This paper is an inquiry into a variety of schemes by which governments have in recent years compelled the private sector to make loans to them.
Section I outlines what is meant here by compulsory lending and tries to differentiate such schemes from those that are close or distant relatives. Section II summarizes the principal characteristics of the schemes of the countries concerned; further details are given in the Appendix. Section III is of a more analytical character, exploring the advantages and disadvantages that such schemes might have with respect to incentive effects, equity, administration, etc. Section IV summarizes some of the lessons of practical experience with these schemes in recent years. Section V asks whether there is a future for such devices. The general answer is that there is on balance a certain amount to be said in their favor, especially if various precautions are taken in their formulation.
I. Definition of Compulsory Lending
Compulsory lending is a phrase that could cover a large number of quite different schemes, and so we must clarify the exact meaning we have in mind. We shall take a compulsory lending scheme to be one where all people coming under some wide classification (e.g., liability to pay income tax) are obliged to deposit with the government a given sum of money for a period of time, on the understanding that it will be returned to them relatively soon. Several features of this definition should be stressed. First, we specify a “wide classification”; this is necessary in order to delimit our inquiry from various types of compulsory lending schemes applying to particular small sections of the community (e.g., reserve ratio requirements for banks in the United States; directions to insurance companies to buy government bonds in Ceylon; compulsory deposits on imports, as introduced in the United Kingdom in November 1968). At the same time, we do not insist that the coverage should be comprehensive; we are thereby able to cover schemes that, for instance, apply to individuals but not to corporations.
Second, the notion of “obligation” is essential. This means that we exclude such schemes as those that have operated over the years in, for instance, Sweden and Brazil, whereby voluntary depositing of funds with the government enabled firms to claim tax relief on investment; and we also exclude the multiplicity of instances where tax relief is given for voluntary saving in prescribed forms. Third, it is specified that the deposit should normally be made in money form; this enables us to leave out the issuance of government bonds in exchange for physical assets acquired from landowners or company shareholders when nationalization of some activity takes place.
Fourth, there is the requirement that the sum deposited will be returned to the depositor—or, of course, his heirs. This enables us to differentiate our field of inquiry from the operations of marketing boards in West Africa and elsewhere in the postwar period. Effectively, marketing boards have at times operated as a means of imposing compulsory lending on large groups of people, but there has never been any rule of repaying such sums to the individual contributors.
Fifth, we stipulate that sums lent should be returned to lenders “relatively soon”; this enables us to separate the schemes that interest us from the general range of social security schemes found in many countries in the world.1 Finally, it should be observed that we have used the term “compulsory lending.” Frequently, such schemes are known by the name “compulsory saving"; it seems to be less misleading to use the former term, for although governments can legislate for schemes of compulsory lending it is not normally within their power to legislate for net additions to the annual flow of saving.
II. Classification of Schemes
The archetype of all modern schemes of this sort was that proposed by Keynes in How to Pay for the War2 and introduced in modified form in the United Kingdom in World War II under the name “postwar credits.” A similar scheme operated in Canada; and the United States considered the idea seriously.3 Recently, schemes falling within our definition have been operating in the years given in the following countries:4
|South Africa||1952… (intermittently)|
|United Arab Republic||1965…|
In addition, a compulsory lending scheme was proposed by one commentator at the time of the heavy budgetary impositions in the United Kingdom in March 1968.5
It may be useful to set out some of the general characteristics of such schemes.6 The basis of the contribution is usually income or, alternatively, taxes paid on income. Detailed practices vary widely with respect to the exact definition of income (e.g., it may or may not be the same as that for income tax purposes) and the categories excluded from any liability (e.g., the exemption limit may be higher than that used for determining income tax liability). The degree of progression also differs considerably from one scheme to another. In one country (Nepal) the basis is agricultural production, but this is understandable in view of the unsophisticated nature of the income tax system in that country.7 Peru’s scheme developed out of a 10 per cent tax on exports. Usually, compulsory loan payments are not deductible for income tax purposes; nor are repayments taxable. But the two schemes (compulsory deposits and annuity deposits) operating successively in India between 1963 and 1968 were an exception in both respects.
Bonds issued as part of compulsory lending schemes have certain characteristic features. They are usually for periods ranging from five to ten years; interest rates are usually of the order of 4 per cent to 5 per cent, tax free, sometimes (e.g., in Guyana and Ghana) with a lottery element as a partial substitute; and, in general, such bonds are nontransferable, though there are exceptions to this. An abortive proposal in Chile in 1967 would have allowed transferability; there has been a degree of transferability in Israel; and the scheme introduced in Turkey in 1961 legally allowed transferability of the ten-year bonds after the first five years.8
Most schemes start from the premise that bonds will be repayable in cash on the due date. But some (e.g., in Chile and Brazil) have provided that repayment can be offset against future tax liabilities. And others have been amended in practice in such a way that, in effect, bonds had to be exchanged for bonds instead of cash. In some countries (Brazil and Israel) there have been provisions for adjusting capital repayments and/or interest in accordance with a price index.
Some schemes were introduced initially for a limited period and in fact did not last longer (e.g., in Denmark, 1963-64). Some were introduced for a limited period but were renewed (e.g., in South Africa). Others were introduced for an indefinite period and have continued operating (e.g., in Turkey); yet others were introduced for an indefinite period but were soon withdrawn (e.g., in Ghana and Guyana). And it has often been the case that the lawmakers were too ambitious, and so schemes have frequently been modified in the light of experience, usually in the direction of reducing their coverage (e.g., in India and Ghana). Peru’s scheme (1968) originated in a 10 per cent tax on exports repayable in eight equal annual installments, at no interest; an option was then given of two-year to four-year dollar bonds carrying 7 per cent interest.
We now look at the reasons given for introducing schemes of this kind. The most usual was the macroeconomic one: that some additional reduction in purchasing power was needed at the time. This has sometimes been in the context of financing additional expenditure (e.g., in Ghana and Morocco); sometimes in the form of stating the alternative as cuts in spending (e.g., in Brazil, 1965); in one country (Denmark) the alternative was stated to be the postponement of tax cuts that had already been announced, and in another (Canada), a compulsory loan was imposed on corporations in order to restrain capital investment.
The macroeconomic reason for restraint may or may not have been fully justified in each of these instances; but let us assume for the moment that it was. We are then left with the further question: Why was this particular form of restraint chosen? This cannot be answered in any simple way from the relevant budget statements or legal enactments, and there has to be an element of rationalization in the answer. One argument is that, in effect, a cruder basis for calculating liabilities is permissible in the compulsory loan case, when people will be repaid in the future, than in the income tax case, when the contribution is sunk without trace. One version is that on these grounds one can take a gross concept of income, without worrying much about exemptions, deductions, etc., as the base for compulsory lending.9 A slightly different line would be to say that, given the crudeness of many personal income tax systems in developing countries, there is a strict limit to the amount that one can hope to raise in income tax form; but it may be possible to raise an additional tranche of revenue if it is levied as a compulsory loan.
Another argument that seems to lie behind these schemes is the intergeneration one. If there is some particular need for more government revenue at the moment (say, to finance a major public works project), it has been maintained that, whereas a tax imposes the burden of payment unequivocally on the present generation, a loan arrangement will make it easier for some future recompense to be made to the survivors or heirs of this generation. This possibility is alleged to have both equity and incentive connotations.
It has also been maintained that compulsory lending may do something to promote the development of capital markets, even though the usual situation is that such bonds are nontransferable. Presumably the proposition is that the issuance of bonds or certificates, even if nontransferable, will acquaint people with such financial assets and prepare them to buy and sell other financial claims.
Finally, compulsory lending is sometimes linked with specific expenditure projects, often of a capital nature. This gets us into such points as the general merits of earmarking, the argument that it is more legitimate to borrow for capital than for consumption purposes, and so on.
III. Analytical Considerations
We must now look at the pros and cons of these schemes from a more analytical viewpoint. It will be convenient to consider a single act of compulsory lending first and look later at any complications that may spring from a repeated process.
Single act of lending
We shall consider compulsory lending principally in relation to the alternative of an income tax. But we shall look at various other alternatives later.
Income tax alternative
Let us take the case where one can raise revenue either by a proportional income tax applying to all incomes or by a proportional compulsory loan on the same basis. The loan is assumed to be repaid, with accrued annual interest, at the end of a given period of time.
In the income tax case, we raise $1.00 in year 1 and that is the end of the matter. In the compulsory loan case we raise $1.00 in year 1; as this is repaid in year n with accrued annual interest, the sum involved will be (l+r)n, where the rate of interest paid (net of any tax liability) is 100r per cent. For the lender the present value of this sum in year 1 will be
Schematically, the situation is as follows:
|Income Tax||Compulsory Loan|
|Present dollar value of||Present dollar value of|
Given this situation, we now investigate the circumstances under which the individual will regard the compulsory loan as equivalent, first, to the income tax alternative, and second, to a pure loan.10 It is clear from the tabulation that compulsory borrowing would be exactly equivalent to the tax, where
One or two further complications should be noted. First, interest payments may not all be accumulated until year n but may instead be paid in the intervening years. This does not alter the general principles of the analysis; while it will not now be necessary to raise as much as $(l+r)n to repay the loan in year n, there will be greater financial needs in the intervening years (relative to the income tax case). So, essentially the analysis is the same. Second, we have to think about the case where compulsory loan payments are tax deductible and repayments are taxable, as in India. In this instance, the analysis is simpler in that effectively such schemes reduce to a system of paying taxes in advance; all we need to know is whether the taxpayer’s subjective rate of discount is higher or lower than the net rate of interest paid by the government.11
The crucial question in judging the relative incentive effects of compulsory lending and income taxation is whether the former is regarded as being nearer to a tax or to a loan. It seems rather clear that in practice it will be between the two limiting positions. On the one hand, it is reasonable to argue that the subjective rate of discount attached to the finance of repayment will be greater than that attached to the repayment itself. Many people will surely think that they will not be called upon to make any mandatory contributions at all (in the form of either extra taxes or smaller government benefits) toward repayment; or that at least there will be new additions to the ranks of taxpayers who will shoulder part of any burden at the due date. Therefore one can assert with some confidence that r" > r’; this rules out the equivalence of compulsory lending to a straight tax. On the other hand, the proposition that compulsory lending can be thought of as equivalent to a pure loan or zero tax is also indefensible. If it were, presumably there would be no need for compulsion; in other words, the rate of interest actually paid by a government is likely to be less than the subjective rate of discount attached to repayment, when distrust of government, lack of negotiability, expectation of price rises, etc., are all allowed for.12
But having made the point that the compulsory loan is unlikely to approximate to either extreme, it is difficult to go further and speak with confidence about the combination of tax and compulsory loan rates to which the population would be indifferent. This is partly because loan market data relating to transferable bonds is irrelevant for this case,13 and partly because lending is being imposed on people who have never lent before in their lives. One simply cannot be dogmatic here; at the same time there is some fragmentary evidence which does suggest that the tax element is substantial but by no means complete.14
Given this general background, what can be said about the savings incentive effects of compulsory lending relative to an income tax? There seem to be some opposing considerations here. First, if people evaluate the reduction in net worth involved in compulsory lending to be less than that involved in income taxation, one would expect the reduction in consumption to be less, too. So, an equal rate of compulsory loan must be expected to reduce aggregate demand less than a given rate of income taxation—always subject to constraints on the ability of people to finance dissaving when they have no capital assets or when there is no ready access to borrowing facilities. On the other hand, there is the well-known point that income taxation may differentiate against savings relative to consumption, depending on such things as the interest elasticity of savings.15 Essentially it is a matter of weighing up the capital effects and the price effects—the former pointing to less, and the latter to more, saving in the compulsory loan case.
There is clearly no way of pinpointing the outcome. But two other relevant considerations should be added. First, insofar as corporations are subject to the additional levy, one would expect a one-year tax increase to be borne largely out of undistributed profits, given the well-known reluctance of corporations to cut dividends. In this instance the whole, and not just part, of the tax increase would be at the expense of saving. Second, the macroeconomic position may be more complex than the simple analysis so far would suggest. In many developing countries, private domestic investment may be closely tied in with private saving,16 and even if this is not true, private saving may be tied in with capital exports, which may in turn affect the capacity to import. So, the full macroeconomic position is very complex.17
In the absence of any clear indications, let us assume that, from a macroeconomic standpoint, the rate of compulsory lending would have to be the same as the income tax rate. The relative effects on incentives to work and to take risks would then depend on the (subjectively estimated) size of the tax element in the loan. Given that people do give some weight to the pure loan element, one would expect differential effects on willingness to work and to undertake risks—depending on the usual balance of income and substitution effects and (in the risk case) the possibility of loss offsets. Insofar as tax and loan rates are progressive rather than proportional, there would be further complications of differential effects on different income groups and of their relative reactions.
To summarize on incentives, we can expect that a compulsory loan will be viewed as a lighter burden than an equivalent rate income tax, though to an unknown extent. The macroeffects of equal rate imposition could go either way, but if one assumes that they are identical, there will be some net advantages for work and risk incentives, depending on the usual income-effect and substitution-effect considerations.18
What of the equity side? If it is not necessary to raise the same amount of revenue by compulsory lending as by income tax, or if the incidence of the two alternatives is different for the various sections of the community, there will be immediate distributional implications. In principle, either method could favor the poor more than the rich, or vice versa. But let us assume for simplicity that we do in fact raise the same amount of revenue by either means (i.e., we abstract from any differential savings incentives under a scheme of compulsory lending) and that the distributional impact is initially the same. If such were the case, any distributional differences would then depend crucially on the method adopted for financing repayment of the compulsory lending.
The problem is a complex one in that one has to think about the distribution of both capital and income. Suppose we consider the solution advocated by Keynes for World War II19—the imposition of a capital levy at the end of the period to repay the government’s compulsory creditors. It would then be reasonable to argue that this would lead to a more equal distribution of private capital—in that the levies to finance repayments would be a function of the distribution of capital and the repayments themselves would be a function of the distribution of income, and the former distribution is normally more skewed than the latter. Ignoring other influences on the size and distribution of private capital over the period, the end result would be the same total as before the scheme of compulsory lending but with a more egalitarian distribution. Similarly, there would be a change in the distribution (but not the size) of income from property with consequent repercussions on the size distribution of total income. But further complications would ensue if the process of repayment by means of a capital levy altered the total of current spending; one would then presumably have to take further action to stabilize the economy, and this would in turn have further distributional effects.
To take another example, let us hypothesize that compulsory loans are repaid by means of borrowing. In this instance (ignoring, as before, other influences on the stock of private capital) the process of compulsory lending, borrowing to finance repayment, and repayment itself will tend to make the total of privately owned capital greater than before the compulsory lending took place. It would not seem, however, that one can be specific about the degree of concentration of this larger total: this will depend on the distribution of ownership of the newly raised loans, as well as of the repayments. The same conclusion must therefore follow for the effects on the distribution of income, quite apart from any complications owing to changes in goods and factor prices associated with the borrowing process.
On the basis of these examples—which could be multiplied indefinitely—one must conclude that the effects of compulsory lending, relative to income taxation, on the size distribution of wealth and income are by no means clear cut. It may be that the need to repay sums compulsorily borrowed would force through some major and unequivocal fiscal measure, such as a capital levy, which would not otherwise have been possible; but unless this is so, one must be very guarded about the likely consequences of a compulsory lending scheme for the size distribution of income or capital.
There is one further matter. The finance of repayment at the end of the period may well fall on people irrespective of whether they were contributors or even born when the lending originally took place. This implies the likelihood of some increase in the wealth of the “old” (or the heirs of the “old”) relative to the “young.” This seems to be a much more certain conclusion than any of those above about size redistribution, and it is where the kernel of truth lies in the notion of “passing on the burden to future generations.” 20
It is sometimes argued that, from an administrative viewpoint, schemes of compulsory lending have much to commend them. For instance, whereas an income tax must normally be refined, taking into account the number of dependents and so on, a compulsory loan can be levied on gross income without worrying about such niceties. There are two comments to make on this. First, if there is a refined income tax already in existence, the above argument cannot apply to those covered by it; it will therefore be relevant only for those newly coming into the net under the temporary surcharge. Second, even though there may be administrative savings at the time of imposing the loan, there will presumably be administrative costs in issuing certificates and registering loan holdings, in transfers between one holder and another (e.g., at death), and in the process of financing and making interest payments and capital repayments. So it is very difficult to conceive that there is much left in this argument for administrative convenience.
Last, arguments have sometimes been put forward about the need to avoid “tax friction,” i.e., unnecessary fluctuations in tax rates.21 This is another old chestnut, like the analogous one about the need to smooth out the price fluctuations that would otherwise result from the strict application of short-run marginal cost-pricing principles to public enterprises. There may well be something in this, but all the same one must question whether people would regard a compulsory loan as so different from, say, an income tax, that one should allow it to count heavily in the scales.
Other revenue alternatives
Let us now look at one or two possibilities other than the income tax alternative to compulsory lending.
Suppose that we compare the latter with a sales tax of some kind. One obvious difference is that a sales tax (or, at any rate, a tax limited to consumer goods and services) cannot be argued against as implying a double taxation of savings—with repercussions on the relative rates of tax and compulsory lending that are equal from a macroeconomic viewpoint. More generally, there will now be a different distribution of initial payments with all sorts of possible consequences for incentives, income distribution, and so on. If one accepts the view that the sales tax is likely to result in higher market prices than the income tax system, there will presumably be differential effects on any cost-push inflationary mechanism.
As another possibility, assume that the government has reached what it considers, rightly or wrongly, to be the limits of taxation yield, and so it simply has to borrow to finance additional expenditure. Borrowing from abroad may be possible to some extent; internal borrowing that does not involve adding to the stock of money may also be possible to some extent; and even if there is an addition to the money stock, there may be spare capacity in the economy. But the most likely alternatives are compulsory borrowing or borrowing involving an addition to the money supply in a situation where there is no spare capacity. Any comparison between these two alternatives must start from the proposition that the amount of borrowing will tend to be less in real terms,22 but greater in money terms, if the borrowing is noncompulsory. We must also expect the distribution of claims on the government to be different in the two instances, depending on the distribution of gains in money income in the inflationary process and the savings propensities of different groups. The most likely result will be a less egalitarian distribution of the increase in private wealth in the inflation case. The further likely repercussions on work and savings incentives, the product-mix, and the balance of payments need not be discussed in detail here, important as they may be. One point, however, should be made. The inflationary alternative might be more conducive to capital formation than a compulsory lending scheme, whether one thinks the main driving force behind capital formation is expected profitability or availability of funds. Whether such additional capital investment would take the form of luxury housing, etc., is another matter.
A revenue-expenditure package
Another comparison is that between raising a compulsory loan and spending it on some government project, on the one hand, and doing without both, on the other. Although the proposition could be explored without designating any particular form of government spending, it is likely to be stronger if one does. So let us arbitrarily say that the expenditure would be on a dam.
There are varying arguments to disentangle here. One is whether the net social benefits of the package outweigh the social opportunity costs;23 this is a factual question that must be investigated separately in each case. Another is that people will more readily accept a cut in private investment or consumption if it is for the sake of a specific recognizable outcome from government spending. However, the relevant question here is whether, say, a compulsory loan-dam package is more acceptable than a tax-dam package. A priori, it is not easy to see why this should be so. There is also the naive proposition that government borrowing in general, and compulsory lending in particular, is more justifiable if associated with capital than with current spending. But we need not explore the fallacies in that argument.
Alternative bases for compulsory lending
We have assumed thus far that compulsory lending is based on some measure of income. This is not inevitably so. Nepal’s system is based on agricultural production, and there are clearly other possible systems.
Suppose, for instance, that compulsory lending were on the basis of net worth. This is not in fact very likely, partly because of difficulties of assessment and partly because some of the people whom one may want to catch may not have any net worth. But if this were the basis, one could then compare it with a tax alternative. There would obviously be advantages relative to an income tax—on the basis of the usual comparisons of capital tax and income tax. Relative to a net worth tax, the advantages are less obvious; presumably the prospect of repayment would encourage dissaving in the loan case relative to the tax one.
Another alternative basis might be consumer spending, e.g., purchases of stamps, redeemable at a later date, might be made compulsory with retail purchases. Compared with a straight retail sales tax, this would again tend to reduce voluntary saving. But given the very limited knowledge of and control over retail trading transactions in most developing countries today, one cannot regard such comparisons as very relevant at present.
Let us now abandon the assumption of a single act of lending and imagine that we have a repetitive process, so that the alternatives are an income tax of $1.00 for each of n years or a series of one-year compulsory loans successively renewed through time. Essentially, the main basis of the comparison remains the same; it will depend on the relative sizes of r, r’, and r“.24 And the same will hold true in the more complex cases, such as where the process is an intermittent one, additional levies being made in some years but not in others.
The comparison between the income tax and compulsory loan alternatives does not remain totally unchanged, however. First, if the prospect is one of a long-continued process of lending and repayment instead of a once-for-all effort, it may well be thought that net worth is diminished less, relative to the income tax case, than earlier—though the precise result will depend on expectations about methods of financing repayments. But on this basis there may easily be some net reduction in the deflationary effects of loans relative to taxes, compared with the once-for-all case. Second, it is sometimes argued that in a country with a growing level of money income per capita (whether owing to rising output or price inflation) for a given population or with a growing population with given money income per capita, or both, the amount of compulsory lending in any one year will tend to exceed the principal repaid, even though rates of contribution remain unchanged. The exact result will be a complex one,25 but the general point is clear: compulsory lending can easily be a way of permanently reducing the level of private demand relative to potential output below what it otherwise would be. However, the question is not the deflationary effects of a continuing borrowing and repayment process but these effects relative to the deflationary effects of an income tax levied at unchanging rates in the context of money income increases, population increases, etc. In this context, it is by no means clear that the deflationary effects of one device relative to the other are markedly different in the continuing case from what they were in the once-for-all case.
As far as incentives to work and to take risks are concerned, the argument would seem to remain substantially as it was before, i.e., it will depend on the size of the tax component of the compulsory loan and on the income and substitution effects of switching from one to the other.
IV. Practical Experience
Ideally, one would like to be able to test some of the propositions enunciated in Section III and see whether there is any reason to think that in practice the schemes operating in recent years have been better from an incentive or equity point of view than the most relevant alternatives. In reality, we cannot do anything of the sort: quite apart from any questions of methodology or appropriate procedures, the data and records simply do not exist for any scientific and exhaustive examination. What follows must necessarily be of an incomplete and even impressionistic character.
One point on which there is certainty is that schemes of the sort discussed have not usually been important sources of revenue. For instance, lending of this type was equivalent to only 1 per cent of total central government current revenue in India in 1963/64, and to only 2 per cent in South Africa in 1966/67. Very exceptionally, receipts in Nepal in 1964/65, the peak year of that scheme, were equal to one third of current revenue; but the Nepalese scheme differs from all the others in a variety of ways. Perhaps the Ghanaian experience should be singled out as an example of what can be done by a country that uses this device in a determined way. Receipts were equal to 10 per cent of current revenue in 1962/63. But the scheme operated for only two years there, and it may be questioned whether this pace could have been kept up for long.
The overwhelming impression derived from an examination of these schemes is that they were introduced as an emergency measure to cope with some financial crisis. Thus, various countries essentially introduced them for a year or so at a time when there was a need for a measure of deflation. Canada wanted to cut back corporate investment in 1966; Denmark was faced with possible inflationary problems in 1963; Nigeria was engaged in a war in 1967. Indeed, it is possible to go a bit further and suggest that, by and large, a compulsory loan is a device taken up by countries already making high tax efforts and suddenly faced with a need for greater revenues. One can obviously debate the measurement of tax effort interminably, but if one takes the findings of Lotz and Morss26 on tax effort in 52 low-income countries, corrected for inter-country differences in per capita income and openness, the result is that of 11 countries 27 with compulsory lending schemes 3 were in the high tax effort group, 7 in the middle group, and only 1 in the lowest. This may give some limited support to the notion that compulsory lending is a device adopted mainly by countries that are already pushing toward their taxable capacity limits.28
Another feature of these schemes, to which we alluded earlier, is their ephemeral nature. As can be seen from the Appendix, a number of countries have used this technique for only a short time; and among those having longer records there have often been major changes in the operation of the schemes. Thus, the scheme in the Sudan operated provisionally for a few months in 1967 but had to be abandoned when the parliament refused to pass the law authorizing it. That in Guyana was introduced early in 1962 but came to an end some three years later when it was declared unconstitutional by the courts. And in other countries (Denmark and Canada) it was intended from the beginning that such schemes should only be temporary.29
Major changes have been introduced in a number of countries. Thus in Ghana, Guyana, and India, the exemption limits had to be raised and/or various classes of people originally within the schemes had to be excluded at a later date. It should be noted that it has been overwhelmingly a matter of excluding some people originally included rather than including some originally excluded; this suggests that the various tax authorities have consistently underestimated the political and/or administrative difficulties of compulsory lending. There have also been a variety of instances where there were major changes in the laws governing the schemes, e.g., in Brazil in the 1950’s and in India and Turkey in the 1960’s.
By and large, these schemes have been closely tied in with the existing income tax systems, in providing lists of lenders and bases of contributions. L.A.Y.E. (lend as you earn) has also been a common feature. However, there are some cases where coverage has been narrower, and others where it has been wider. Thus, foreigners have been exempted frequently (e.g., in Nigeria), and even when they have been subject to the levy, there has usually been provision for repayment on departure from the country. On the other hand, attempts have been made to widen the coverage, e.g., by including payers of a flat-rate direct tax (in Nigeria) or cocoa farmers (in Ghana). Exceptionally, Turkey applies the system to receipts from many types of capital transactions, lottery prizes, gifts, and inheritances, as well as to incomes, but the comprehensiveness of the coverage achieved is not known.
Quite apart from the indirect evidence presented by the “on-off” character of some of these schemes, there have been some specific references to administrative difficulties. Thus it has been reported that the system of presenting coupons for interest payments in Turkey was far from satisfactory. And in Ghana, considerable doubts were expressed about the accounting control of the scheme, and instances were quoted where local authorities had deducted contributions from employees’ wages and salaries but had not handed them over to the Central Government.30
On the important questions of incentives, it is not possible to say anything with certainty. However, there are some pointers to people’s views about the relative importance of the pure tax and pure loan components in these schemes. Although bonds issued in Turkey under the 1961 law 31 were not legally transferable for the first five years, it is understood that they were commonly changing hands at about 30 per cent to 33 per cent of face value, notwithstanding tax-free interest payments at 6 per cent per annum. This point was expressly taken cognizance of in 1967 32 when among the major changes enacted was a provision that all old bonds must be exchanged for new ones, with a penalty tax on those who could not prove that they were the original owners. As another piece of “evidence,” it is reported33 that in October 1961 the compulsory loans levied in Brazil, under legislation of 1951 and 1956,34 were being sold on the Rio de Janeiro Stock Exchange at a discount of 40 per cent of the face value.
Obviously, there were many influences at work affecting the price at which people were willing to dispose of assets—not least the fact of the illegality of transference—but, as a minimum, one can say that people did not regard the bonds as worthless on either of these occasions. So, there was some element of loan in them in most people’s minds. On the other hand, there is certainly no evidence to suggest that, where permissible, anyone ever voluntarily took up more of the bonds than the minimum required of him.
Informed observers with whom the matter has been discussed have unanimously taken the view that there are psychological advantages for incentives to work, etc., in giving people bond receipts rather than tax receipts. Whether these advantages can be retained for long is another question. It would certainly appear to be the case in Nepal—where the receipts were re-lent for specific purposes, mainly agriculture—that the scheme was judged to have worked fairly satisfactorily. But, as explained before, the Nepalese scheme is so different from the rest that no generally applicable conclusions can be deduced from it. And it should be noted that an assessment of wartime experience in the United Kingdom and Canada did not reveal any important incentive effects.35 Nor, finally, can we come to any certain conclusions on the equity effects, as these would inescapably involve the reconstruction of what might have happened to the fiscal pattern in the absence of compulsory lending schemes.
V. The Future
It is easy enough to mount a case against compulsory lending schemes. It might be argued, first of all, that their raison d’être is to be found in the prospect of a short, sharp need on the part of a government to increase its revenue intake, the pre-eminent example being a war that is expected to last for a short time. And if the postwar period is likely to be one of slackness in the economy, thereby facilitating repayments, the case is pro tanto stronger. On the other hand, the common situation in developing countries is not this at all, but rather a long slow haul during which necessary steps toward modernization may be taken; nor is the prospect of Keynesian underemployment equilibrium at all likely in the near future for most of them. War and development conditions are not necessarily mutually exclusive (e.g., the Nigerian civil war that began in 1967), but in general it is true to say that the conditions under which the original Keynes plan of 1940 was derived have little applicability to most developing countries.36
Another proposition might be called “the dilemma of the income tax list.” In order to operate a compulsory lending scheme, one has to have a clearly defined list of people on whom the levy can be imposed in the first instance. One of the few sources of such a list in many countries is the record of income-tax payers. But if this is the only list, what is the strength of the argument for collecting a loan rather than more income tax? In other words, without a long list one cannot introduce compulsory lending; with a list, one may not need to.
In support, it might be further argued that one is simply deluding oneself in thinking that anything effective can be done by compulsory lending to develop capital markets in developing countries. If it is insisted that bonds be nontransferable and that they carry rates of interest of the same order of magnitude as those prevailing in more developed countries, then it might be said that one is doing very little to unify the capital market or to break down the barriers among its different segments.
On the equity side, if one seeks a conservative view one cannot do better than refer to a statement by the Board of Inland Revenue in the United Kingdom in 1940 apropos the social aspects of the Keynes wartime proposals.
These factors do not, in the Board’s judgment, warrant the abandonment of well-established principles of taxation for ones which are unsound and which involve enormous administrative difficulties.37
When one adds to these arguments the general record of such schemes in recent years (low yield, short-lived experiments, administrative difficulties, and so on), it is small wonder that some are inclined to write them off as being a time-expired survivor from a bygone age. When asked to explain why such schemes are repeatedly taken up afresh, the reply might well be that finance ministers are prone to take an easy way out even though they may be storing up trouble for their successors who have to finance the repayments; or, alternatively, that such schemes are a reflection of the assiduity with which the finance department in one country studies innovations in the revenue system in another—but with a lag of two or three years.
The proponents of compulsory savings schemes can also deploy a number of arguments to support their case. The first question may be to ask, What is the relevant alternative to compulsory lending? Lower public expenditures may be impossible politically, whatever the economic arguments. Pure loans may involve such high rates of interest that governments would be liable to attack on the grounds either of favoring the propertied classes or of discouraging private investment unduly. More help from abroad may also be impossible. In the end this would leave the alternatives of higher taxes or the imposition of direct controls along with higher taxes. And both of these might well be even more unpopular than a system of compulsory lending. So, as a minimum, one can put up a case for saying that compulsory lending may be the least of a number of evil alternatives.
In fact, one can go further than this. There probably is something in the incentives argument, and in the propositions that the holding and handling of bond receipts will do something to familiarize people with the institutions of the capital market, that one does not have to stick rigidly to the income tax list, and that such schemes should more properly be thought of as being slightly avant-garde rather than as discreditable gimmicks.
Perhaps a fair over-all view is to say that although such schemes should not be regarded as hailing the dawn of the new Jerusalem,38 neither should they be dismissed as being completely worthless. If this be the case, we must try to make suggestions about the most appropriate direction in which they should go if used in the future.
One general principle that affects such schemes in a variety of ways is that the connection with the income tax system should be minimized. Unless this is done, people are obviously much more likely to put emphasis on the tax, as distinct from the loan, element in compulsory lending. This means, for instance, that one should not have income tax payments as a base for compulsory loan levies. It is preferable to have income as a base, and even more preferable to have some concept nearer to gross income (before exemptions, etc.) rather than to whatever is the precise net concept used for tax purposes. Where it is possible to extend the coverage of the compulsory loan beyond the population covered by the income tax system, this is desirable, but one must recognize that there are sharp limitations to this. Similarly, if it is possible to avoid making compulsory loans collectible from wages and salaries in the same way as income tax, this is also desirable. If, for instance, there is a separate social security system to which employers and employees contribute, it may be better to operate through that mechanism. It is also desirable that government accounts should treat receipts of this kind as a capital item rather than a current one,39 and also that titles such as “income tax bonds” should be avoided.
A word or two needs to be said about interest payments. It is probably better, if not administratively prohibitive, to pay interest on an annual basis rather than to allow it to accrue until the date of redemption. Whether a lottery element should be incorporated in interest payments (or bond drawings) is a more open question. Some countries have certainly had difficulties running a lottery ingredient in a compulsory lending scheme; on the other hand, lottery schemes do play important roles in some countries.40 This must be a matter for individual countries to decide, without the aid of any very general economic principles. The question also arises whether interest payments (and, for that matter, capital repayments) should be linked to an appropriate domestic price index or to a foreign currency. This raises many questions of the level of interest rates paid on compulsory lending relative to those ruling for other securities, the prevalence of index-linking practices generally in a country, and the still wider question of the recorded past and the expected future price levels.41
Various features of the bonds themselves must now be considered. The first is whether they should be transferable. One suspects that the authorities in some countries are not facing reality on this subject: if they really want to encourage a capital market, they are likely to get further by allowing transactions in such bonds than by forbidding them. In practice, edicts about nontransferability may not work; and the claim sometimes made that such restrictions are necessary to prevent poorer people selling them at much less than face value is somewhat disingenuous. So there is surely a great deal to be said for negotiability.42 Whether one should countenance bearer form is another matter. Bearer securities do undoubtedly cause all sorts of administrative difficulties, and therefore one can understand any reluctance to add to the existing stock of them. But one possible solution would be to give people a choice between a bearer investment, fully transferable, and a negotiable bond—the latter carrying a higher rate of interest than the former. This would, at any rate, be a less restrictive and less negative attitude than that usually adopted.
There are a variety of other possibilities with respect to the form taken by the loan. One might be to allow people to hold it in some special blocked account;43 another might be to allow convertibility at some point into holdings in state enterprises. Perhaps the most important point here is a negative one: repayment should not be bound up with the income tax system. On those grounds, one must question those schemes that approximate to an advance payment of income tax (e.g., in Chile) or those that make contributions tax deductible and repayments taxable (e.g., in India).44
About the length of life of loans, there is probably not much to be said in general terms—this must very much depend on governments’ requirements and contributors’ reactions. One can speak with more certainty about the desirability of a fixed date of repayment. It would seem likely that unless the government fixes this (at least within narrow limits) the tax element would far outweigh the loan element in the contributor’s eyes, even if the rate of interest were somewhat higher in compensation. On the other hand, it is also reasonable to argue that the allowable occasions for premature repayment should be kept to an absolute minimum; otherwise, the administrative complications would be legion.45
Whether schemes of this sort should be tied into specific expenditure commitments is more a matter of psychology than economics. We prefer to remain neutral on this; if it is felt that a particular scheme in a particular country can be more acceptable in this way, we should not oppose it, but, on the other hand, we should not wish to do anything to encourage the notion either.
Finally, and in some ways most important of all, schemes of this sort should not be introduced without a good deal of preparation and discussion. This might seem such an obvious point as to be not worth making. But the history of so many countries is so replete with examples of revenue levies (both of the compulsory loan variety and more generally), which have in practice gone off at half cock, that no opportunity of making such an apparently trivial statement can be allowed to slip.
|Country||Years of Operation||Individuals||Companies||Basis of Contribution||Maximum Duration of Loan (.years)||Interest Rate Paid (percent)||Special Features|
|Brazil1||1951…||Yes||Yes||Income taxes||5-25||5 (minimum)||Indexing element; repayment offset against taxes|
|Canada||1966-67||No||Yes||Corporate profits (gross)||2-3||5||—|
|Chile||1968…||Yes||Yes||Income taxes||7||—||Repayment offset against taxes|
|Earmarked for housing and steel|
|Ghana||1961-63||Yes||Yes||Income; cocoa sales||10||4|
|Payment deductible; repayment taxable|
|Israel1||1959-68||Yes||Yes||Income; income tax||15||4|
|Indexing element; some transferability; lottery element|
|Nepal||1963…||Yes||No||Agricultural production||5||5||Earmarked for lending purposes|
|Pakistan||1966…||Yes||Yes||Income tax payments||10||5||Some transferability|
|South Africa 2||1952…||Yes||Yes||Income; income tax||5-7||4½|
|Peru||1968…||Yes||Yes||Exports||4||7||Optional eight-year annual repayment at no interest|
|Sudan||1967||Yes||No||Income||5||…||Scheme only in force temporarily|
|United Arab Republic||1965…||Yes||No||Income||5||5||Wage and salary earners only|
Variety of schemes during years listed.
Variety of schemes during years listed.
Systèmes d’emprunts forcés
Cette étude constitue une enquête sur un ensemble de procédés par lesquels les gouvernements ont, dans les années récentes, contraint le secteur privé à leur consentir des prêts.
La Section I définit le champ de l’étude. L’emprunt forcé est par définition l’opération par laquelle toutes les personnes appartenant à une vaste catégorie sont obligées de déposer entre les mains du gouvernement un montant déterminé d’argent pour une période donnée, étant entendu que ce montant leur sera restitué dans un délai relativement court.
Dans la Section II et en Annexe sont décrites les principales caractéristiques des types d’emprunts forcés utilisés dans 19 pays au cours des dernières années. Les détails donnés sur ces opérations concernent notamment leur date, leur base, leur durée maximum et le taux d’intérêt paye.
La Section III est une étude analytique. L’auteur établit des comparaisons entre le système de l’emprunt forcé et les différents autres moyens dont disposent les gouvernements, notamment l’impôt sur le revenu, au regard des effets d’incitation, de l’équité et des formalites exigees. En matière d’incitation, il conclut qu’un emprunt forcé répresente une charge plus lègeré qu’une imposition sur le revenu équivalente; en revanche, aucune conclusion aussi simple n’est donnée en ce qui concerne l’équité et les formalités exigées.
La Section IV rend compte de l’expérience déjà acquise dans ce domaine. Bien que dans la plupart des pays cités de telles mesures aient ete prises arm de faire face d’urgence à une crise financiere, on peut neanmoins en tirer un certain nombre de leçons.
L’avenir de ce système est envisagé dans la Section V. La conclusion générate est que, tout bien considéré, l’emprunt forcé présente des éléments positifs dès lors qu’un certain nombre de précautions ont été prises.
Esquemas de préstamos forzosos
Este trabajo consiste en el estudio de una serie de esquemas mediante los cuales los gobiernos en años recientes han obligado al sector privado a que les proporcionen préstamos.
En la Sección I se define el campo exacto del estudio. Se entiende que un préstamo es forzoso cuando cada una de las personas comprendidas en una amplia clasificación estáan obligadas a entregar, por un cierto plazo, al gobierno una suma de dinero, en el entendido de que dicha suma les séra devuelta con relativa prontitud.
En la Sección II y en el Apéndice se exponen las principales caracteristicas de los esquemas que se han encontrado en funcionamiento durante los últimos años en 19 distintos países. Se dan detalles, por ejemplo, sobre el périodo de funcionamiento, la base de las contribuciones, la duratión máxima del préstamo y la tasa de interés abonada.
La Sección III es analítica. Se efectúan comparaciones entre el prestamo forzoso y varias otras alternativas, especialmente la del impuesto sobre la renta en lo que respecta a sus efectos en cuanto a incentivos, equidad y conveniencia administrativa. Con referencia a los incentivos la conclusión general es que al préstamo forzoso probablemente se lo considere como una carga áas liviana que una tasa equivalente de impuesto sobre la renta; pero no existen conclusiones sencillas en cuanto a la equidad y a la conveniencia administrativa.
En la Sección IV se examina la experiencia práctica. En muchos casos estos esquemas constituyeron medidas de emergencia para hacer frente a una crisis financiera; sin embargo, se puede obtener un buen numero de enseñanzas.
En la Sección V se examina el future de este procedimiento. Una vez examinados todos los aspectos del esquema, la conclusión general es que algo puede decirse en favor del sistema de préstamos forzosos, especialmente si al establecerlo y administrarlo se toman varias precauciones.
Mr. Prest is professor of political economy at the University of Manchester. This paper was prepared while he was a visiting scholar in the Fund’s Fiscal Affairs Department during the summer of 1968. He is the author of Public Finance in Theory and Practice and Public Finance in Underdeveloped Countries, as well as other books and articles in economic journals.
See Franco Reviglio, “Social Security: A Means of Savings Mobilization for Economic Development,” and “The Social Security Sector and Its Financing in Developing Countries,” Staff Papers, Vol. XIV (1967), pp. 324-68 and 500-40.
It could also be argued that social security schemes differ from compulsory lending in that repayments do not necessarily bear any close relation to inpayments. Employer and government contributions to such funds are common; and even if pension repayments were actuarially equal to employee inpayments for the totality of those concerned, there would be large discrepancies for individuals, depending, e.g., on the length of time elapsing between the date of joining a scheme and the date of death.
John Maynard Keynes, How to Pay for the War: A Radical Plan for the Chancellor of the Exchequer (London, 1940). The same proposal was made more or less simultaneously by H.O. Meredith in “The Finance of War,” The Political Quarterly, Vol. XI (1940), pp. 59-73.
See the discussion on policy problems in the United States at that time, e.g., Milton Friedman, “The Spendings Tax as a Wartime Fiscal Measure,” Kenyon E. Poole, “Problems of Administration and Equity Under a Spendings Tax,” and Carl Shoup, “Problems in War Finance,” in The American Economic Review, Vol. XXXIII (March 1943), pp. 50-62, 63-73, and 74-97.
For an appraisal of wartime experience, see Walter W. Heller, “Compulsory Lending: The World War II Experience,” National Tax Journal, Vol. IV (1951), pp. 116-28.
For an account of this device in earlier years in the U.S.S.R., see Franklyn D. Holzman, “An Estimate of the Tax Element in Soviet Bonds,” The American Economic Review, Vol. XLVH (June 1957), pp. 390-96.
See Alan T. Peacock, The Times, London, March 16, 1968, p. 13, and March 22, 1968, p. 25.
The Appendix summarizes some particular features of different country schemes.
Nepal is also exceptional in allowing payment in cash or in kind.
In fact, there were many illegal transfers before the necessary five years elapsed. See page 44.
See Nicholas Kaldor, “The Choice of Taxes in Developing Countries,” in Economic Development in Africa, ed. by E.F. Jackson (Oxford, 1965), pp. 156-66.
Meaning a loan to which people would have been willing to subscribe voluntarily at the prices currently ruling in factor and goods markets.
This can be seen by a numerical example. Suppose that a taxpayer has the choice of either paying 100 in tax in year 1 and in year 2, or paying tax plus a loan of 10 in year 1 which will be repaid with interest in year 2, with consequent repercussions on tax liabilities in the two years. Let us assume that the marginal rate of tax is 50 per cent and the gross rate of interest is 10 per cent. In year 1 his total liability in the second instance will be 105 (i.e., 95 tax plus 10 loan), and in year 2 it will be 94.5 (i.e., 105.5 tax minus 11 for repayment with loan interest). The question is then whether an individual prefers the combination of paying 100 in both years to that of paying 105 and 94.5 in successive years.
If interest payments and/or repayments have a lottery element in them, further questions of risk aversion (or preference) also arise.
See Carl S. Shoup, “Forced Loans,” Chapter XI, in Curbing Inflation Through Taxation (Tax Institute Symposium, New York, 1944).
See page 44.
See Nicholas Kaldor, An Expenditure Tax (London, 1955), and A. R. Prest, “The Expenditure Tax and Saving,” The Economic Journal, Vol. LXIX (1959), pp. 483-89.
For opposing views about the interest elasticity of savings, see Bent Hansen, “Tax Policy and Mobilisation of Savings,” in Government Finance and Economic Development (papers and proceedings of the Third Study Conference on problems of economic development, Athens, 1963), ed. by Alan T. Peacock and Gerald Hauser (Organization for Economic Cooperation and Development, Paris, 1965), pp. 143-55, and M.S. Feldstein and S.C. Tsiang, “The Interest Rate, Taxation, and the Personal Savings Incentive,” The Quarterly Journal of Economics, Vol. LXXXII (1968), pp. 419-34.
With the consequence, however, that, insofar as a reduction in corporate saving in turn affected corporate investment, there would still be a deflationary effect from a tax increase.
It should be noted that Reviglio, on page 341 of “Social Security: A Means of Savings Mobilization for Economic Development” (cited in footnote 1), was able to cite some evidence that compulsory social security schemes enlarge the total of personal saving.
It should be noted that Heller’s conclusion was that incentive effects of compulsory lending in the United Kingdom and Canada in World War II were not substantial; see Heller, op. cit.
Keynes, op. cit., Chapter VII, pp. 44-51.
Three further points need to be made on this. First, insofar as older people own more capital per capita than do the young, this element of transfer might work in the wrong direction from a short-term egalitarian viewpoint (a life-cycle view might be a different matter). Second, a system of compulsory lending is not the only way of securing such a redistribution. One could finance some extraordinary expenditure by income taxation and still have a subsequent levy and grant system that would secure the same distribution between “old” and “young.” However, this would be a complex operation administratively; and the inescapable necessity to finance repayment is much more likely to lead to action. Third, we abstract from the question whether compulsory lending is more likely than income taxation to pass on a burden to future generations in the classical sense, i.e., by leading to a lower rate of growth of the capital stock in an economy.
See Richard A. Musgrave, The Theory of Public Finance: A Study in Public Economy (New York, 1959), p. 567.
Insofar as progressive taxation leads to a larger tax/income ratio in an inflationary situation.
It should be remembered that curtailment of private saving in developing countries may have repercussions on the level of private investment or on the level of capital exports.
For example, with interest paid annually the present value of the repayment stream will be
and that of the finance of repayment stream
For example, it would depend on whether the sector covered by compulsory lending is growing, relative to other sectors of the economy.
Jørgen R. Lotz and Elliott R. Morss, “Measuring Tax Effort’ in Developing Countries,” Staff Papers, Vol. XIV (1967), pp. 478-99.
Only 11 countries, out of the 19 covered in this paper, were among the 52 covered by Lotz and Morss.
If it is felt that crude tax-income ratios, unadjusted for per capita income or openness, are more appropriate, we find 3 in the highest group, 5 in the middle one, and 3 in the lowest. But in our view the adjusted ratio is the more relevant one in this context.
It should be added that the Canadian one was abandoned earlier than originally intended.
See Ghana, Report and Financial Statements by the Accountant-General and Report Thereon by the Auditor-General for the Year Ended 30th September, 1962 (Accra, 1964) and Report and Financial Statements by the Accountant-General and Report Thereon by the Auditor-General for the Year Ended 30th September, 1963 (Accra, 1965).
Law No. 223 of 1961.
Law No. 980 of 1967.
Gertrude E. Heare, Brazil: Information for United States Businessmen (U. S. Department of Commerce, Bureau of International Programs, Washington, 1961).
Law 1518, November 24, 1951, and Law 2973, November 26, 1956.
See Heller, op. cit.
The point could be made, however, that real private investment and consumption are likely to grow faster in a nonwar situation, thereby making easier the financing of repayments of compulsory loans.
Quoted by R.S. Sayers in Financial Policy, 1939–45 (London, 1956), p. 81, fn. 2.
The merits of compulsory loans have perhaps been exaggerated on some occasions. For instance, “… the scheme outlined must not be considered in the same way as taxation, since the amounts taken are fully repayable with accrued interest. The contributor will benefit doubly—by saving for his own and his family’s future and by helping to make a future for himself and his family by the country’s development,” 1962 Budget Speech of the Minister of Finance of British Guiana, Sessional Paper No. 2/1962 (Georgetown, 1962), p. 28.
Practices vary in this respect (e.g., India has treated them as capital, and Pakistan as current receipts).
Juanita D. Amatong, “The Revenue Importance of Government Lotteries” (unpublished paper, April 16, 1968).
If the idea is acceptable, it can be applied to redemptions as well as interest payments, i.e., those successful in the lottery would get not only interest in this form but also an early repayment of capital. This technique was used for many years in the U.S.S.R.; see Holzman, op. cit.
For an account of recent Israeli experience, see Amotz Morag, “Value-Linking of Loans in Israel,” Fiscal and Monetary Problems in Developing States, ed. by David Krivine (New York, 1967), pp. 148-60.
There would be no question of discounting them at the central bank, of course.
This was originally suggested for the 1940 U.K. scheme.
In such a situation, and with a progressive income tax system, people with rising incomes through time are particularly caught.
The definitive answer on this issue was given in the debate on the U.K. Finance Bill in 1941, when the Treasury view, in response to questions, was stated to be that such moves must be rejected on the grounds that they would represent “the encroachment of the tide,” Parliamentary Debates, House of Commons Official Report (London, 1941), Vol. 372, No. 78, Cols. 1248-53.