The broadest definition—total domestic borrowing

If it is assumed that (a) all domestic receipts other than borrowing curtail purchasing power and demand, (b) all expenditure adds to aggregate demand, (c) domestic borrowing has no impact either way, and (d) foreign receipts make available additional real resources for the economy, then a budget balance reflecting the stimulatory/contractionary impact on aggregate demand may be derived from the difference between total expenditure, on the one hand, and the sum of domestic “current” receipts and foreign receipts, on the other hand. That is, the expansionary deficit

In a closed economy or in a country where foreign borrowing by the government is nonexistent, negligible, or relatively small, the expansionary deficit as just defined can obviously be equated with the overall deficit, which is often used as a first approximation in attempts at measuring the impact of the budget. However, in countries where government borrowing from abroad forms a significant proportion of total borrowing, the overall deficit will not be an accurate measure of the net impact of the budget on aggregate demand.

The impact of budgetary transactions involving foreign receipts or expenditure varies according to the types of receipt and expenditure that are combined in given transactions. Four different combinations can be envisaged: ²⁰ (a) Foreign receipts, such as grants from abroad, combined with foreign expenditure—net impact is neutral with respect to both GNP and the balance of payments. (b) Foreign receipts combined with domestic expenditure—an expansionary impact on GNP but a countervailing favorable impact on the balance of payments. (c) Domestic receipts other than borrowing combined with foreign expenditure—a contractionary impact on GNP but an adverse impact on the balance of payments, (d) Domestic borrowing combined with foreign expenditure—a nil impact on GNP but an adverse impact on the balance of payments.

As indicated earlier, in the context of the use of the simple (unweighted) budget balance, an expansionary impact is defined in terms of an increase in the demand for domestic goods and of the impact on the balance of payments. Given this definition, the net impact of budgetary transactions can only be obtained by combining the effects on GNP and on the balance of payments. Thus, not only (a) but also (b) and (c) are examples of neutral impact, while (d) is a case of expansionary impact. Since expenditure abroad is added to domestic expenditure in computing the unweighted budget balance, foreign receipts, all of which have a favorable impact on the balance of payments, need to be placed “above the line” in the derivation of the expansionary deficit. As regards the immediate impact on aggregate demand, there is no difference between foreign grants and foreign loans. However, in a longer-term assessment of the overall financial position of the government, a distinction must be made between grants and loans from abroad. For, unlike foreign grants, borrowing from abroad adds to liabilities, giving rise to problems of future interest burden and of transformation of domestic savings involved in repayment of loans in foreign currency. In other words, the favorable impact of borrowing from abroad on the balance of payments may be only temporary, unless the expenditure of the loans were to produce a more lasting effect (in terms of productivity and exports).

Treatment of government lending

In calculating the expansionary deficit (on the basis of the broad definition), domestic lending has been included in expenditure (placed above the line) and government borrowing from the nonbank private sector has been excluded from the receipts side (placed below the line). This procedure implies the assumptions that government lending, or net acquisition of financial assets by the government, increases aggregate demand in the same way as any other expenditure, and that government borrowing from the nonbank private sector has no effect on private consumption or investment. As discussion in the literature has shown, these propositions are subject to a considerable degree of uncertainty, and hence the need to suggest alternatives to the broad definition given earlier as the point of departure for discussion.

The assumption with regard to government lending is taken up first, and the treatment of government borrowing—which is, in a sense, a related question—is discussed next.

The discussion is confined to domestic lending. If lending abroad by a developed country is largely in the form of tied aid, it really should be treated as part of domestic expenditure. Lending abroad is hardly significant in developing countries.

The treatment of domestic lending in the derivation of the budget balance has been the subject of considerable discussion in the literature.²¹ Only the main arguments are recapitulated here. These revolve around the question whether in lending (or relending) money, the government is merely engaging itself in financial intermediation or whether the intervention of the government in the capital market gives rise to net additional expenditure in the private sector. On the one hand, those who favor the inclusion of lending in expenditure argue that loan programs establish claims on real resources that are hard to distinguish from those that arise from other types of expenditure, such as for grants or subsidies. On the other hand, those who wish to exclude lending argue that payments that affect only liabilities and assets of individuals should not be treated on the same footing as income-creating expenditure, to which indeed they are not additive, and that in any case lending represents mostly financial intermediation, that is, lending of borrowed money. Two different questions seem to be at issue here. First, is financial investment less demand creating than an equivalent amount of direct investment by the government? Second, do not the effects of borrowing neutralize the effects of lending, or, in other words, is there justification for treating lending to and borrowing from the same sector asymmetrically? These two questions may be considered separately to avoid confusing the results of one operation with those of the other.

Acquisition of financial assets by the government can take many forms, ranging from the repurchase of existing government securities to lending money to potential investors who are held back for want of funds (the so-called submarginal borrowers). Some forms of acquisition of financial assets are likely to be much less demand creating than others. A good example is that of buying existing government or private securities. Such expenditure, the direct impact of which is to change the asset portfolio of the private sector, may be considered to create very little additional demand, unlike direct investment by the government.²² There is a good case for excluding that kind of financial investment from expenditure.²³ At the other extreme, it could be argued that lending to a submarginal borrower would most likely give rise to a net increase in total demand almost equal to the amount of lending because, in the absence of the government loan, the submarginal borrower without access to the private credit market would have been unable to undertake the investment (now made possible by the loan); it could also be argued that since he is outside the private credit market anyway, others dealing in it are not affected by his actions in either case. In between the two extremes would come loans to supramarginal borrowers whose investment in real assets would probably not increase by the full amount of the loans. For example, if the government lends $100 to a supramarginal borrower who would have obtained $75 from the private credit market in the absence of government action, the net increase in his expenditure would be only $25. However, his withdrawal from the market releases $75 for use by other people. In other words, nonmarginal loans ease conditions in the credit market and put downward pressure on interest rates, thereby enabling some submarginal borrowers to become marginal borrowers.²⁴ Thus, even though the direct increase in real investment is only $25, the total effect flowing from the loan is likely to be much larger, unless the government siphons off the unused credit by some means.

In developing countries, government loans are often given to sub-marginal borrowers, such as small-scale farmers and producers, and to promote investment that would not have taken place otherwise. Typically, the credit market is undeveloped, and large investments would generally be extremely difficult to undertake without the government’s financial support. Furthermore, if the question (of the effects of government lending) is considered not in terms of comparative statics but in the context of growth, it is likely that the market would have adjusted itself to a certain proportional magnitude of government lending in every (annual) period. The stoppage of lending in one period would introduce a deflationary tendency. The continuation of government lending programs may be said to contribute to the growth in aggregate demand. These considerations suggest that there is a strong case for including government lending for new investment as part of expenditure in analyzing the impact of the budget in developing countries. It is true that different lending programs may have greater or smaller demand effects, but since in computing the simple budget balance all items included above the line are to be taken as being equivalent, these differences in impact cannot be taken into account. The only possibility is to include items that are believed to have sufficient effect to warrant inclusion and to exclude others. As already noted, acquisiion of financial assets representing the purchase of existing assets may be excluded. Also, in the French-type treasury system, where the treasury engages in banking activities, lending under these operations should be excluded from expenditure.

In national income accounting and flow-of-funds analysis, real investment is assigned to the sector that undertakes it and not to the one that finances it. Hence, if government lending is added to government expenditure for fiscal analysis, the concerned magnitude should be taken out of the figure of private investment in an overall analysis of the economy in order to avoid double counting.

Treatment of government borrowing

In the broadest definition of expansionary deficit (given earlier), all domestic borrowing has been included in the deficit with the implication that no act of borrowing has a negative impact on demand. It is, of course, granted that such an assumption is not, strictly speaking, equally valid for all types of borrowing, but it is believed by those who adopt this definition that there is a clear difference as regards impact on demand between taxation, on the one hand, and any type of borrowing, on the other hand. However, it seems necessary, at least in respect of the developing countries, that a distinction should be made between borrowing that reduces purchasing power in the hands of the public and that which does not. Let us consider the three major types of borrowing in turn.

Borrowing from the central bank. Borrowing from the central bank may be said to have no direct impact on aggregate demand, because the central bank is not under a compelling necessity to reduce credit elsewhere when it extends credit to the government.²⁵ From this it follows that domestic expenditure combined with central bank borrowing has an expansionary impact on aggregate demand. Demand may be expected to increase not only because of the increase in income but also because of the possible effect of the additional liquidity on consumption through the operation of the Pigou effect. The increase in money supply implicit in this means of financing may turn out to be more than the increase in desired cash balances resulting from the expansion in income that was created by the increase in government expenditure. This is especially likely because part of the newly created cash paid by the government to the public would be deposited with the commercial banks, the credit multiplier would operate, and the increase in aggregate demand would be larger than was originally caused by government expenditure as such. In such a case, the attempt on the part of the public to get rid of the excess cash balances would lead to a rise in the price of income-yielding assets and a downward pressure on interest rates, which in turn might induce additional private investment and/or a balance of payments deficit.

These are the possible results if the problem is viewed in terms of comparative statics. Let us now introduce growth in real income and consider the effects of government expenditure financed by borrowing from the central bank when the economy is growing at a given percentage rate. The borrowing has no negative impact on demand, and hence the domestic expenditure financed by it would have an expansionary effect on aggregate demand as in the comparative static situation. But, whereas in the latter situation if the initial state is one of formal full-employment equilibrium ²⁶ there will be a rise in prices accompanied by an increase in money incomes only, in a situation where growth is allowed for, the additional demand created by credit-financed government expenditure may be met by the increase in real output and hence there need be no inflationary rise in prices. Nevertheless, the effect of credit-financed domestic expenditure is expansionary in both cases in terms of incomes directly generated. But the results are different in terms of the effects of additional liquidity created. Whereas under comparative statics it has been shown that the increase in money supply might produce both the Pigou effect on consumption and the Keynesian effect on the interest rate, in the dynamic situation, with the growth in real income and progressive monetization of the economy, the demand for cash balances will have increased, and the growth in money supply represented by the credit extended by the central bank (if less than the increase in demand) would go toward meeting part of this increase in demand and would not then cause a rise in the price of assets as under comparative statics. The increase in the demand for cash by the public may also be looked upon as the desire to hold a certain proportion of savings in the form of liquid assets, and the government then pre-empts part of these savings for its own use through credit creation by the central bank.²⁷ Insofar as the increase in cash balances resulting from government use of central bank credit is limited to a part of the increase in wealth desired to be held liquid, no Pigou effect could be attributed to government action. Nor will there be an inducement to get rid of cash balances, because there is an increase in demand to which, it is assumed, the increase in supply is limited.

To conclude, domestic expenditure financed by central bank credit is expansionary. In the framework of comparative statics, borrowing from the central bank will not have a negative impact on aggregate demand; instead, it may induce an increase in aggregate demand through the effects of the additional money stock created. In the context of growth in real income, domestic expenditure financed by central bank credit will probably be less expansionary, because the increase in money supply will be counterbalanced, at least in part, by the increase in the demand for cash balances.

If borrowing from the central bank is used to finance government expenditure abroad, the expansionary impact is reflected in the adverse effect on the balance of payments. In this case, there are no additional effects created by a change in liquidity conditions.

Borrowing from commercial banks. Government borrowing from the commercial banks will have no impact on aggregate demand if the banks have excess reserves, and therefore domestic expenditure financed by such borrowing will have a direct expansionary effect in exactly the same way as expenditure financed by central bank credit.²⁸ If the commercial banks do not have excess reserves when the government offers them its bonds, they can lend to the government only by curtailing credit to the private sector. This would exert a contemporary pressure offsetting the expansionary effect of the additional government expenditure.²⁹ However, instead of curtailing credit to the private sector, the banks may turn to the central bank for accommodation. If the central bank extends credit to the commercial banks under such circumstances, the result would tend to be the same as if the government had been granted the credit direct—except that the cost of acquiring credit by the commercial banks would probably result in less expansion. It is possible to adopt a convention by which government borrowing from the commercial banks can be treated as borrowing from the central bank, to the extent that there is extension of credit to the banks at the same time. Even then, whenever the extension of credit by the central bank to the commercial banks is less than the amount of government borrowing from the commercial banks, it would still be necessary to ascertain their excess reserves position. This will not be easy to identify, because, in practice, it is not easy on the basis of ex post data to distinguish between an autonomous fall in the credit extended to the private sector and deliberate curtailment of credit by the banks themselves.

Net borrowing from the commercial banks continuously (in successive periods) will be possible only if the government or the central bank continues to create high-powered money.³⁰ If real income is rising from period to period, the demand for cash balances rises and the creation of high-powered money provides the basis on which bank credit can expand. Thus, in the context of growth, commercial bank credit to the government can also be looked upon as the means by which part of the savings of the public is diverted to the government. This is the part that the public is willing to hold in the form of bank deposits, and the amount that can be diverted without a rise in prices depends both on the proportion of savings that the public wishes to hold as bank deposits and on the rival claims of the private sector for bank credit. But even if government borrowing is limited in every period to this amount, expenditure financed by it acts as an expansionary force. On a balance of considerations, therefore, for practical purposes, commercial bank credit to the government may be treated as having no negative effect on private demand. However, a situation may exist to which this general conclusion does not apply; for example, during a given period the banks might have curtailed credit to the private sector in order to subscribe to government securities in the absence of accommodating action by the central bank.

Expenditure financed by central bank credit may be considered to be potentially more expansionary than that financed by commercial bank credit, insofar as the former adds to the reserves of the banking system. However, we are dealing here only with the direct budgetary impact.

Borrowing from the nonbank private sector. In most developing countries, borrowing from the nonbank private sector, that is, the public, will have a contractionary effect on private demand under several plausible circumstances. Hence, great importance is attached in those countries to efforts at mobilizing private saving for economic development through such means as the establishment of a government savings bank and the issue of savings certificates suited to the small investors. Diversion of funds through these channels is looked upon as the utilization of genuine voluntary savings as opposed to the use of bank credit and is considered, therefore, a noninflationary means of financing economic development. Borrowing from the public, however, will clearly have a less contractionary effect on private demand than taxation, in as much as the taxpayers feel that they are worse off for taxes paid, while bondholders may feel neither worse nor better off since they would have only exchanged one form of asset for another. Nevertheless, borrowing from the public in developing countries may have a significantly more contractionary effect than using bank credit to warrant consideration of excluding it from the measure of expansionary deficit. Again, this question must be considered both in terms of comparative statics and in the context of growth.

Complications in analyzing the effects of government borrowing from the public arise mainly because of the possible countervailing forces emanating from an elastic credit structure and from monetary policy actions. In order to isolate the effects of government borrowing, per se, let us assume a 100 per cent reserve system and no credit creation by the central bank. The analysis under these hypothetical conditions will be carried out only in terms of comparative statics.

(a) If the public were to lend to the government from idle balances (”kept under the mattress”), the borrowing would have no significant impact on aggregate demand. There might have to be a slight increase in the rate of interest to induce “dishoarding,” but unless private investment were to be quite interest elastic, the impact on private demand might be negligible. This source of subscription may not be discounted altogether, but it may not be of much significance except in cases of once-for-all, or infrequent, government borrowing, because the public cannot continuously buy financial assets out of past savings kept as cash.

(b) If the subscription is out of current income, the results will depend on the respective interest elasticities of savings and investment. Government borrowing will lead to a rise in the rate of interest. If savings are interest elastic but investment is not, only consumption will fall. If investment is interest elastic but savings are not, only private investment will fall. If both are interest elastic, both will decline, so that the expenditure financed by borrowing will draw resources from both private consumption and investment. In the realistic case, most of the diversion is likely to be from private investment. Then, if the government sterilized the proceeds of borrowing, a deflationary tendency would be created because private investment would have fallen and money supply would have been reduced.

The government competes with the private entrepreneurs for investible resources, and what the government is able to secure, someone else has to go without.

(c) If the public wishes to subscribe to government securities by selling some of the private securities that it is holding, the market rate of interest will rise and the same reasoning as in (b) applies.

In reality, both a fractional reserve system and the possibility of credit creation by the central bank exist. These must be allowed for. With these realistic assumptions, we shall also consider borrowing in the context of growth. The problem is looked at first in terms of comparative statics.

If no excess reserves exist in the banking system, government borrowing from the public, by itself, reduces aggregate demand (mostly investment demand), reserves of the banking system, and money supply. The effect will be contractionary. The same results would follow even if excess reserves existed, except that the potential private borrower who has been outbidden by the government may be able to obtain credit from the banking system and the effects of government borrowing may be partly neutralized by the expansion of commercial bank credit. If excess reserves do not exist but the central bank extends credit to the commercial banks so that they continue to meet the private demand for credit, government is, in effect, borrowing from the central bank. The result is almost the same if the central bank is required to give price support to government securities through open market operations. Whenever the central bank intervenes in this manner, either the borrowing should be treated as being from the central bank, or the expansionary effect of monetary policy should be set against the contractionary effect of borrowing.

It follows from the reasoning just given that expenditure financed by borrowing from the public may result in a net expansionary impulse if excess reserves exist in the banking system, or if the government follows an expansionary monetary policy.

Let us now introduce growth and government borrowing in successive periods. As stated earlier, with growth in real income and positive savings in every period, the public would wish to hold a part of its savings in the form of cash balances or demand deposits. If opportunities are provided, it would also wish to keep a part in government savings bank deposits and in different types of government securities. Through borrowing from the public, the government may be said to be diverting a part of such savings. In several developing countries, the nonbank contribution to government securities comes mainly from pension funds, social security contributions,³¹ insurance funds, and government savings institutions. The level of private investment can be expected to adjust itself to the government’s appropriating savings in these forms. And if in one year the government should suddenly sterilize these funds, a deflationary impact would most likely be created. Also, excess reserves cannot be expected to exist in the banking system year after year unless the central bank engages in continuous expansion of credit, or unless a continuing balance of payments surplus exists.

On the other hand, it should not be assumed that the borrowing of the government ex ante would always match the excess of private sector saving over intended private investment. An attempt could be made, in an overall financial plan for the economy, to make the level of government borrowing consistent with the projected level of private investment. But, since fluctuations in private investment (departing from the projected levels) could materialize through the existence of an elastic credit structure, even in the context of growth in income, government borrowing from the public cannot always be said to be a contractionary measure.

In principle, there is not much difference between the way in which monetary conditions and monetary policy modify the effects of borrowing and the way in which they modify the effects of taxation. In computing the unweighted budget balance, it is implicitly assumed that there is a one-to-one relationship between taxation and the reduction in aggregate demand. In fact, however, one dollar of taxation would reduce consumption only by a fraction of the dollar (equal to the marginal propensity to consume); the remainder of it is supposed to come from a reduction in private investment. But this need not happen if commercial banks have excess reserves, or if the central bank is willing to supply additional reserves to the commercial banks on easy terms.³² Again, if there is excess liquidity in the economy, additional taxation may not have the expected full effect even on consumption. Moreover, monetary policy actions, such as relaxing the terms (or reducing the cost) of the hire-purchase system, may partly neutralize the effects of taxation.

An alternative definition

The foregoing discussion of the effects of government lending and of expenditure financed by different types of borrowing seems to lead to the conclusions that in computing an unweighted measure of expansionary deficit (a) borrowing from the entire banking system should be included but (b) borrowing from the nonbank public may be included or excluded depending on circumstances and on the policy assumptions most appropriate to a given situation. A corollary of the inclusion of nonbank borrowing in the deficit is that it would ensure that all net lending is included in expenditure. Exclusion of nonbank borrowing would mean, in effect, that borrowing of that kind is regarded as an offset to lending to the nonbank private sector, where one exists, and that only lending out of bank borrowing is included in expenditure.³³ Total net domestic borrowing³⁴ probably indicates the upper range of estimate of expansionary impact, while net borrowing from the banking system gives the lower range of that estimate.

Industrial countries have a highly developed (and integrated) capital market and an elastic credit structure, with a clear dichotomy between savers and users of savings. In these circumstances, short-run levels of private investment are not constrained by the prior availability of savings, and it is possible for investment to be sustained on the basis of increased credit in the face of an increase in the government’s diversion of household sector savings. Given these institutional factors, there is a good case for including borrowing from the nonbank public in the measure of expansionary deficit in fiscal analysis for developed countries.³⁵ However, in regard to a typical developing country, the question may be raised whether borrowing from the banking system alone does not more closely approximate the desired estimate of the expansionary impact of the budget. The following considerations are relevant to the question and seem to support the choice of the lower-range estimate.

(1) Whereas situations of excess ex ante savings over forthcoming investment are common in developed capitalist economies, a shortage of savings in relation to investment is more likely in less developed countries. Thus, it can be said that whereas in the developed countries there exists a situation of closer approximation to the Keynesian framework of investment “determining” the level of income and savings, in the developing countries it is a situation of closer approximation to the classical framework of investment adjusting to changes in available savings. Since, in addition, the credit structure is poorly developed or undeveloped in the less developed countries, and the small farmers and other producers have no easy access to the banking system, if the government should divert a larger part of the savings of the households, it would not be so easy as in developed countries for the level of private investment to be sustained.

(2) It is believed that speculative demand for money is unimportant in most developing countries. If this is so, a rise in the rate of interest induced by the government selling bonds to the public may lead to no change in the demand for money and, hence, to no change in income velocity. In this situation, the monetarist’s argument that since money supply is unchanged when government expenditure is financed by non-bank borrowing there can be no increase in demand or income would acquire validity. Then, as income cannot increase, with a rise in government expenditure financed by nonbank borrowing, there must be a fall in private investment or consumption (or both), depending on their relative interest elasticities.

(3) In a typical developing country, the government undertakes a number of productive activities that are normally left to the private sector in the industrialized private enterprise economies. These activities necessitate the undertaking of industrial and other business investment that often supplants or replaces private investment, especially in view of the limited available foreign exchange that needs to be allocated.

(4) In some of the developing countries that do not have a social security system, there exists a provident fund system that entails compulsory contributions by employers and compulsory savings by employees in both the public and the organized private sectors. Investment in life insurance policies is normally promoted through concessions under the income-tax law, which substantially benefit mainly the taxpayers at the lower end of the taxable income scale. Almost the whole of the provident funds and a good proportion of the life insurance funds are invested in government securities.³⁶ In a developing country, most contributors to provident funds may have little of other savings, so that variations in the laws affecting provident funds, and perhaps in those offering tax concessions for insurance premiums,³⁷ will most likely affect consumption. The contributions by the employers are, of course, akin to taxes. It follows that the utilization of these funds by the government, that is, borrowing from the so-called captive market, may have a significantly contractionary impact.³⁸

In the light of the foregoing considerations, for purposes of fiscal analysis for developing countries, the use of net bank credit may be regarded, conceptually, as the more appropriate summary indicator of budgetary impact on aggregate demand. Thus, the alternative definition of expansionary deficit (giving the lower-range estimate) may be derived as follows:

In practice, however, it is hard to obtain a precise estimate of this magnitude because of the difficulties involved in separating out the amounts borrowed from the bank and nonbank sectors. During the course of the budget year, the public may resell to the banking system bonds taken up during the earlier part of the same year, or, conversely, the banks may resell to the public. Ex post, it is possible to look at the “net absorption” of government bonds by each of the three major nongovernment sectors: the central bank, the commercial banks, and the nonbank private sector. Moreover, the net increase in the combined claims of the three sectors can be expected to equal the net domestic borrowing of the government, if the market value of the “old” securities does not change. However, the net absorption of government securities by any one sector will not be equal to the net borrowing of the government from that sector, because the central bank might have been carrying out open market operations during the year; hence, if these operations have affected the holdings of the nonbank private sector, the estimate of net bank credit or increase in the net claims of the banking system against the government will reflect the impact of not just budget policy but of a mix of budgetary and monetary policy.³⁹ Besides, if the measurement of the increase in the net claims of the government were to be affected by changes in the market value of government bonds, the increase in the net claims would, of course, not equal the net domestic borrowing by the government.

To the extent that the capital market is undeveloped, that open market operations affecting the nonbank public are insignificant or nonexistent, and that the commercial banks do not have much scope for selling securities to the public, the estimate of increase in the net claims of the banking system against the government, derived from banking statistics, will approximate the amount of the government’s net borrowing from the banking system. Insofar as conditions in several less developed countries are like those just described, the government’s use of net bank credit could be employed as the basic measure of the direct expansionary impact of the budget. But wherever the nonbank private sector plays even a moderately important role in the capital market, it will be necessary also to give the total of net domestic borrowing by the government as an alternative upper-range estimate of the expansionary budget balance. For industrialized countries, total net domestic borrowing would be the more adequate basic measure of expansionary deficit; if foreign borrowing is not significant, the overall deficit itself may be used to reflect the impact of the budget.

The weighted budget balance

The use of the simple budget balance, however defined, as a measure of the impact of budget policy on demand is based on a number of simplifying assumptions, which generally are not made explicit. Some of the limitations of the use of this measure arise from these assumptions. First, there is the implicit assumption either that foreign transactions are not significant or that it is not necessary to distinguish between the impact of the budget on GNP and on the balance of payments. Subject to other limitations to be discussed later, a composite measure can be used as a broad measure of the immediate impact of a change in budgetary stance, but further (secondary, tertiary) repercussions on the economy will obviously differ according to the relative weights of the two components of expenditure. Second, it is obvious that the simple budget balance approach is based on an implicit model in which all items of expenditure and receipts that are included to compute the balance have equal (and one-to-one) impact on aggregate demand, negative and positive, respectively, and that those items included in the deficit have no direct impact on demand. In the earlier discussion, it has become apparent that different types of lending and borrowing tend to have an unequal impact on demand. This is also true of various forms of taxation and different kinds of expenditure. That is, quite apart from the differing weight of the impact on the balance of payments in the total demand impact, equivalent amounts of different items of revenue (or of expenditure) are likely to have unequal impact on GNP. For example, expenditure on goods and services and on transfer payments: the impact of the former on GNP plus imports is equal to the amount of expenditure itself, whereas that of the latter is reduced by the savings leakage. On the revenue side, the negative impact on consumption demand of a dollar of a given tax depends on the marginal propensity to consume of the taxpayer, and the way in which it affects his economic position; the impact will rarely be equal to the amount of the tax. From all of this follows the proposition that a balanced budget may exert a positive or negative impact on domestic demand and that consequently an equal increase in revenues and expenditure generally cannot be considered to be “neutral” in its impact.

In order to get a more satisfactory measure of the impact (or first round effect) of the budget, an attempt may be made to construct a weighted budget balance. This balance is derived by assigning a specific weight to each budgetary item and then summing up the items multiplied by their respective weights. The weight assigned to each item would be designed to reflect its impact on GNP or on demand for domestic goods and services,⁴⁰ as to both magnitude and sign. For example, if expenditure on imports and savings are the only leakages to be taken into account for the personal income tax, the weight for this tax would be — (l-m)c, where m is the marginal propensity to import of the income-tax payers and c is their marginal propensity to consume. A budget item multiplied by its weight may also be called the first-round effect of that item, since its total effect can be obtained by multiplying this product and the general multiplier. The weights for the different budget items are also indicative of (one aspect of) their relative effectiveness for demand management.

A relatively simple way of assigning weights to different budget items would be to use a basic analytical framework for deciding what leakages should be allowed for and then giving them numerical values derived from existing estimates of the relevant propensities, or assigned on the basis of judgment. A more satisfactory way of working out a weighting system would be to construct a somewhat disaggregated fiscal policy model suited to the type of economy that is being studied and to derive the formulas for the weights from the model. This procedure will not only ensure consistency of the weights used but will also make possible, wherever necessary, an econometric estimation of the relevant propensities.⁴¹

Even in the absence of a general weighting system, it may be desirable to work out a limited “weighting” exercise designed to show separately the impact on GNP and on the balance of payments. Basically, what is involved is the separation of the total budget balance into domestic and foreign components in such a way that the former will reflect the direct impact on demand for domestic goods and the latter the direct impact on the balance of payments.

Domestic balance. The measure reflecting the impact on GNP equals the total of expenditure on domestic goods and services minus the sum of domestic current revenues, depreciation, and the sale of domestic physical assets. That is,

where G. 1. – F. 1.	=	net decrease in domestic cash balance
and F.2. – G.2.	=	net increase in foreign cash balance.

View Table

where G. 1. – F. 1.	=	net decrease in domestic cash balance
and F.2. – G.2.	=	net increase in foreign cash balance.

The measure reflecting impact on GNP is thus seen to be equal to the sum of (a) net domestic borrowing, (b) decrease in domestic cash balance, and (c) net sale of foreign exchange to the banking system. Whereas the expansionary deficit (in terms of equation (3)) excludes foreign borrowing and any current receipts from abroad, in the computation of the measure reflecting only domestic impact, foreign receipts are included to the extent that they are used to finance domestic expenditure.⁴³

Direct impact on balance of payments. It is obvious that to the extent that the measure of equation (5) is accurate it excludes the impact of the budget on the balance of payments. An approximate measure of the direct (adverse) impact of budgetary operations on the balance of payments may be said to equal

Most governments do not present, and probably do not maintain, separate accounts of domestic and foreign expenditure.⁴⁴ Hence, a fiscal analyst can at best make only rough estimates of direct expenditure abroad. (It is probably much easier to obtain a separation of the two components of receipts.) Besides, there are two other limitations to the measures of impact on GNP and on the balance of payments as suggested in equations (5) and (6), respectively. First, equation (6) could give an indication of only the direct impact of the budget on the balance of payments. It could be argued that if a separate measure of the impact on the balance of payments is to be worked out at all it should include both the direct and indirect effects, and that if this is not possible, one summary measure of impact on total demand is to be preferred. Second, it is arbitrary to exclude from the domestic component all expenditure directly incurred abroad but to include domestic purchases straight from an importer. More generally, there is an import content of domestic expenditure as defined and measured in Table 2 that should properly be added to expenditure abroad. The first of these limitations can be overcome only through a proper weighting exercise, but the second can be remedied by estimating the import content of a government’s domestic expenditure and by adding it to the direct expenditure abroad. This in turn would involve the estimation of the marginal or average propensity to import of the government sector or of the economy as a whole, and in that sense would take one farther than the presentation of budget accounts and the derivation of balances, which is the concern of this paper.

Timing or recording of transactions

The whole group of questions relating to the timing of transactions or to the dates to be attached to transactions can be touched upon only briefly. The size of the balances for a given year will differ according to the principle of timing chosen, that is, according to whether the accrual or the cash basis is adopted. For deriving the measure reflecting the impact on money supply, this problem is not relevant, as this can be meaningfully derived only on a cash basis. The accrual basis seems to be the most appropriate one for the balance reflecting government saving, since the figure of government saving is to be generally used within the framework of national accounts, which are (sought to be) kept on an accrual basis. For deriving the balance reflecting the impact on demand, ideally, each budget item should be recorded according to the “reaction-date” of the private sector for each type of budgetary measure. This would mean, in effect, a mixture of cash and accrual recordings, because in some cases the main impact is at the time of accrual (e.g., corporate income tax), while in others it is probably at the time of actual payment or receipt (e.g., personal income tax, transfer payments to households). In practice, the accrual system of keeping accounts poses a number of conceptual and theoretical problems, especially on the expenditure side. At present, accounts on a cash basis only are available in most developing countries. The cash basis statistics are generally considered more useful for macroeconomic analysis, particularly if fiscal and monetary analyses are to be integrated. As progress with accrual accounting is made, the budget balance, with some major items (whose reaction-dates are close to accrual) adjusted on that basis, can be presented as a supplement to the budget statement.

The budget balance as a measure of impact of budget policy

The budget balance approach, whether the weighted or the unweighted balance is used, can give an idea only of the initial impulse imparted to the economy. Since neither the total effects nor the time profile of effects can be determined on the basis of this approach, it can be used only to obtain a broad indication of the size and direction of the budget impact in different years for comparative purposes. That is, the budget balances, or changes therein, can be used only for ordinal comparisons of budgetary impact in different years and not as cardinal measures of the effects of policy in given years.

Apart from the limited use to which the budget balance can be put in demand analysis, a major objection to its use for evaluating budget policy is said to be that the budget outturn for any year partly reflects the impact of the economy on the budget. The reasoning is that if the budget balance itself is partly determined by the level of economic activity, then it would be inappropriate to use the balance for evaluating the impact of budgetary policy. According to this view, a distinction should be made between automatic and discretionary changes, and the evaluation of budget policy should be based more on the discretionary changes. The concept of the full employment budget surplus has been developed mainly to eliminate the effect of changes in income on the budget outcome and to derive a measure or indicator of the strength and direction of discretionary policy. Since even a deficit caused by an exogenous decline in income will serve as a cushion and will act to arrest the decline in income, or to raise it, in the subsequent period, there seems to be a case for employing also a total budget balance to get an idea of the total budgetary stance including the impact of automatic stabilizers. Nevertheless, the limitations of the budget balance approach discussed in this section and the detailed considerations put forward in the discussion of the definition of the expansionary deficit will have pointed up the circumspection with which conclusions about the impact on demand should be drawn on the basis of the simple budget balance.