V. Public Expendituer Measurement
- Ke-young Chu, and Richard Hemming
- Published Date:
- September 1991
How is measured public expenditure affected by the choice of definition of the public sector?
Which activities should be included in measures of public expenditure?
Is the size of public expenditure a good guide to the role of the public sector in the economy?
No single measure of the size of government can serve every purpose for which such a measure might be needed. Government activity takes a variety of forms—consumption, production, redistribution, regulation—and each influences the economy and society in different ways. Ideally, a measure of government size should be tailored to the activity that is of interest. In practice, however, there has been a tendency to describe the size of government by reference either to the level of public expenditure or its share in national income. The purpose of this note is to show that a number of problems arise with this approach. As a result, comparisons of expenditure levels and ratios across time and between countries have to be treated with caution.
The Level of Government
For most countries, data are readily available on expenditure by the central government. But expenditure is also undertaken by local (and regional) governments and public enterprises. Expenditure can therefore be defined also for the general government (i.e., central and local governments) or the public sector (i.e., general government and public enterprises). It would, however, be incorrect to add the expenditure recorded by the central government, local governments, and public enterprises to yield total public expenditure. Part of central government expenditure typically reflects transfers to local governments and public enterprises which are in turn spent by the recipients. Transfers between different levels of government have to be eliminated to avoid double counting. In the case of public enterprises, current spending on wages and salaries, purchases of goods and services, etc., is not included in public expenditure, the argument being that it is only to the extent that this expenditure is not covered by internally generated income that public enterprise activities impinge upon the economy and society.
In defining public expenditure, the preference should always be for properly consolidated data (i.e., with intergovernmental transfers eliminated) relating to the public sector as a whole. However, in many cases only more limited data are available and the resulting incompleteness of coverage, combined with the complexity of financial relationships that normally exist between the covered and omitted levels of government, will necessitate careful interpretation of these data. To the extent that expenditure and expenditure-type activities are not reflected in reported data, even greater care is needed. This problem arises especially in the context of measures of government expenditure derived from budget data.
Off-budget expenditure derives from the operations of a wide range of extrabudgetary accounts, special funds and decentralized agencies. To the extent possible, all expenditure of the government should be reflected in measured expenditure even if some of it is not included in the expenditure budget. Common omissions include expenditure by social security schemes, government lending, and noncommercial expenditure by public enterprises.
Social security expenditure and the corresponding taxes are sometimes included in the budget. However, in many cases there is a separate financial agency, often a social insurance fund, through which social security revenue and expenditure flows. There is some ambiguity as to whether funded social insurance payments should be included in public expenditure only to the extent that they are not covered by revenues, as would be the case with the expenditure of other public financial institutions, or in total, as with social security expenditure in the budget. The compulsory nature of social insurance tends to point to the latter treatment.
It is not uncommon for government lending operations to be outside the formal budget, except insofar as the lending institution or program receives funds directly from the budget. However, in many cases the government is on-lending foreign loans and this has a direct budgetary impact only if borrowers are delinquent, if the government lends on more concessional terms than it receives, or if the government is required to cover foreign exchange losses. Ideally, these lending operations should be fully reflected in expenditure data for the central government, as part of net lending. Inasmuch as the central government lends to other levels of government—especially public enterprises—on-lending is netted out upon consolidation. But if the borrower is the private sector, and in many countries the government has been forced to inject capital into failing nonfinancial enterprises and financial institutions, such lending should be included in public expenditure.
Noncommercial activities of public enterprises
The above-mentioned treatment of current spending by public enterprises assumes that the activities concerned are of a commercial nature, as is the case in the private sector. Yet public enterprises are charged with a wide range of social objectives—employment generation, supplying goods and services to remote areas, charging low prices—and are required to concede profit in the process. Sometimes enterprises will be paid a subsidy to cover the cost of a specific noncommercial activity; other times an operating subsidy will be paid to loss-making enterprises; and profitable enterprises will simply make reduced profit transfers to government. Generally, the cost of the public policy objectives pursued by public enterprises is not included in expenditure data. The note on Fiscal Activities of Public institutions discusses how these and similar activities of public financial institutions can be better reflected in such data.
Other Off-Budget Activities
In addition to off-budget expenditure, the government also undertakes activities which affect the economy in ways that are underestimated by their contemporaneous expenditure requirements. Regulation and government loan guarantees are examples.
Only the direct—mainly administrative—costs of government regulatory activity are reflected in the fiscal accounts. This is clearly a highly imperfect measure of the scope and influence of such activity. While there have been many attempts to measure the impact of regulatory policy, these are contentious at best. Efforts to measure private expenditure mandated by regulatory policy (e.g., the cost of fitting catalytic convertors to motor cars) come closest to what might be considered reasonable. There are, however, obvious difficulties in attempting to measure the private costs associated with many other regulations, such as those regarding environmental health, urban planning, and consumer product safety. Moreover, it would be inconsistent to aggregate a measure of the cost impact of regulatory policy on the private sector with public expenditure, the consequences of which for private spending are not taken into account. When regulation takes the form of wage and price controls (the latter also including fixed exchange rates and administered interest rates), this will also distort measures of public expenditure. Corrected measures should reflect the use of appropriate shadow prices, which are discussed in more detail in the note on Public Investment. However, given measurement difficulties, the best one can usually hope for is some appreciation of the significance of regulatory activities when discussing the impact of the government on the economy.
Government loan guarantees
In addition to its direct lending activities, the government also guarantees loans to other levels of government, and in particular public enterprises, and to the private sector. At the time a guarantee is given no expenditure is involved; it is a contingent liability which will only give rise to a payment in the event of default. Such guarantees nevertheless have an economic impact, and to ignore them is to understate the influence of government on the economy. However, like regulation, it is difficult to reflect the importance of loan guarantees (as distinct from payments to which such guarantees give rise) in measured public expenditure; while loan guarantees can be valued on a probability basis, this is not the same as expenditure. But there is nevertheless a strong case for making provision for the likely expenditure implications of guarantees. Contingent liabilities are discussed further in the note on Fiscal Activities of Public Institutions.
Expenditure figures are also misleading when the same objective is met through the expenditure budget in some countries and through the tax system in others. For example, financial assistance to families with children can be provided either by paying a cash benefit or by granting additional tax-free income in the form of an allowance or a tax credit. While these alternative methods of provision can have the same impact on the family concerned and on public finances, one is treated as an expenditure item while the other is an offset against tax revenue. There is a general consensus that these offsets are to all intents and purposes expenditure—hence the term tax expenditures. By the same token, items of recorded expenditure in some countries, such as tax refunds, should properly be treated as offsets against revenue.
Measuring tax expenditures is not straightforward. First, a question arises as to what is the “normal” tax structure from which a tax expenditure represents a departure. There may be many features of the current tax structure that reflect the existence of tax expenditures, and if the latter were eliminated the former would also be removed. Second, there is disagreement as to whether the impact of tax expenditures should be quantified on a revenue forgone basis (calculated on ceteris paribus assumptions), a revenue gain from removal basis (which incorporates behavioral responses) or an outlay-equivalent basis (by estimating the expenditure necessary to achieve the same objectives). From the point of view of measuring and comparing expenditure, it is probably the latter that is most appropriate. Estimates of tax expenditures are available for only a few industrial countries; these do reveal, however, that such expenditures can be quantitatively significant.
While off-budget expenditure and tax expenditures lead to an underestimation of expenditure, cost recovery could constitute an at least partially offsetting influence. Public expenditure can be defined on a net or gross basis. Many goods and services provided by public agencies are subject to fees and charges, and a case can be made for adjusting expenditure to reflect the extent of cost recovery. The imposition of fees and charges makes public sector activities largely indistinguishable from those undertaken in the private sector, and as such they could justifiably be excluded from public spending. Nevertheless, the standard procedure is to define government expenditure on a gross basis, and as indicated above, only the current outlays of public enterprises on a net basis.
Full adjustment of expenditure data to reflect off-budget spending and tax expenditures may not be possible, but a minimum requirement is reasonably consistent treatment both over time and between countries. Similarly, there should also be consistency in reflecting the activities of other levels of government in public expenditure aggregates. However, in interpreting the resulting expenditure data, due account needs to be taken of off-budget activities, such as regulation and the provision of loan guarantees, that have an influence on the economy far greater than associated expenditure figures suggest. In addition, it is necessary to look at the composition of expenditure, and in particular the role of transfers.
The Impact of Transfers
Public expenditure can be separated into exhaustive spending and transfer payments. Exhaustive spending reflects purchases of goods and services while transfer payments redistribute purchasing power between different members of society. For a given level of total expenditure, the split between exhaustive spending and transfer payments affects the extent to which the public sector absorbs real resources, and to many people it is the latter that provides the most meaningful indicator of the economic impact of public expenditure. But the way public expenditure is paid for is also of economic significance, and a definition of public expenditure which attempts to capture the influence of taxation and borrowing associated with a particular level of expenditure on the economy should include transfers. Clearly, the appropriate definition depends upon the purpose to which the resulting measure is put. The choice of definition and measure also implies that expenditure to national income ratios have to be interpreted with special care. Since exhaustive expenditure is reflected in national income but transfers are not, similar levels of total expenditure can correspond to widely different expenditure to national income ratios at the same level of private sector activity.
Consider, for example, the following possibilities:
|Public expenditure/national income||33%||40%||50%||100%|
In case A public expenditure is half the size of private expenditure, and all public spending is on consumption and investment. Public expenditure is therefore 33 percent of national income. However, if public expenditure is split equally between exhaustive spending and transfers, as in case B, the public expenditure ratio rises to 40 percent, reflecting the exclusion of transfers from national income. When the government does nothing but redistribute income through transfers, as in case C, the public expenditure ratio climbs further to 50 percent. With a particularly redistributive government—that is, one that imposes heavy taxes to finance large transfers—the public expenditure ratio can rise to 100 percent, as in case D, and higher. The implications of different mixes of exhaustive spending and transfer payments in any expenditure aggregate are therefore that: (i) high public expenditure, in particular a high public expenditure ratio, need not correspond to a large share of public spending in aggregate demand; and (ii) similar expenditure ratios can have quite different impacts on the economy and society. It is therefore important to know the broad composition of the aggregate, even for macroeconomic analysis.
Public Expenditure and Public Sector Output
Ideally, changes in public sector output should be judged by measuring public goods and services valued at appropriate prices. Public expenditure is only the cost of providing these goods and services. Thus, a 10 percent increase in nominal expenditure does not translate into a similar increase in the volume of goods and services provided by the public sector. Expenditure has to be deflated by the relevant price index, which should in turn reflect the price indices for the various components of expenditure. In general, these will differ from indices corresponding to private sector output, the usual argument being that the price of public sector output rises faster than that of private sector output. While wage costs are similar, productivity growth is slower in the public sector, reflecting the low level of competition, the inefficiency of complex bureaucracies etc. There is therefore a built-in tendency for public expenditure to grow relative to national income without any increase in output, because of this relative price effect. To assess properly the underlying output growth associated with a particular rate of expenditure growth, it is necessary to construct an appropriate output price deflator, or shadow price. But because public output is largely unmarketed, an input price deflator is often all that is available. Comparison based upon changes in the relative prices of public sector inputs and private sector outputs (or inputs)—which define away productivity improvements in the public sector (and, in the care of input price comparisons, also in the private sector)—are therefore not uncommon.
International Monetary Fund, A Manual on Government Finance Statistics (Washington, 1986).
Saunders , P., and F. Klau, The Role of the Public Sector, OECD Economic Studies, No. 4 (Paris: Organization for Economic Cooperation and Development, 1985).