THE FORM in which a government’s budget is presented reflects to some extent official attitudes toward the various components of receipts and expenditures, and it also influences popular interpretations of the government’s program. In this respect the treatment of capital receipts and expenditures is especially important. A number of countries maintain separate capital and current budgets, and other countries accord special treatment by other means to at least a part of capital items. The advisability of introducing a capital budget has recently been under public discussion in several countries that have not used the device. At the same time, there have been indications that in some countries with established capital budgets the financial implications of the system are being reconsidered.


THE FORM in which a government’s budget is presented reflects to some extent official attitudes toward the various components of receipts and expenditures, and it also influences popular interpretations of the government’s program. In this respect the treatment of capital receipts and expenditures is especially important. A number of countries maintain separate capital and current budgets, and other countries accord special treatment by other means to at least a part of capital items. The advisability of introducing a capital budget has recently been under public discussion in several countries that have not used the device. At the same time, there have been indications that in some countries with established capital budgets the financial implications of the system are being reconsidered.

THE FORM in which a government’s budget is presented reflects to some extent official attitudes toward the various components of receipts and expenditures, and it also influences popular interpretations of the government’s program. In this respect the treatment of capital receipts and expenditures is especially important. A number of countries maintain separate capital and current budgets, and other countries accord special treatment by other means to at least a part of capital items. The advisability of introducing a capital budget has recently been under public discussion in several countries that have not used the device. At the same time, there have been indications that in some countries with established capital budgets the financial implications of the system are being reconsidered.

This paper explores the economic implications of a dual government budget system involving separate capital and current budgets. It reviews the purposes that such systems appear to be intended to serve and criteria for distinguishing between capital and current items. The influence of a dual budget system on the amount and composition of government expenditures and means of financing is examined. Attention is also given to the problem of measuring government expenditures, receipts, and surplus or deficit when a separate capital budget is employed.1

General Nature of Capital Budgets

The scope of government capital budgets in different countries and the financial practices associated with them vary widely. The one essential characteristic is that some or all of the general government’s expenditures for “capital” items are segregated and presented in a special budget separate from the budget covering operating expenditures.2 Capital items for this purpose should be understood broadly to cover both physical and financial assets which are expected to yield a return in future years. This return may be in the form of services, as from public buildings, or of receipts from sales, interest, dividends, etc. Few if any governments attempt to include all transactions in capital assets, in this broad sense, in the capital budget, but some items of each type are often included. Moreover, a few governments include in the capital budget items that would not ordinarily be regarded as assets.

The establishment of a separate capital budget carries a strong presumption that the method of financing the expenditures included in it will normally differ from that used for other government expenditures. Usually, borrowing is considered more appropriate for capital expenditures than for current expenditures; in the Union of South Africa, for example, the capital budget is called the “loan account.” In inflationary periods, however, some countries, including South Africa, make a special effort to cover a part of capital outlays by a current account surplus. In some capital budgets, the principal item on the receipts side is proceeds of borrowing; in other systems, borrowing is not listed as a capital budget receipt but is considered an appropriate means of covering any excess of capital outlays over capital receipts. Other common items of capital receipts include transfers from the operating budget, proceeds of sales of state property, and recoveries of loans. It is sometimes suggested that revenues from death duties, capital levies, and other taxes that are presumed to be paid from private capital be assigned to the capital budget. Certain user taxes, such as those on motor fuels, are occasionally allocated to the capital budget.

Governments that use capital budgets often provide for depreciation or amortization allowances against certain assets. These allowances are usually recorded as receipts in the capital budget and as expenditures in the operations budget3 or the enterprise budgets. In Sweden and Ecuador initial write-offs of the nonremunerative part of capital expenditures are included in current expenditures when the outlays are made. In some countries—for example, Sweden—provision is made for rental charges for occupancy of public buildings, which are recorded as expenditures in the current budget and as current receipts in the enterprise budgets.

Governments that do not employ formal capital budgets may approximate some of the results of capital budgeting by the establishment of autonomous government entities which carry out public investment and whose activities are in the main outside the general budget. In Mexico, for example, a sizable fraction of government investment has been financed through a national finance corporation and the government banks. Lease-purchase arrangements for public buildings and other facilities also resemble capital budgeting in the sense that the government’s operations budget reflects the annual cost of the facilities (consisting mainly of operating costs, interest, and depreciation) over a period of years rather than the construction cost at the time when the facilities are built. In the United States, where these arrangements have recently been adopted for government office buildings and post offices, the original construction cost is financed by private investors, and the government leases the building, receiving title after an agreed number of years.4

Purposes of Capital Budgets

Inasmuch as there are usually no detailed official statements of the purposes that are expected to be served by the establishment of a capital budget, the discussion of these purposes must be conjectural, to some extent.

Reflection of government’s assets

An advantage that is often claimed for a separate capital budget is that the distinction between expenditures which do and expenditures which do not result in the acquisition of an asset aids in the interpretation of figures on government finance. This approach commends itself to many persons as an application of private accounting principles. In private accounting, outlays for assets that are expected to yield a return over a period of years are ordinarily “capitalized” (i.e., charged or debited to asset accounts). Capitalized outlays, in contrast to current expenses, are not immediately deducted from receipts in computing annual profits of a private firm. (If the asset is depreciable, its cost is deducted in installments over a period of years which in principle coincides with the expected useful life of the asset.)

By analogy, it may be argued that there is a notable difference between government expenditures for assets and for current operations. In the government as well as in the private sphere, expenditures for assets give rise to the expectation of a future return. They represent collective investment, whereas operating expenditures represent collective consumption. The analogy with private accounting suggests that the inclusion of investment outlays in government expenditures, without recording the matching asset, will give a misleading impression of the state of government finances.

From the analogy with private accounting, it is an easy step to conclude that an excess of government expenditures over revenue receipts when incurred in order to acquire capital assets should not be regarded in the same light as when it is attributable to current operations. If the excess expenditures are covered by borrowing, there is no change in the government’s net asset position (net worth) when the expenditures are for capital assets, the increase in indebtedness being matched by an increase in assets. When current expenditures are met by borrowing, however, the government’s net assets or net worth (assets minus liabilities) decline. Similarly, financing capital expenditures by the use of government cash resources or the sale of property does not change net assets, but financing current expenditures in this way reduces net assets.

The extent to which private accounting principles are properly applicable to government is questionable, however. The primary objective of private business is to increase net worth and distributable profits. Both current expenses and capital outlays are merely a means to that end and should be held to the lowest level that is consistent with maximizing profits (and net worth). The objectives of government are more complex, and the extent to which they are achieved cannot be measured simply by reference to a balance sheet.5 In this respect, government operations are more closely analogous to those of a household than to those of a business firm. Expenditures by government to provide current services for the community resemble a family’s consumption expenditures in that both types of expenditure represent a proximate end rather than a means. Acquisition of assets which will increase future capacity to consume does not necessarily have a rational claim to priority over current consumption in the plans of a government or a household.6

There are important practical difficulties in valuing many government assets because they serve purposes which are not market-oriented or profit-seeking. In addition to the conundrums presented by attempts to value such items as art museums, roads, and battleships, elements of complexity and uncertainty would be involved in any effort to report tax revenues on an accrual basis. In some cases, final accounting of government transactions has been greatly delayed, and figures have been frequently revised in an effort to assign receipts and expenditures to the year in which they accrued rather than to the year in which the cash receipts or payments occurred.

A possible disadvantage of government capital budgeting is that the analogy with private accounting may be interpreted as suggesting that acquisition of assets is generally preferable to expenditures for current operations. Such a preference may be strengthened if the financing of capital outlays is politically easier than the financing of current expenditures, as may well be the case when capital items are covered by borrowing.

Allocation of costs over time

In government accounting as well as in private accounting, the practice of capitalizing certain outlays may serve as the basis for a convenient method of allocating costs over time.7 The most familiar example is provided by durable plant and equipment. In private accounting, outlays for these items are capitalized in the year in which they occur and are written off in the form of depreciation or amortization allowances over the period of years during which the plant and equipment are expected to contribute to income. This procedure reduces irregularities in the reported costs and net income, and is universally considered to result in a more accurate statement of business costs and profits in each time period.

The same reasoning may be applied to government costs. If, for example, the government spends 1,000 to construct an office building that is expected to last 50 years, a more useful statement of the cost of office space may be obtained if this is reported as, say, a series of items of 20 a year for 50 years than if it is recorded as 1,000 in the first year. Under the accrual method, it is easier to compare the cost of erecting a government building with the cost of renting a privately owned building and also to appraise the cost of additional office space in the light of expected gains in convenience and efficiency.

The procedure, it should be noted, requires both the segregation of capital outlays when made and their distribution over the estimated service life of the assets by means of depreciation or amortization allowances. A budget system, such as is in effect in several countries, which provides for segregation of capital outlays but not for depreciation or amortization allowances, will not achieve any significant improvement in cost allocation over time.8

To be sure, the method of accounting does not alter the fact that the original construction imposes a draft on economic resources and thus represents a real cost to the community, in terms of foregone opportunities, when the work is carried out. It is now generally recognized that this real cost cannot be passed on to future generations. The original outlay also presents the government with a financing problem at that time. These considerations are two of the principal reasons for obtaining accurate figures on total government costs; both suggest the reporting of capital as well as current expenditures in the period when incurred.

On the other hand, an analysis of the costs of particular government programs or services may be facilitated by capital accounting and the spreading of depreciation allowances. When the timing of capital outlays is irregular, the inclusion of both capital and current outlays when they are originally made will result in wide fluctuations in reported annual costs of the program and will complicate the already difficult process of comparing annual costs and benefits. Under a capital budget system providing accrual accounting, annual program costs as reported in the operations budget would include depreciation or amortization (and possibly interest) on capital equipment.

This phase of program costing stems entirely from the irregularity of capital outlays. In a perfectly stable program involving a great number of comparatively small capital outlays, these outlays would equal normal depreciation allowances each year.9 It would therefore be a matter of indifference whether capital costs were reported on the basis of cash outlays or of accrued depreciation. When, however, comparatively large and infrequent capital outlays are required, the two methods of costing would yield rather different year-by-year figures. Since irregularity of capital outlays is likely to be more characteristic of small programs and small governmental units than of large programs and units, capital budgeting and accrual accounting may be especially revealing for the former. Furthermore, the two methods of accounting may yield sharply different cost figures for rapidly expanding or contracting programs and in the years immediately after a new capital budget is adopted.

Another aspect of the measurement of program costs relates to the degree to which capital equipment is specialized rather than to the regularity with which it is acquired. When capital equipment is to some degree nonspecialized and hence capable of being diverted from one governmental use to another or of being sold to private investors, it may be difficult to allocate capital costs among various programs at the time when the equipment is acquired. For example, the government may buy earthmoving equipment, employ it initially for road building, and later use it to construct irrigation canals. In this case, an accurate allocation of the cost of the equipment between road construction and irrigation works may not be feasible when the machinery is purchased. The cost of fully specialized equipment, on the other hand, can be definitely assigned to a program at the outset.

Despite the theoretical advantages of capital budgeting in measuring annual program costs and in allocating resources among competing uses, there is little evidence that these potentialities have so far received much attention or application. In most countries, more emphasis has been given to broad aggregates, which are more difficult to interpret and to use as policy guides.

Administrative Management and Control

Inasmuch as accurate statements of program costs are indispensable to efficient management and control at the program level, the possible advantages of capital budgeting and accrual accounting for these purposes, given certain conditions, are evident. The degree of specialization of the capital assets is again a highly relevant question. Administrative management and control are much more likely to be facilitated by including depreciation accruals in program costs when the capital items are unspecialized than when they represent resources already irrevocably committed to a particular program. In the former case, the fiscal authorities have to consider periodically whether the services of the non-specialized capital equipment should continue to be devoted to the existing use or should be made available to another program which is judged to have a stronger claim. This decision is parallel to that with respect to, say, the employment of personnel and can be most effectively made by reference to annual costs, represented by depreciation accruals. With respect to specialized equipment, on the other hand, past capital outlays need be taken into account only to the extent that they influence the service yield of current expenditures for labor and materials.

In the general government sector, capital budget techniques have not yet been widely applied for management and control purposes at the program level. In the public enterprise sector, the merits of accrual accounting for capital outlays have achieved more general recognition. State enterprises customarily employ commercial accounting which does involve accrual of capital costs. Application of this procedure to enterprises does not require a capital budget for the government proper.

Information on capital formation

A capital budget may serve the purpose of providing information on capital formation in the public sector. Although it is undeniably true that many so-called current expenditures of government—as well as some current expenses of business firms and part of private consumption—yield a significant future return, special interest attaches to outlays for tangible items of a durable nature which have this characteristic to a peculiar degree. These last items are included in estimates of capital formation in the private sector, and many systems of social accounts call for information on these items in the government sector as well, to arrive at comprehensive estimates of national capital formation.

In order to provide information on gross capital formation in the government sector, a separate capital budget is not essential. All that is required is to make a distinction between capital and current expenditures, which may be supplied by classification within a unified budget or by an informational table. (An informational table is appended to the budget of the United States, for example; and in Canada a list of physical assets valued at original cost is presented.) This is not enough, however, to provide information on net capital formation. For this purpose, it is necessary to present figures on depreciation, either as part of a formal capital budget system or by ad hoc estimates.10

For social accounting purposes, it is also necessary to distinguish between transactions in government assets which represent a real accretion to the community’s wealth and those which do not. The former consist of acquisitions of newly produced durable physical assets, such as machinery and buildings; they are a component of national capital formation. The latter consist of existing physical assets acquired by purchase, such as land and secondhand machinery and buildings, and financial assets, such as loans and securities. Government capital budgets, being designed on the same lines as private accounting systems, ordinarily do not show this distinction and hence are not readily adaptable for social accounting purposes.

It may be concluded that a properly oriented capital budget system can make a helpful contribution to the provision of social accounting data on capital formation in the public sector. A formal capital budget, however, is not indispensable to the preparation of complete social accounts. Estimates of government capital formation can often be made on the basis of collateral data. The budget and public accounts, moreover, are drawn up primarily to facilitate government management and policy making. Considerations of statistical convenience are secondary in importance.

Political acceptability

Capital budgeting may have the effect of increasing the political acceptability of government capital expenditures and of government borrowing. Opposition to tax increases and to budget deficits may pose a limitation to government expenditures in most democratic countries. When a dual budget system exists, taxation requirements are likely to be decided mainly by reference to the operations budget. Expenditures in the capital budget may be covered to a large extent by borrowing; but, since assets are being acquired, this borrowing may not be considered a form of “deficit” finance. Expenditures assigned to the capital budget may therefore escape partly or entirely the usual political checks.

As already explained, however, a full-fledged capital budget involves not merely the removal of capital outlays from the operations budget but the addition of estimated depreciation or amortization requirements to the expenditure side of the operations budget. If the rate of capital outlays were constant, these allowances would, after a time, equal new outlays for depreciable assets, and the size of the operations budget would be completely unaffected by the segregation of outlays for such assets in a separate capital budget.11 The amount of expenditures most directly subject to the normal political checks—i.e., expenditures in the operations budget—will be smaller under a divided budget system than total expenditures under a unified budget to the extent that outlays for depreciable assets exceed annual depreciation allowances, owing to an irregular or cyclical fluctuation or to secular growth. Exclusion of outlays for nondepreciable assets, such as land and certain financial assets, will tend to hold down the size of the operations budget in all years unless the purchase of nondepreciable assets is charged in some other way to the current budget. One way of doing this would be to include in current expenditure the subsequent debt interest and amortization charges relating to land purchases financed by borrowing.

Capital budgets have been favored by those who believe that when a unified budget is employed government expenditures tend to remain below the optimum level. During the 1930’s some advocates of the capital budget in the United States believed that its introduction was desirable to bring about a relaxation of traditional checks on government expenditures, which they felt were interfering with compensatory fiscal action. Although the problem of capital consumption allowances was often neglected in these discussions, advocates of the capital budget might have argued that, since in a period of depression government capital outlays should expand more rapidly than depreciation accruals, the inclusion of depreciation allowances would not pose a serious obstacle to the political acceptability of the policy that they wished to pursue. Other advocates of the capital budget have held that in rich countries there is a chronic tendency for government expenditures to fall below the optimum level, and that introduction of the capital budget would make it easier to overcome this tendency.

Many critics of the capital budget regard the political arguments in its favor as evidence of a cynical lack of confidence in public understanding of economic affairs. They argue that any appropriate countercyclical adjustment of government expenditures and means of finance can be made within the framework of a unified budget, and they favor reliance on education to break down traditional opposition to such a policy. In regard to the secular problem, some of the critics deny the need for a relaxation of the balanced budget norm and rely on monetization of public or private debt to provide any required growth of liquidity.

Workable rule for government borrowing

A somewhat more sophisticated version of the political arguments in favor of the capital budget is the contention that capital budgeting offers a workable rule for regulating the amount of government borrowing. According to this line of reasoning, government borrowing has positive advantages, provided it can be kept within reasonable limits. These limits, it is said, can be enforced by adopting a divided budget and two rules for financing. The first of these rules is that current expenditures should normally be covered by current revenues. This norm of balancing the operations budget is intended to provide a desirable element of discipline and to help prevent extravagance. The second rule is that capital outlays will normally be financed by long-term borrowing12 and that proper depreciation allowances on capital assets will be entered as expenditures in the operations budget. This rule is intended to prevent an excessive increase in government debt and service charges.

The first advantage claimed for the plan of borrowing to finance capital outlays is that it avoids the frequent tax revisions which would have to be made if a fluctuating volume of capital outlays were to be covered out of current revenue. Although fluctuations in capital outlays are likely to be greater for a small unit than for a national government, capital outlays are probably more irregular than operating expenditures for virtually all governmental units.

A second argument in favor of borrowing to finance capital outlays is that it allows a better distribution of the cost of government services among beneficiaries and taxpayers. By adopting loan finance, the government can defer final allocation of capital costs within the community and can levy charges or taxes designed to recover capital costs from users of the services. Tax financing, on the other hand, brings about a final allocation of costs at the outset. This consideration should be sharply distinguished from the fallacious notion that the total social cost of government expenditures can be passed on to future generations by means of borrowing. The real costs of capital outlays, as measured by foregone opportunities, are necessarily incurred at the moment when the facilities are produced.13 Under a program of loan finance, however, these costs are voluntarily assumed at that time by creditors of the government. They are finally allocated when the creditors are repaid out of the proceeds of taxes or charges collected from other members of the community.

A third argument in favor of loan financing for capital outlays is that it results in a better division of the national product between consumption and investment. To the extent that taxes are raised out of income that would otherwise be used for private consumption, tax financing of current expenditures of government will bring about a substitution of public consumption for private consumption. On the other hand, long-term government borrowing from the public tends to reduce private investment by restricting the availability of funds and raising interest rates. Hence, the financing of government net capital formation by borrowing will tend to reflect a substitution of public investment for private investment. It can be argued that this is a reasonable substitution and that loan financing involves a minimum of interference with the freely exercised choice of the public regarding the division of national production between total consumption and investment.14 At the same time, in a growing economy the secular increase in government capital outlays will provide for an expansion of government debt to satisfy the public’s demand for such assets.

Most proponents of the divided budget and the financial rules mentioned above would concede that, under conditions of incipient or actual unemployment and deflation, it is not advisable to attempt to finance government net investment in a way that will displace private investment. Under these conditions, the real (opportunity) cost of government investment and also of current services is reduced and may become virtually zero or possibly negative. Hence it may be desirable to finance government capital outlays, and perhaps even a part of current operating expenditures, by short-term borrowing15 or by an expansion of central bank credit.

In opposition to the claim that the divided budget system provides a workable rule for government borrowing, it may be contended that loan financing of capital outlays introduces an inflationary bias at all times. The curtailment of private expenditures that is caused by even long-term borrowing is ordinarily less than that brought about by taxation; and this tends to be more true if the government maintains easy credit conditions in an attempt to hold down interest costs on the public debt. An underdeveloped country, in particular, may have few opportunities for borrowing from genuine savings of the public, and loan financing of capital outlays may involve recourse to the commercial banks and the central bank, leading to an excessive expansion of credit and to chronic inflationary pressure. Furthermore, an underdeveloped country may be interested more in increasing total capital formation than in respecting the wishes of the public regarding the division of the national product between saving and consumption, and therefore may not hesitate to finance government investment by means of taxation.

Distinction Between Capital and Current Items

In practice, the criteria employed for allocating items between the capital and the current budget are often not carefully stated and, to a considerable extent, they appear to be arbitrary. This section discusses some of the criteria that have been used or suggested.

Asset criterion

The fact that capital expenditures change the form but not the net total of government assets, while current expenditures reduce the net assets of the government,16 provides one general criterion. A capital expenditure for the acquisition of physical or financial assets involves the exchange of cash (or an increase in liabilities) for these assets. Similarly, many capital receipts reflect merely a change in the form of government assets and liabilities. For example, receipts from the sale of physical or financial assets initially involve an increase in cash and an equal decrease in other assets; and net borrowing increases assets (cash) and liabilities (public debt) by equal amounts. However, receipts from capital taxes, which are sometimes included in the capital budget, reflect a net increase in government assets. Current expenditures involve a decrease in government assets and current receipts involve an increase.

This approach parallels private accounting usage, but it merely throws the question back to the definition of “assets.” Although a fairly explicit set of conventions exists in private accounting for identifying assets and distinguishing between capital outlays and current expenses, some difficult questions arise in this connection. In government accounting, conventional rules for allocation are less well settled and more problems are involved.

In both private and government accounting, the time factor is the primary basis for identifying assets. An asset represents basically a claim for or an expectation of future income or services. Usually the claim or expectation is associated with a physical good, such as a building or piece of machinery, or with a security, such as a bond or share, and these items are customarily referred to as “assets.” To some extent, however, private accounting principles recognize the existence of assets which represent merely accounting entries indicating the expectation of future income. An example is the capitalization of research expenditures. Goodwill of a business may also appear as an asset on the firm’s books even though, unlike most other assets, it may not be specifically associated with identifiable past expenditures. It can be confidently asserted, however, that no widely used accounting system provides for the capitalization of all expenditures that can reasonably be expected to add to the entity’s future income. Thus disbursements that are classified as ordinary operating expenses contribute indirectly to the earning of future as well as of current income, if only through the preservation of the firm as a going concern or the continuity of the unit.

In the more inclusive government capital budgets, expenditures for durable goods and for the acquisition of financial claims on other sectors (which include securities and loans) are classified as capital outlays. Expenditures for the acquisition of goods under a stockpiling program are also sometimes considered capital investments,17 but inventories of office supplies and other materials are seldom treated as capital assets except in government enterprises which employ commercial accounting systems. As a matter of convenience, comparatively minor items, such as office furniture and equipment, are often omitted from government capital accounts.

Outlays for military equipment are also commonly omitted from capital accounts. This omission may reflect in part the belief that military items have a high rate of obsolescence and an uncertain life. It may also be associated with the attitude that these items are not “productive” in the same sense as civilian capital goods. The specialized nature of much military equipment means that, for reasons stated above, expenditures on this account can be unequivocally assigned to the defense program.

The absence of a clear-cut profit standard may help to explain the fact that governments seldom classify as capital items expenditures for certain services that may reasonably be expected to yield a return to the community in the future as well as the present. Examples of this type of item are expenditures for education, public health, and research. The special tabulation of investment expenditures of the U.S. Government, however, includes outlays for research, education, and health (and also on military equipment).

“Productivity” criterion

“Productivity” is sometimes suggested as the basis for identifying capital assets, but on examination this proves to be a slippery concept. In the broad sense, all government expenditures that satisfy wants of the community must be regarded as productive. The underlying idea seems to be to distinguish between items which are instrumental in the sense that they contribute to the provision of other goods and services—either in the government or the private sector—and items which represent final goods and services. The instrumental goods and services contribute to the production of national income as ordinarily measured; the final goods cannot be associated with the production of any particular part of future national income. Some items can be classified fairly easily on this basis, but many government expenditures cannot be so classified.

Self-liquidating feature as a criterion

A somewhat related but less ambiguous basis of classification is to assign only self-liquidating items to the capital budget. The clearest case of a self-liquidating project is one for which sales of services to the private sector are expected to return initial and operating costs over a period of years. Examples are toll roads, irrigation projects, rental housing, and the like. On the financial side, a loan is an example.

A self-liquidating program may be financed by earmarked user taxes. Taxes on motor fuels, for example, are often assigned to road construction. In principle, this arrangement is similar to exacting a price for services, the user tax being considered more convenient to collect than tolls.18 If the earmarked taxes are not specifically related to use, the program is not self-liquidating in a meaningful sense; it has merely been assigned a preferred position in the budget.

A government project that provides to other government agencies services that would otherwise have to be bought from the private sector may be regarded as self-liquidating, particularly if formal arrangements are made for the using agencies to remit regular payments for the value of services received. The most usual example of this type of project is a building erected by one government agency or state enterprise and occupied by other government agencies.

The Swedish practice, for example, is to charge to capital account only that part of expenditures which is self-liquidating (“remunerative”) by virtue of charges collected from the private sector or from other government agencies. For items that are partly self-liquidating, an immediate write-off against revenue account is taken for the portion of the cost which is not expected to be recovered. It appears that the same principles underlie the Ecuadoran system.

In a looser sense, self-liquidating expenditures are sometimes said to include outlays which are expected to give rise to an increase in the tax base and in revenues sufficient to recoup the original outlay over a period of years. The increase in the tax base would be due to an increase in private income produced with the aid of the publicly provided services or capital. It should be recognized, however, that in order to be self-liquidating by this standard an investment would have to bring about a very large increase in national income, relative to its cost, inasmuch as marginal tax rates on income as a whole are always well below 100 per cent. The identification of any investments satisfying this requirement would be very difficult, and it seems preferable to restrict the term “self-liquidating” to projects involving service charges, prices, or user taxes.

Capital receipts

The most usual receipts items in capital budgets are (1) surpluses transferred from the revenue budget, (2) proceeds of borrowing, (3) proceeds of sales of government property and repayment of loans, (4) depreciation allowances on fixed assets, representing transfers from the revenue budget, and (5) regular appropriations from the revenue budget.19

It is sometimes suggested that taxes which are paid out of private capital should be assigned to the capital budget. Although all taxes may come out of potential private savings to some extent and thus in a sense represent a transfer of capital to the government, death duties and capital levies are often considered to have this characteristic to a special degree and hence to be properly assignable to the capital budget. Denmark and the Union of South Africa, for example, assign receipts from death duties to the capital budget. The arrangement helps assure that the taxes in question will not reduce the total capital stock. The objective may imply either a social priority for capital formation as distinguished from consumption, or a desire to avoid interference with private choices regarding the division of the national product between consumption and investment.

In Ecuador taxes on gasoline and international telecommunications, as well as a school stamp tax and a portion of customs receipts, are allocated directly to the capital budget.

Implications for Aggregate Demand and Resource Allocation

Government capital budgeting has implications for fiscal policy and for aggregate demand, mainly because of the presumption that expenditures segregated in a separate capital budget will normally be financed in large part by borrowing. Special reliance on loan finance is not a necessary feature of the capital budget. A United Nations publication, which is generally favorable to the dual budget, states:

… asset acquisition, taken by itself, is not an adequate justification for government borrowing, and a separation of current and capital accounts should not be undertaken for the purpose of rationalizing government debt creation. Rather, the separation of current and capital account transactions should be undertaken where it will facilitate proper management of government affairs and where it will contribute to an analysis of the economic significance of government activities. Such a classification should, however, never be taken as a fiscal policy guide, since the choice between financing from taxes or from borrowings must be based on purely economic considerations.20

The practices of most countries that maintain a separate capital budget, as distinguished from a purely informational tabulation of government investments, strongly suggest, however, that the arrangement ordinarily reflects the attitude that loan finance is more acceptable for capital outlays than for operating expenditures. Several of the considerations advanced in favor of capital budgeting imply that borrowing is a peculiarly appropriate means of financing capital outlays and should not be viewed in the same light as a deficit on current account. To be sure, governments often make a special effort to realize a current account surplus in order to cover a part of capital outlays during inflationary periods. Under a dual budget system, however, attention tends to be focused on the balance between revenue and expenditure in the operations budget, and the attempt to maintain a current account surplus often meets opposition. For example, the Minister of Finance of the Union of South Africa, in presenting the 1954–55 budget, took note of criticisms directed at the Government “for taxing the people in order to finance the loan programme [i.e., capital outlays]” and found it necessary to explain the considerations underlying the Government’s policy.21

That the rule of loan finance for the self-liquidating portion of capital expenditures has the support of at least one eminent economist is shown by the following quotation from Professor Pigou:

… consider government expenditure devoted to producing capital equipment—a national electricity plant, municipal gas works or tramways and so on—the fruits of which will subsequently be sold to purchasers for fees. Here it is generally agreed that the required funds ought to be raised by loans. If this is done, violent and sudden changes in tax rates are avoided: the people who benefit from the service which the new capital equipment renders pay for it in proportion to their use; and, provided, of course, that the fees charged are sufficient to wipe out the principal of the loan during the lifetime of the capital equipment, no additional taxation has ever to be raised on account of it. Upon this matter, as upon that of the proper method of finance for normal recurrent expenditure, there is no room for controversy.22

Critics of the capital budget argue that the appropriate relation between total government expenditures and taxation and borrowing should be determined in the light of general economic and monetary conditions rather than by reference to the composition of government expenditures. They emphasize that borrowing is ordinarily less anti-inflationary than taxing and conclude that a dual budget system, which relies on borrowing to cover a large part of capital expenditures, is likely to have an inflationary bias. As a subsidiary argument, some critics contend that a dual budget system, involving a mixture of cash and accrual items as well as transfers between the capital and current accounts, has the disadvantage of obscuring the relation between government income and outgo.

Although the proposition that borrowing is less deflationary than taxation is generally valid, some types of borrowing may approach taxation in this respect. The sale of long-term securities to the general public is usually the most deflationary type of borrowing. Borrowing from the commercial banks may also displace a considerable amount of private expenditures, particularly private investment, when the banks lack reserves in excess of the legal or customary minimum. Borrowing from the central bank, on the other hand, seldom decreases the availability of finance for the private sector and may even increase it by adding to commercial bank reserves.

The borrowing-spending combination will nearly always be more expansionary than an equal amount of taxing and spending, for two reasons: First, even when borrowing reduces private liquidity (or disposable funds) as much as taxation, it does not reduce private assets as taxation does. To the extent that private expenditures are a function of wealth as well as of liquidity or disposable income, private demand will remain higher with borrowing. Second, lending to the government is usually voluntary (although compulsory lending has occasionally been employed)23 and lenders are likely to revise their private spending plans less extensively as a result of the transaction than are taxpayers, who are compelled to contribute to the support of government.

Opportunities for borrowing from the public are usually more limited in underdeveloped countries than in wealthy countries, because savings are low and financial institutions are not well developed. Inasmuch as many underdeveloped countries already exhibit a tendency toward chronic inflation, a government capital budget—by justifying increased reliance on loan finance, which in the circumstances is likely to come from an expansion of central bank or commercial bank credit—may have unfortunate consequences. In some countries, to be sure, observation of the rule that borrowing may equal, but not exceed, government net investment would call for less reliance on loan finance than has been customary. But in countries strongly inclined toward deficit spending, the prospects seem dim for establishing any simple budget balancing rule, whether related to a unified or to a dual budget.

When aggregate demand tends to fall below the full employment level, it is generally agreed that not all government expenditures need be met by taxation. Under these conditions, financing either capital outlays or current expenditures by borrowing may be helpful. It will, however, be only by accident that the amount of capital expenditure will be exactly equal to the gap between revenues and total expenditures that is appropriate to help ensure the maintenance of adequate total demand. Furthermore, the usual capital budget principles relate the maturity of the debt to the expected length of life of the asset and may suggest long-term borrowing at a time when less deflationary short-term borrowing would be preferable from the monetary viewpoint.

Under potentially inflationary conditions, it may not be feasible to cover all expenditures by taxation, and the government may find it advisable to resort to a comparatively deflationary type of borrowing, such as the sale of long-term securities to the general public. The size of the current expenditures, however, has only a political relation to the amount of taxation that can safely be raised, and the size of capital outlays is a poor measure of the need for deflationary borrowing.

It is possible to conduct a compensatory fiscal policy, designed to maintain an appropriate level of demand, within the framework of a dual budget. But decisions with respect to taxation and different types of borrowing would have to be made primarily by reference to the total of government expenditures and to conditions in the private sector. It is hard to avoid the conclusion that the division of the budget into current and capital accounts is to some extent an obstacle to compensatory policy. In particular, a dual budget system is likely to complicate the stabilization problem during periods of incipient or actual inflation.

The importance to be attached to the weakness of the divided budget from the standpoint of fiscal policy turns, to a considerable extent, on a judgment concerning the frequency with which deliberate compensatory fiscal actions will be necessary. If market forces plus monetary policy and built-in flexibility of government revenues and expenditures are considered strong enough to bring about reasonable stability and high employment most of the time, the advantages of capital budgeting for resource allocation and administrative management may be given greater weight than the disadvantages from the point of view of countercyclical adjustments. If, however, very frequent countercyclical adjustments are expected to be required, the opposite evaluation may be reached. Furthermore, capital budgeting is likely to be considered more desirable by those who believe that an expansionary bias in the fiscal system is acceptable as a means of offsetting secular tendencies toward inadequate demand than by those who foresee a tendency toward secular inflation.

Measurement Problems

Given a dual budget system, what measures of government receipts, expenditures, and surplus or deficit should be selected for use in an analysis intended to shed light on the influence of government finance on aggregate demand and the monetary situation?

For this purpose, it is clear that comprehensive figures covering both current and capital accounts should be used. Regardless of their effect on the government’s balance sheet, many capital expenditures and receipts have the same influence on aggregate demand and the monetary situation as current transactions. Some governments that maintain capital budgets recognize this and provide consolidated statements as a basis for discussion of fiscal and monetary policies. The Netherlands, for example, publishes a consolidated statement of capital and current receipts and expenditures as well as an analysis of cash transactions.

Suitable figures cannot be produced simply by adding together the totals of receipts and expenditures in the operations and capital budgets. It is necessary in most cases to make adjustments to eliminate double counting, intragovernmental transactions, and certain entries of a purely bookkeeping nature. In this category are two items that often bulk large in dual budget systems: depreciation accruals on government property and transfers between the operations budget and the capital budget. These items are ordinarily entered as expenditures in the operations budget and receipts in the capital budget. Computing government expenditures and receipts by merely adding together the operations and capital budgets will therefore overstate the totals but will not affect the balance between expenditures and receipts. A similar item that is sometimes involved in the dual budget system is rental payments by government agencies for the use of buildings and other facilities owned by other government agencies.

Another phase of the preparation of comprehensive measures of government transactions is the treatment of state enterprises and trading departments. Owing to diversity of practices, this subject is especially important in international comparisons. Generally, it seems desirable to exclude from the consolidated statement of government receipts and expenditures the accounts of state-owned enterprises which derive their current receipts primarily from transactions with the public and which are accorded a large degree of autonomy from detailed treasury control of their operations. Railway and postal, telephone, and telegraph systems often fall in this category. In these cases the general government accounts would include in expenditures purchases of goods and services from the enterprise, capital subscriptions to it, and subsidies to cover losses; general government receipts would include dividends and interest paid to the treasury by the enterprise. Accounts of enterprises which deal mainly with government departments and which lack financial autonomy may properly be consolidated with general government accounts. Examples of such enterprises are organizations which construct buildings and lease them to government departments and various supply agencies. Admittedly, the application of these criteria leaves much room for differences of opinion. However, inasmuch as the subject is only marginally related to capital budgeting, it will not be considered in further detail here.

Certain items which are not related to the existence of a capital budget should be eliminated from the data if they have been included. The objective is to eliminate all entries which do not represent transactions between the consolidated government accounts, on the one hand, and the general public, including government entities which are not included in the consolidation, on the other. There are two general types of entry to be eliminated: (1) entries representing transactions between government accounts included in the consolidation—for example, transactions between the treasury and special accounts, such as social security funds—which inflate equally government revenue and expenditure (and therefore do not affect the surplus or deficit); (2) entries that affect the size of the surplus or deficit but do not represent cash transactions of the government with the public—for example, on the receipts side, profits from revaluation of government holdings of gold and foreign exchange, and, on the expenditures side, accrued but unpaid interest.

The treatment of public debt transactions and of additions to, and withdrawals from, reserve funds raises the question of the definition of the surplus or deficit. In actual usage, different definitions are employed, and the terms may have value-judgment implications. The predominant usage, however, assumes the elimination from government receipts of proceeds of borrowing and withdrawals from reserve funds, and the elimination from expenditures of repayments of the principal of debt and additions to reserve funds. The surplus or deficit is then taken to be equal to the difference between combined current and capital receipts and expenditures as adjusted.24

It may be questioned whether the surplus or deficit so defined is the most useful single measure of the contractionary or expansionary influence of the government’s finances. The chief drawbacks are the failure of the figure to distinguish among receipt and expenditure items having rather different impacts, and the lumping together of the use of cash balances and all kinds of borrowing under “deficit” (and all kinds of debt repayments and additions to cash balances under “surplus”). For purposes of monetary analysis, some authorities would prefer to draw the line at a different point and to assimilate proceeds of long-term borrowing with taxation, and repayment of long-term debt with other expenditures.25 They reason that long-term debt transactions, like taxes and expenditures, change the liquidity of the public and thus have a much greater influence on private consumption and investment than do short-term debt operations, which involve merely changes in the form of liquid assets. According to this view, the surplus or deficit should be a measure of the change in the liquidity of the private sector attributable to the government’s budgetary policy. The surplus or deficit would equal the change in the government’s short-term debt and cash balances.26

In order to minimize confusion, however, it seems best, particularly when writing for an international audience, to employ the broader definitions of surplus and deficit and to equate this item with the change in the government’s total debt and cash balances (after suitable adjustments to eliminate noncash items). It may be helpful to characterize the figure as the “cash” surplus or deficit in order to distinguish it from other concepts. In some instances, the colorless and somewhat awkward phrases “excess of receipts” and “excess of expenditures” may be preferable. These phrases have the advantage of being comparatively unambiguous, and the preparation of the figures involves fewer questions of judgment, on which opinions will differ, than does the preparation of figures based on alternative concepts.

The principal adjustments that may be necessary to consolidate the operations and capital budgets may be summarized as follows:27

1. Receipts

  • (a) Operations budget receipts Subtract:

    • (i) Any proceeds of borrowing or withdrawals from reserve funds which may be included

    • (ii) Noncash and intragovernmental items

  • (b) Capital budget receipts Subtract:

    • (i) Depreciation accruals as provided in operations budget

    • (ii) Other transfers from operations budget, including initial capital subscriptions, reimbursements for capital losses, etc.

    • (iii) Proceeds of borrowing

    • (iv) Withdrawals from reserves and balances from prior years

  • (c) Adjusted total receipts

2. Expenditures

  • (a) Operations budget expenditures Subtract:

    • (i) Depreciation accruals for government departments

    • (ii) Other transfers to capital budget

    • (iii) Repayment of debt

    • (iv) Additions to reserve funds

    • (v) Other noncash and intragovernmental items

  • (b) Capital budget expenditures Subtract:

    • (i) Repayment of debt

    • (ii) Additions to reserve funds

  • (c) Adjusted total expenditures

3. Cash surplus or deficit = item 1(c) minus item 2(c)

The cash surplus or deficit derived in this manner should be reconcilable with figures on changes in cash balances and in the public debt. In addition to making such a reconciliation, it is often helpful to introduce more detailed figures on the disposition of the cash surplus or the financing of the cash deficit, as well as other quantitative or qualitative information that may aid in the interpretation of bare summary figures.

A distinction between long-term and short-term debt transactions may be especially illuminating when the country’s monetary policy maintains a significant difference in the liquidity of the two kinds of securities. The distinction may not be meaningful, however, when the monetary authorities are vigorously supporting the price of long-term securities in order to prevent a rise in interest rates or when the government itself stands ready to redeem nominally long-term securities. These practices prevailed in a number of countries during and immediately after the war, and they tended to reduce or obliterate the difference between long-term and short-term debt.

Another classification of global figures that may be enlightening is that of government expenditures into (1) expenditures for newly produced goods and services, (2) purchases of existing physical and financial assets, and (3) transfer payments. Item (1) represents the direct exercise by the government of a claim on national output and constitutes in itself a part of aggregate demand. The other two items affect demand for the current final product only insofar as recipients of government payments use the funds to buy goods and services. Expenditures for newly produced goods and services are therefore usually more expansionary than purchases of existing assets and transfer payments.

In the present context, it is worth emphasizing that none of the classifications of financing items or of expenditures that are likely to be useful for measuring the impact of government finances on aggregate demand and monetary conditions will necessarily parallel the distinction between the operations and the capital budget. In some instances, capital budget receipts and expenditures will include long-term but not short-term debt transactions, but even in these cases other adjustments to current and capital accounts are likely to be necessary to develop the most meaningful summary figures.


Mr. Goode, Chief of the Finance Division, was educated at Baylor University, the University of Kentucky, and the University of Wisconsin. Formerly economist at the U.S. Bureau of the Budget and the U.S. Treasury and a member of the faculty of the University of Chicago, he is the author of several articles in various economic journals.

Mr. Birnbaum, economist in the Finance Division, was educated at Ohio State University and the George Washington University. As a member of the Fund’s Statistics Division, he was formerly responsible for the statistics on government finance published in International Financial Statistics.


In its original form this paper included a series of notes describing the budget and accounting systems of Sweden, the Netherlands, Ecuador, the Union of South Africa, India, the United Kingdom, Canada, and the United States. The notes also contain some discussion of the significance that appears to be attached in official policy statements to the form of the budget. Mimeographed copies of the notes may be obtained from the Secretary, International Monetary Fund, Washington 25, D.C.


State trading or industrial enterprises usually employ commercial accounting methods, involving capital accounts, regardless of whether a separate capital budget is maintained by the administrative departments of the general government.


The terms “operations budget,” “current budget,” and “revenue budget” are often used interchangeably, and this practice is followed in the present paper.


An analogous procedure has been followed for power-generating facilities. The full construction costs of many projects, however, are still included in the budget when incurred.


For a discussion arguing that commercial accounting principles have only a limited relevance to government accounting, see Committee on the Form of Government Accounts, Final Report (London, Cmd. 7969, June 1950), pp. 74–76.


The analogy between a government and a household, of course, should not be pressed too far. Decisions concerning the amount of economic resources over which the government will attempt to gain command in any period are based on considerations and opportunities fundamentally different from those presented to a household.


This procedure involves asset accounting, but it carries no necessary implication that the capitalized cost exactly represents the “value” of the item or the future benefits from it.


Strictly speaking, interest costs on invested capital or a time discount for future benefits should also be imputed, but this refinement has usually been neglected.


Provided annual outlays for new capital had been constant for a period of years equal to the average depreciable life of the assets.


For preparation of a comprehensive estimate of net capital formation in the government sector, it would be necessary to take a complete inventory of state property at the time when a capital budget was adopted, and to make estimates of annual depreciation of the old items as well as to recognize the depreciation allowances set up for new assets.


This point has been stressed by James A. Maxwell in “The Capital Budget,” The Quarterly Journal of Economics (Cambridge, Mass.), Vol. 57 (1942–43), pp. 450–65.


As often stated, the rule holds that the maturity of the debt should ordinarily be approximately equal to the expected life of the assets. For many capital investments, this implies long-term borrowing.


The method of financing, of course, may have some future effects on capital supply and productivity.


To be sure, the policy involves proportionately large changes in private investment, but the division of national product between total consumption and total investment (including private and government items in both magnitudes) may be regarded as more fundamental.


Short-term securities are more liquid than long-term securities; hence their sale to the public and the banks is likely to bring about a smaller contraction of private expenditures.


United Nations, Department of Economic Affairs, Manual for the Classification of Government Accounts (New York, preliminary, July 1954), p. 127.


When stocks are acquired in connection with a government price-support program, their value may be somewhat problematical.


Provision of road services could be regarded as analogous to a government enterprise with user tax receipts being assigned directly to the road administration and the general government accounts reflecting only transfers between the treasury and the road administration; or receipts from user taxes could be first entered in general receipts and then transferred to the road administration on the assumption that they are roughly equivalent to amortization and maintenance of roads.


The first two items are often considered means of financing a deficit in the capital budget rather than capital budget receipts in the strict sense.


United Nations, Department of Economic Affairs, Budgetary Structure and Classification of Government Accounts (New York, February 1951), p. 12.


The Minister stated that increased reliance on taxation was justified by the greater maturity of the economy and suggested that, in deciding the extent to which capital outlays should be covered by current revenue, the country should be guided by the availability of voluntary private savings and the degree of inflationary pressure (Union of South Africa, Budget Statements, 1954–55 (Cape Town, 1954), pp. 12–13).


A. C. Pigou, A Study in Public Finance (London, 3rd ed., 1949), p. 36. Pigouconcedes that loan finance may also be appropriate for certain other nonrecurrentor irregular expenditures, but he does not suggest that any special attention should be given to the question whether they are capital items.


A peacetime example is the system introduced in the Union of South Africain the 1953–54 budget and abolished in the 1954–55 budget.


This is the practice followed by the International Monetary Fund in its International Financial Statistics, and by the United Nations in the compilation of statistics on “cash operations” published in its Statistical Yearbook. In the latter source, the difference between receipts and expenditures is called “balance.”


See, for example, Nederlandsche Bank, Report for the Year 1953 (Amsterdam, 1954), pp. 58–59, and Government of India, Planning Commission, The First Five Year Plan (New Delhi, 1952), pp. 59–60.


For a discussion favoring this definition over others, especially for underdeveloped countries, see K. N. Raj, “Definition and Measurement of Deficit Financing,” Indian Economic Review (Delhi), August 1954, pp. 34–51.


It is not intended to deal comprehensively here with questions relating to the scope of government budget accounts; hence no suggestions are offered on such questions as the adjustments necessary to incorporate properly the transactions of social security funds or other funds frequently unconsolidated in government budget accounts.