribution of factor ownership, and therefore the final distribution of income and welfare associated with a Pareto optimum, to be acceptable.
The market failure that results when the assumptions of the competitive model do not hold takes a variety of forms. In its strictest sense, market failure results from the violation of assumptions (i), (ii), and (iii) above. The most common issues that then arise relate to the provision of public goods (for which market prices are inappropriate) and collective goods (which are characterized by externalities), the emergence of natural monopoly (when there are continuous economies of scale) and the need for social insurance (arising from information asymmetries). The inability of the free market to generate a socially optimal distribution of income and welfare, which violates assumption (iv), can also be regarded as an example of market failure. This issue, however, will be discussed separately.
Public goods are characterized by nonrivalry in consumption. Nonrivalry in consumption refers to the idea that the benefits of a good can be enjoyed by more than one person simultaneously and that the cost of accommodating additional consumers is zero. Nonexcludability is also claimed to be a characteristic of public goods. Nonexcludability exists when a person can enjoy the benefits of a good whether or not there has been payment for its use. In other words, there are no clearly established property rights which can be assigned to particular individuals. The possibility of exclusion is to some extent a technical question; with rapidly changing technologies, there are increasing possibilities for limiting or charging for the consumption of goods heretofore considered as public goods. Nevertheless, it is undesirable on efficiency grounds to exclude access to nonrival goods, given that marginal cost is zero. It is therefore nonrivalry that primarily defines public goods. Generally, the private sector would tend not to provide public goods since market prices cannot be used to ration them. Therefore, if these goods are worth producing, it is up to the public sector to ensure their provision either directly, or by encouraging or contracting for private provision (see the note on Privatization). In general, public goods are produced by the public sector. However, charging for public goods represents a problem for the public sector since individuals who cannot be deprived of consuming a public good have an incentive to minimize their potential contribution to financing its provision. The free-rider problem implies that public goods would be underprovided if the government tried to link charges to some indicator of consumer preference, because individuals would not reveal their true preferences. Such goods therefore tend to be financed from the budget. While the theory of public goods has a powerful conceptual appeal, when one looks to the real world there are not many examples of pure public goods. Defense and air pollution abatement clearly fit the technical characteristics of public goods (for further discussion see the notes on Military Expenditure and Public Expenditure and the Environment). Flood control, some public health services, weather forecasting, and lighthouses are also often-quoted examples.
Collective goods and externalities
While there may be relatively few pure public goods, there are many goods which have partial public good elements to them. These are goods which are largely rival in consumption and for which exclusion is possible, but where some of the benefits or costs associated with their provision accrue to or are borne by other individuals or by society as well. Goods exhibiting externalities in consumption or production are often referred to as collective or mixed goods. The existence of externalities forces a wedge between market prices and social valuation. Thus, although the private sector can and does provide goods and services characterized by externalities, the quantity provided will not correspond to the socially desirable level. The individual who demands education is not likely to consider or be willing to pay for the additional benefit derived by others, and private markets offer no mechanism whereby one individual’s preferences for consumption by another individual can be readily expressed. In such situations, social benefits exceed the private benefits that a buyer would appropriate for himself; private markets would tend to supply too little compared with the amount that would emerge if all individual preferences could be fully captured by private markets. As a consequence, government action may be needed to help society approach the socially optimal level of output. Where there are external benefits—as in the case of education, vaccination, waste disposal, and certain forms of transportation, for example—the government can subsidize consumption and/or production. Where there are external costs—as in the case of air pollution and noise, for example—the government can levy compensating taxes. However, there are alternative strategies. For instance, the government could define the relevant property rights, thus establishing that polluters have to compensate the public who “own” clean air. But as with public goods, the government produces many collective goods directly. It also uses its regulatory powers to contain activities that give rise to external costs.
Natural monopoly describes a situation where one firm dominates an industry subject to continuously decreasing average costs of production. The incumbent firm can always expand capacity at a cost lower than that at which a new firm can create the same additional capacity. Efficient pricing requires that prices be set equal to marginal cost. However, under such circumstances, setting prices equal to marginal cost will result in losses, since marginal cost is necessarily below average cost. Therefore, if society is to benefit from the lower price and higher output than a profit-maximizing private monopolist would choose, either the private producer must be paid a subsidy, or the public sector must take over the industry and bear its losses directly. Examples of decreasing-cost industries include gas, electricity and water utilities, telecommunications, and mass transportation, all of which are characterized by extensive networks (gas pipelines, electricity grids, railway lines, etc.) that would be costly and inefficient for competing suppliers to replicate. Public production in these areas is commonplace.
Most types of insurance can be purchased on the private market; these markets are, however, very imperfect. The fact that the insured typically have more relevant information than the insurer gives rise to moral hazard and adverse selection problems. Moral hazard arises when the insured can affect the liabilities of the insurer without the latter knowing—with full medical cover, the insured will tend to demand too much medical care because the marginal cost is zero, the health care system typically has an incentive to provide it, and the insurance companies cannot monitor need. Insurers respond by using coinsurance and deductions. Adverse selection arises where low-risk individuals refuse to buy insurance at an actuarially fair price. Only high-risk individuals buy insurance, which forces up the price. In the end, only the worst risks would require insurance, and insurers would be reluctant to remain in the market. If they did so, they may attempt to discriminate against those in the greatest need. It is because complete private markets are unlikely to exist that government provision of unemployment insurance, health care for low-income families, and retirement pensions is often defended by reference to the benefits of risk pooling.
In these cases, however, merit good arguments are also used to justify intervention. Even if private insurance were generally available, it is argued that individuals would often not act in their own self-interest and purchase it. Government intervention is therefore justified on paternalistic grounds. It should be noted, however, that defending government intervention by reference to a paternalism motive involves the subordination of individual preferences. For this reason, arguments for public expenditure based on paternalism are uneasy ones for economists who otherwise argue for the efficiency of the market mechanism.
Toward Allocative Efficiency
Public goods, externalities, natural monopoly, and the need for social insurance provide government with an allocative role. However, the way in which the government influences the economy in pursuit of social efficiency is important. For example, the theory of the second best implies that when the conditions for optimality do not generally pertain, there should be no presumption that an attempt to correct for any one of many deviations from optimality will be welfare improving. In principle, the optimum associated with any compensated market failure is characterized by a completely new set of marginal conditions that can depart significantly from those that define a first-best optimum. Moreover, to move the economy to this new optimum, the government must use taxes, subsidies, regulatory controls, etc., that are themselves distortionary. Government policy has therefore to trade off the inefficiency of market failure against the inefficiency associated with the compensating policies, and the costs of administering them. The design of such policies is the principal challenge facing policymakers having to judge the appropriate role of government in influencing resource allocation.
As indicated above, there exists a Pareto optimum associated with each initial distribution of factor ownership—the resulting distribution of income and welfare reflects the initial distribution of the capital stock and ability. If this distribution is unacceptable, society will sanction redistributive measures. While there are private agencies that can effect some redistribution—charitable organizations, for example—without the government’s power to tax and transfer, the amount of coordinated redistribution that can be achieved is modest. Nevertheless, complicated issues still arise. Progressive taxation and cash transfers are the most direct redistributive instruments used by governments. But other policies are either intentionally redistributive or have redistributional implications. The public provision of health, education, housing, and other social services lies at the core of redistribution policy—there are, however, problems in assessing the amount of redistribution that takes place through the provision of public services. There is also a question as to whether the government should provide these goods and services directly, or provide income supplements and let people purchase the amounts they want in a market, if it exists or can be created. These issues are taken up at greater length in the notes on the Distributional Impact of Public Expenditure and Poverty and Social Security. In addition to its explicitly redistributive activities, governments affect the distribution of income and welfare in other ways. In particular, regulatory activities such as consumer protection, anti-monopoly legislation, and safety laws—many of which are justified by reference to problems associated with collecting, disseminating, and interpreting information, difficulties in organizing collective action, and the costs of establishing and enforcing minimum standards—have distributional implications. As in the case of policies directed toward improving resource allocation, measures aimed at addressing distributional objectives have an efficiency cost. There is therefore a trade-off between equity and efficiency that implies a limit to the amount of redistribution which can be undertaken.
Public Expenditure Growth
The preceding discussion provides a basis for government intervention by reference to its beneficial impact on resource allocation and equity. Public expenditure should therefore reflect the extent of market failure and the importance of distributional objectives. It is then interesting to ask whether the observed positive correlation between public expenditure shares and income, both between countries and over time, is linked to these factors (Table 1). The conclusion is unclear. It seems reasonable to conjecture that the richer a country, the more it can afford general redistribution. Also, as development proceeds income support is provided less by the extended family, and formal social security is needed in its place. Variations in social expenditure support such a conclusion. The link between other market failures and income is unclear, although it is often argued that poorer countries are characterized by more extensive market failure. Inadequate financial and capital markets are a case in point. Hence the dominant role of government in investment and production in many developing countries. But while market failure may be a bigger problem in poorer countries, they rarely have the resources to support all the necessary compensatory policies. On balance, market failure and distributional concerns probably suggest that public expenditure should increase with income and development. Specific theories of expenditure growth contribute a fuller explanation of expenditure trends.
Wagner’s Law—named after its nineteenth century originator—suggests that as a society becomes industrialized, the set of social, commercial, and legal relationships within it becomes more complex. Governments will occupy a more prominent role in establishing and running institutions to control this complexity. More recently, it has been argued that in a democracy there is a notion of the tolerable burden of taxation, and governments are severely constrained from increasing expenditure dramatically. However, during periods of social disturbance, such as war, famine, or some natural disaster, the level of tolerance for taxation is greater and, consequently, government expenditure is able to expand. After the social disturbance, the tolerable rate of taxation does not fully return to its original level; public expenditure therefore increases permanently due to the displacement effect.
Other approaches emphasize factors that influence the demand for public programs. For example, the role of demographic factors is particularly important in both industrial and developing countries. Aging populations in many industrial countries over the last three decades have been an important factor in explaining the growth of social expenditure on pensions and health care for the elderly. This is also expected to be the predominant demographic trend influencing social expenditure over the next forty years in these countries (see the note on Pensions for more detail). By contrast, developing countries have experienced high birth rates and declining infant and child mortality rates in the 1960s and 1970s, and many are still experiencing these trends. The result has been a rapid growth in their population as a whole and particularly the younger age cohorts. This has led to increased spending on education and health care. The role of supply-side factors in explaining the growth in public expenditure is emphasized in models of unbalanced productivity growth, which imply that the price of the goods and services provided by the government rises more rapidly than the price of privately produced goods and services (see the note on Public Expenditure Measurement for further discussion).
Much recent attention has focused on attempts to explain the growth of public expenditure in terms of political processes. Consumers reveal their preference for public programs through their voting behavior. The views of elected representatives are then mediated by the decisions of the bureaucrats responsible for policy and program implementation. Public choice theorists have advanced a number of hypotheses on the political and bureaucratic processes by which government decisions are made. It is argued, for example, that majority rule may lead to an over-provision of public programs because coalitions form to push expenditure above levels consistent with willingness to pay. Expenditure programs, however, do not come into existence merely because some interest group wants them. They must be provided by a government bureau. Thus, government expenditure may grow not because increasing expenditures are demanded by its citizens and interest groups, but because they originate with the bureaucracy supplying such programs, whose power and prestige is enhanced by larger budgets.
Analyzing Public Expenditure
A clear distinction needs to be drawn between normative rules describing the extent to which governments ought to intervene and positive explanations of the degree to which they actually do intervene. Market failure and distributional concerns constitute good arguments for public expenditure. Some of the theories of expenditure growth reflect such arguments. If the government has a regulatory role in the economy, the more complicated the economy becomes the greater the degree of regulation that may be needed. Similarly, if there is a compelling case for public provision of retirement pensions, then aging populations will put pressure on expenditure. But other theories are based upon bad arguments for public expenditure. There is little merit in public programs and high levels of expenditure that reflect misconceptions about the potentially damaging effects of increased taxation, low public sector productivity, the influence of pressure group politics, and the misconceived objectives of public bureaucrats.
The distinction between good and bad arguments for public expenditure is central to assessing the level and composition of public expenditure. In seeking to analyze a public expenditure program from a microeconomic perspective, the objective should be to distinguish between good and bad expenditures using arguments that derive from a clear view of the justification for government intervention. The focus of attention should be (i) the extent of market failure; (ii) distributional objectives; and (iii) the efficiency cost of compensating interventions. However, as the note on Public Expenditure Productivity reveals, the scope for clear-cut judgments may be rather narrow.
Atkinson, A.B., and J.E. Stiglitz, Lectures on Public Economics (New York: McGraw Hill, 1980).
Bohm, P., Social Efficiency ( London: Macmillan, 1973).
Mirrlees, J.A., “Arguments for Public Expenditure”, in Contemporary Contributions to Economics, ed. by M. Artis and A.R. Nobay (Oxford: Basil Blackwell, 1978).
Saunders, P., and F. Klau, The Role of the Public Sector, OECD Economic Studies, No. 4 (Paris: Organization for Economic Cooperation and Development, 1985).