SDRs and the Working of the Gold Exchange Standard1

THE SPECIAL DRAWING RIGHTS scheme is designed to provide a supplement to existing reserve assets in order to ensure an appropriate growth in global liquidity. Article XXIV, Section 1(a), of the Fund Articles of Agreement states:


THE SPECIAL DRAWING RIGHTS scheme is designed to provide a supplement to existing reserve assets in order to ensure an appropriate growth in global liquidity. Article XXIV, Section 1(a), of the Fund Articles of Agreement states:

I. Introduction

THE SPECIAL DRAWING RIGHTS scheme is designed to provide a supplement to existing reserve assets in order to ensure an appropriate growth in global liquidity. Article XXIV, Section 1(a), of the Fund Articles of Agreement states:

In all its decisions with respect to the allocation and cancellation of special drawing rights the Fund shall seek to meet the long-term global need, as and when it arises, to supplement existing reserve assets in such manner as will promote the attainment of its purposes and will avoid economic stagnation and deflation as well as excess demand and inflation in the world.

Thus, the volume of allocations of special drawing rights (SDRs) is to be determined after allowing for the contribution to global reserve growth made by the expected development of other reserve assets; this was the procedure followed in the proposal by the Managing Director for allocation of SDRs for the first basic period, in the amount of 9½ billion during the three years 1970–72.2 Since the supply of SDRs is under administrative control in a way that the supply of other reserve assets is not, the availability of SDRs as a supplement for those assets provides a degree of control over the growth of total liquidity in the long term. The availability of SDRs does not provide an equivalent degree of control over the composition of global reserve growth. Yet, for the long term, this composition will clearly play a central role in determining the nature of the international monetary standard: the reserve system will have a different character if the role of SDRs is essentially or mainly to “back up” the growth of liquidity in other forms than if SDRs themselves are the predominant component of global reserve growth, and thus eventually also of the reserve stock.

This paper discusses the main influences involved in the relationship between SDRs and other reserve assets in a context of future reserve growth, and suggests certain general conditions that may be necessary for SDRs to become the predominant source of reserve growth. The question is considered in terms of expected relationships over the long run, extending beyond a single basic period of up to five years that is laid down as the normal period for a given rate of allocation of SDRs. The paper rests on the following main assumptions relating to other aspects of the international monetary system, as these operate after the introduction of SDRs.

(1) The system contains three reserve assets—gold, dollars, and SDRs (although other foreign exchange assets may be substituted for dollars as a result of endogenous forces operating within the system).

(2) The amount of gold in the system remains fixed in volume.3

(3) The price of gold in terms of currencies is not changed on a general basis, i.e., other than as a result of upward and downward changes in individual currency parities.

(4) Countries regard SDRs as an acceptable substitute for gold as an increment to their existing reserves. SDRs share, by design, certain attributes of gold as a reserve asset: they are defined in units of a fixed gold weight and do not constitute the liabilities of any particular country; in other obvious respects, monetary claims represented by special drawing rights differ from gold metal in their characteristics as reserve assets. The argumentation presented below rests on the widely held assumption that the characteristics of similarity will be sufficient to induce countries to regard SDRs as an acceptable substitute for additions they would otherwise wish to make to their reserves in gold. Countries may continue to have a strong preference for holding some absolute minimum of their reserves in gold itself. The central relationship to be considered is therefore between the growth of SDRs and gold, on the one hand, and reserves held in foreign exchange, on the other hand.

(5) It is assumed that, as a result of the increased elasticity given to the world reserve system by the introduction of SDRs, countries other than the United States adjust, over the long term, their holdings of dollars or other foreign exchange reserves to their portfolio preferences, so that they avoid or eliminate felt excesses of dollars in their reserves. This in turn assumes that financing of U. S. balance of payments deficits is not open ended, and that the prevalence of such deficits would spur adjustment—if not by countries other than the United States in response to reserve accumulations, then by the United States in response to reserve losses. If, instead, and over the long-term period under consideration, countries other than the United States took no steps to regulate the level of their dollar reserves when these exceeded some desired minimum, accepting such dollar flows as resulted from a deficit in the U. S. balance of payments, and neither they nor the United States took measures to adjust the underlying reciprocal balance of payments, then global reserve creation would be on a fundamentally different basis than in the system assumed in this paper. The pace of reserve growth would be almost wholly dependent on the U. S. balance of payments, although the potential availability of SDRs, if the SDR scheme survived such a regime, would provide a floor for global reserve growth.

(6) The general assumption presented in (5) is based on the following broad view. While the SDR scheme will not itself assure that the system works in this way—i.e., that the United States operates in the system as a large country rather than as the system’s prime reserve banker—the SDR scheme or something analogous to it permits such an option and may demand it for the long-run acceptability of the international reserve facility itself. In the absence of a mechanism for reserve creation similar to the SDR scheme, and in the absence of a general increase in the price of gold, expansion in global reserves would be dependent on continued expansion in international holdings of a reserve currency, such as the U. S. dollar. There would then be pressure for the reserve currency to become the primary reserve asset of the system, because expansion of the system could not continue if the supply of reserve currency had to be curtailed to assure continued convertibility into a reserve asset, such as gold, that was in inadequate supply. With the introduction of SDRs, global reserve growth is no longer dependent on expansion of reserve currency holdings. Through additions to net global reserves in the form of SDR allocations, room can be created to absorb adjustment of payments deficits of the reserve center, which need no longer be at the expense of reserve losses of other countries. It follows that the validity of a “dollar-centered” view of the international monetary system, e.g., that associated with the Despres-Kindleberger-Salant school [2],4 has been weakened by the establishment of the SDR scheme. (Adherents of this school who argue that pursuit of a “passive” balance of payments policy by the United States is consistent with evolution toward a “pure SDR standard,” with all reserve growth over time taking place in SDRs [Krause, 14, pp. 356–57], imply an improbable world in which countries other than the United States not only undertake the whole decision cost of adjustment but also adjust to the extent necessary to keep their official dollar holdings stable over time.) The opportunity costs of a dollar standard have become greater with the availability of a more attractive alternative.

Thus, it is assumed that, over the long term, countries other than the United States decide on their preferred ratio for accruals to their reserves as between dollars and SDRs, and that allocations of SDRs adjust over the long term to the flow of dollars resulting from the application of that ratio to the U. S. payments deficit or surplus, plus any accrual of other foreign exchange into reserve holdings. The central focus of the paper is the influence exerted on the relative demand for foreign exchange assets (i.e., against gold or gold-like reserve assets) by the establishment of the SDR scheme itself.

II. The Cycle of the Gold Exchange Standard: Pre-SDRs

Increases in reserves in U. S. dollars and other currencies have in the past occurred under the regime known as the gold exchange standard, under which a large number of countries have held substantial proportions of their reserves in a national currency convertible by them into gold at a fixed price. The central question considered in this paper can be approached by asking in what ways the availability of SDRs as a supplement to other reserve growth should be expected to influence the basic forces operating under the gold exchange standard. This question will in turn be approached by recalling the nature of these basic forces, which led to the buildup first of sterling and then of the U. S. dollar as a major component of international reserves, and of the accompanying strains on the international reserve system, which led, inter alia, to the establishment of the SDR facility.

The gold exchange standard, considered as a reserve standard under which a large number of countries accumulate reserve balances in a national currency that is in turn convertible into gold at a fixed rate, has often been shown to be dynamically unstable. The following interpretation draws, selectively, on the analyses of Triffin, Kenen, Holtrop, Johnson, Gilbert, and Officer and Willett.5 The internal working of the gold exchange standard involves a cycle of expansion, followed by strain, followed by crisis (or transmutation). The cycle is set off by the fundamental characteristics of the gold standard itself, of which the gold exchange standard may be considered a natural outgrowth. The underlying cause of the accumulation of reserve balances in a national currency lies in the property of gold as a noninterest-bearing asset (and as a cumbrous means of payment). This makes it especially advantageous for countries to hold balances in a national currency, as long as this offers associated advantages of convenience and does not involve undue risk. These latter provisos initially will tend to channel such holdings to the currency of the dominant economy internationally; this will tend to maximize convenience where that currency is internationally the most widely used for transactions purposes, and will minimize risk of losing value in terms of available supplies of goods as well as eliminating exchange risk on indebtedness expressed in that currency.

Where the currency of the dominant economy in the world is in a position of initial strength (in the sense that no serious expectations exist that its value is likely to decline in the discountable future in relation either to gold or to other important currencies), this set of circumstances will initiate a phase of expansion for the gold exchange standard. Countries other than the reserve center country will accumulate a substantial portion of accruing reserve gains in the currency of that reserve center. In the ensuing analysis, the reserve center country will be referred to as the United States; the analysis assumes the presence of only one active (i.e., expanding) reserve center at a time, an assumption that conforms broadly to the position since World War II.

The practices of individual countries in determining their reserve portfolios may vary considerably, according to the weight given to the advantages of interest earnings and convenience offered by reserves invested in the reserve center in contrast to a variety of contrary considerations (these may include the desire to maintain reserves in physical form, and perhaps also under national territorial control; the desire to avoid any risk of depreciation vis-à-vis gold; and the general desire to avoid undue dependence on decisions taken by the authorities of the reserve center country). For countries other than the United States, the aggregate ratio of reserve gains retained in the form of dollars will be referred to as the “dollar proportion,” or α (it is in fact the marginal dollar proportion). This is a key ratio in the analysis that follows. It is reflected in Chart 1 in the slope of the thick line, which represents the path of reserve composition for countries other than the United States. The level of reserves for these countries is indicated in the chart by the equi-reserve diagonals, at the points cut by the thick line.

Chart 1.
Chart 1.

Countries Other Than Reserve Center: Path of Reserve Composition

Citation: IMF Staff Papers 1971, 002; 10.5089/9781451947335.024.A001

In the initial, expansionary phase of the gold exchange standard, this curve, OA, will have a gentler slope than in later stages of the cycle, because at this stage the risk of capital depreciation of the reserve currency vis-à-vis gold will be small, zero, or conceivably even negative (if appreciation of the reserve currency vis-à-vis gold is in prospect); accordingly, the interest return on currency balances will exert a strong pull. Admittedly, this pull itself, by attracting funds into holdings in the reserve currency, will tend to reduce the interest rate on these holdings below what it would otherwise be. But in a major domestic money market, the interest rate will be determined primarily by a number of other influences, including the policy of the authorities and the condition of the domestic economy, which will not be equally connected with the prospect of a change in the currency’s relationship vis-à-vis gold. (To take an extreme example, in the case of an expected appreciation vis-à-vis gold, one would not expect the interest rate to be negative, in the absence of deliberate policy measures to that effect.) The analysis in this paper therefore assumes that changes in expectations concerning the future depreciation or appreciation of the reserve currency against gold will not be fully compensated by changes in the interest rate on reserve currency balances; accordingly, such changes in expectations will be reflected, other things being equal, in changes in the quantity of reserve currency outstanding.

The level of reserves outside the United States will be determined by other countries’ collective balance of payments with that reserve center, including as a credit item their net acquisition of previously nonmonetary gold (i.e., production of newly mined gold minus flows of gold into private holdings or plus private dishoarding). The reserve position of the United States, and the movement of global reserves, will depend on (1) the U. S. balance of payments, (2) the amount of newly accruing monetary gold in the world, and (3) the rest of the world’s dollar proportion (change in dollars as a percentage of change in the rest of the world’s total reserves), which along with (2) will determine the financing of the U. S. payments balance.6 These three determinants, which are in part interconnected in the ways noted below, are influenced by the internal structure of the gold exchange standard.

In a growing world economy with a long-run tendency to rising prices, accruals of monetary gold will tend to decline over time. While the impact of global inflation on final consumption of gold in nonmonetary uses and on gold production is probably smaller and less direct than is often suggested,7 the likelihood of this development occurring at some stage will give rise to anticipatory speculative influences, which may themselves become dominant, so that the precise incidence of effects on production and on final consumption will be of less consequence than otherwise.8

As accrual of new monetary gold decelerates, expansion of reserves of countries other than the United States will depend increasingly on their earning a collective surplus with the United States. Insofar as the increased pressure on other countries to earn surpluses at the expense of the United States is not resisted, these tendencies will involve a deterioration in the U. S. balance of payments and in its net reserve position (gold reserves in relation to dollar liabilities). Additions to gold stocks of other countries will then come increasingly from the existing gold stock of the United States rather than from new supplies.9 The reduction in U. S. gold reserves and the associated decline in the ratio of gold reserves to dollar liabilities is likely to cast increasing doubts on the continued convertibility of the dollar into gold at its existing parity; other things being equal, this in turn will induce other countries to reduce α in order to raise the proportion of gold in their reserves.10

The system will then enter Phase II, of stress, with reserves of other countries following some path, such as AB, in Chart 1. As expectations of gold revaluation add to speculative and precautionary demand among both official and private holders, the drain on the gold stock of the United States (shown in Chart 2) will give rise to increasing and self-aggravating pressures. At a point on Chart 1, such as B, continuation of such pressures becomes apparently insupportable, as private and official holders of dollars judge that the further prospective decline in the U. S. gold stock and in the ratio of this stock to dollar liabilities would oblige the U. S. authorities to suspend gold convertibility at the existing rate.

Chart 2.
Chart 2.

Global Gold Accrual and Gold Reserves of the United States and Other Countries

Citation: IMF Staff Papers 1971, 002; 10.5089/9781451947335.024.A001

The system will then enter Phase III, where two very different outcomes are possible. On one course, (a), the system may follow the crisis path BC (dotted), with conversions of dollars into gold forcing suspension of convertibility at the fixed gold parity; a straightforward revaluation of gold could then conceivably begin the cycle at a new expansionary phase, although many variations are possible, and analogies with the period after 1934 (when something like this did happen after a considerable time lag) must allow for the influence of a learning process.

An alternative path (b) from point B, which corresponds to the path followed in the 1960s, reflects the fact that the international monetary system is dominated by a small number of decision-making units that can and do allow for the consequences on other units of their own actions. In this sense, it has been pointed out, behavior of monetary authorities in their reserve policy and their decisions on portfolio adjustments is better explained by reference to decisions under oligopoly (and by small numbers of players in “mixed motive” games) than by reference to atomistic investors seeking to maximize returns or minimize risks in a perfect capital market [16 and 17].

Thus, from about the early 1960s, a number of large reserve holders exercised restraint in reducing their dollar proportion, this restraint reflecting self-interest (allowing for reactions) as well as pure “cooperation.” At B on Chart 1, countries other than the United States may have an unconstrained portfolio preference shown by the slope of OR but refrain from moving to this position, which would involve converting BX dollars into XY gold. Moreover, insofar as further increases in their reserves even at the existing average dollar/gold ratio OR′ would involve unacceptable losses of gold by the United States, the dollar proportion actually adopted for ensuing reserve gains may be bent further toward the horizontal. Alternatively or additionally, ways may be found to increase reserves involving neither the risk for other countries of holding dollars nor the strain on the United States of giving up gold—witness the flowering of credit arrangements of the 1960s, through the development of central banking swaps, issuance of monetary claims denominated in creditor currencies (Roosa bonds), and increased recourse to credits from the International Monetary Fund.

These developments created new reserve assets of varying quality but carrying the relevant advantages of protecting the gold stock of the United States while offering “better-than-dollars” security to the creditors. These expedients are sometimes said to have weakened the discipline of the system, particularly as it applies to the financing of deficits of the United States, and to some extent also of the United Kingdom [5 and 6]. However, the criterion for such a judgment is unclear: while the United States would be permitted by these expedients to finance a given deficit with a smaller gold loss than would be entailed in these circumstances in the absence of collective or “oligopolistic” restraints on gold conversion, the proportionate gold financing may still be larger than in an earlier, expansionary phase of the gold exchange standard.

The real problem in the phase of expedients and restraint may be less the softness of financing by the United States than the indeterminacy in the degree to which sacrifices in preferred financing media, undertaken to prevent a crisis of the system, should be borne by the United States, on the one hand, and by its creditors, on the other hand. In this phase, the system for the first time becomes “managed” but lacks a management rule book.

III. Impact of SDRs

The SDR scheme itself may be seen as a culmination of the earlier, ad hoc expedients. It allows countries to increase their reserves in forms other than U. S. dollars without draining the gold stocks of the United States or of any other country. However, the impact of the SDR scheme on long-term portfolio behavior of reserve holders is likely to be far more extensive than the preceding ad hoc expedients, for two reasons. First, SDRs provide a creditor claim of a more “gold-like” quality than swap claims and Roosa bonds, being denominated in gold value and offering full liquidity and effective transferability for the settlement of payments deficits, although also (in conformity with their more gold-like characteristics) carrying an interest rate that is currently well below rates on foreign exchange or swap claims.11 Second, by its intended permanence as a supplement to alternative sources of global liquidity, and by providing a source of additional “net” or primary reserves unencumbered by associated short-term monetary liabilities, the SDR scheme removes or reduces, in a way that ad hoc schemes could not, the prospect that a shortage of primary reserves must eventually be redressed by increasing their valuation. For this reason, the SDR scheme reduces the pressures for an increase in the price of gold in terms of currencies in general. The supply of SDRs, unlike the physical supply of gold, and unlike the supply of reserve claims stemming from the granting of credits decided upon independently, is fully under administrative control. To the extent that SDRs are acceptable as increments of international reserves as an alternative to additional gold, this element of control over the potential quantity of international reserves reduces or eliminates a future need to act on price in order to achieve the desired effects in terms of value.

The establishment of the SDR scheme itself, therefore, may introduce a major change in the characteristics of the gold exchange standard. To the extent that SDRs are regarded as a substitute for additional gold, the SDR scheme provides in effect an infinitely elastic potential supply of gold substitute at the fixed price. The effect of this on the demand for and supply of dollars and other actual or potential foreign exchange assets is likely to be profound.

This impact may be illustrated by referring back to point A on Chart 1 and the ending of the “expansionary” phase of the gold exchange standard. This check was attributable to the slackening of new gold accrual and to consequential pressures on the U. S. balance of payments in the tightened global system. The expansionary phase would be stopped in a different way if the United States succeeded in resisting or reversing such pressures on its payments balance, as this would check reserve growth of other countries short of the diagonal passing through A. This is the second prong of the Triffin dilemma, under which continued expansion under the gold exchange standard necessitates either an unsustainable weakening in the reserve position of the United States or a cessation of global reserve growth. The availability of SDRs breaks this dilemma: global reserve growth is no longer dependent on a U. S. deficit (although still, as discussed below, influenced at least in its composition by the U. S. payments balance, together with its financing as determined by α).12 Equally, for any given U.S. payments position, the assurance of adequate alternative liquidity, in the amount that this may be demanded in the future to make up a given target of global reserve expansion, greatly diminishes the risk of a given depreciation of dollar assets in terms of gold (and now also SDRs). The risk may be reduced to zero in the event of a sustained U. S. payments equilibrium and may become negative (implying the prospect of appreciation) in the event of a sustained U. S. surplus, which the potential availability of SDRs would now make compatible with a tolerable reserve situation in the rest of the world. Even a continuing positive risk of depreciation of dollars vis-à-vis gold would be likely to be related to a smaller degree of depreciation, connected with a possible rise in the dollar price of gold designed to effect a depreciation vis-à-vis a small number of other currencies rather than with a rise in the dollar price of gold designed to restore a viable relationship between the dollar (among other foreign exchange assets) and gold as the components on which a continuing expansion of global reserves rested. (The prospect of a depreciation in the value of the dollar in relation to other currencies may of course in itself reduce demand for dollars, compared with a situation in which no such expectations were held; but in these circumstances a partly or wholly offsetting increase in holdings of the prospectively appreciating currencies would be likely.) In all these respects, the position will tend to revert to that under the expansionary phase of the gold exchange standard.

On these grounds, the introduction of SDRs must be expected to increase countries’ relative appetite for dollars, and for foreign exchange in general, compared with the postexpansionary phases of the gold exchange standard, and to flatten their desired OR lines.13 At point B on Chart 1, the “unconstrained” OR line may shift from OR to OR″; this would turn the earlier potential excess of dollars, BX, into a deficiency, BZ, as countries other than the United States shifted ZD of gold plus SDRs to the United States and moved to D. Conceivably, these countries could move further down the diagonal BD, or its equivalent, after an increase in reserves; their preferred portfolio position could be pushed below a continuation of OA. Thus, the critical bend at A could be eliminated or turned back toward the base line, involving an increasing component of dollars (or eventually of other foreign exchange) in global reserves.

Of course, the outcome will be affected by a number of factors aside from the changed risk situation, and different countries will act differently. Thus, many countries may set some minimum for their gold and SDR holdings, and in some cases for their gold holdings or SDR holdings alone.14 Considerations of portfolio balance—with weight given not only to the interest return on assets, and to allowance for particular risks and uncertainties, but also to achieving a spread of holdings and risks—will affect the size of the response of reserve composition to given changes in the interest return or in the imputed risk on particular assets. But all the various factors that go to determine national preferences in reserve composition would be likely to be affected in some degree and over time by a decrease in uncertainty about the relative valuation of foreign exchange reserves vis-à-vis gold-SDR reserves.

The reduction in uncertainty over the future relative valuation between gold and currency values that can be expected to result from establishment of the SDR scheme has two contrasting effects in terms of stability of the reserve system. In the short term, without doubt, it increases stability, in the sense of taking the system further from its possible crisis path, BC on the chart.15 But over the longer term, the removal of uncertainty and avoidance of the potential upward bend in the reserve path at A on the chart also removes the earlier check on unbridled expansion of the gold exchange standard, and in this sense removes an earlier “control” in the system.16

IV. The Argument in Brief

The argument thus far can be recapitulated as follows. A system in which the basic unit of reserve holding is gold will tend to develop into a system with ancillary reserve holdings in national currency. This tendency stems primarily from the property of gold as a noninterest-bearing asset. In phases in which gold does not appear undervalued vis-à-vis major currencies, and accordingly offers little or no implicit return through potential appreciation, many monetary authorities feel it profitable to hold at least a substantial portion of reserves in interest-bearing balances in a currency convertible into gold rather than in gold itself. Subsequently, gold accrual is likely to slacken in an expanding world economy in which prices are rising; the accumulation of balances in the reserve center will then involve a weakening in its reserve position and will cast doubt on continued convertibility of the reserve currency into gold at the existing parity. The consequential prospect of appreciation of gold in terms of the reserve currency will provide gold with an interest equivalent that will eventually become higher than rates on reserve currency balances, which will be affected primarily by other influences. This process will involve either a crisis of the system, perhaps including a formal revaluation of gold, or (as happened in practice in the 1960s) the application of restraint in policies of gold conversion, and the development of various credit substitutes, leading eventually to the introduction of an international reserve asset designed as a supplement to gold, and endowed for that purpose with a gold value guarantee. However, insofar as introduction of such an asset, which has taken the form of SDRs, removes or substantially reduces the prospect of a revaluation of gold, this will remove or substantially reduce the implicit interest rate on gold and the interest equivalent of the gold value guarantee on the international asset. In this measure, gold, and the new asset in the absence of an increased explicit interest return, must be expected to become less attractive vis-à-vis assets bearing market rates of interest than would be true with continued uncertainty over a general revaluation of gold.

As a result, insofar as these influences hold sway, the cycle of reserve currency expansion will tend again to be set in motion. In the event that holdings of existing reserve currency become unattractive because of risk of depreciation against other currencies, corresponding pressure will be exerted for a buildup of reserve balances in the latter currencies, on which the prospective rate of return looks attractive vis-à-vis gold and SDRs. Such tendencies may be restricted to some extent by deliberate national or international policies of restraint in the holding of reserve currencies; but powerful forces of national self-interest, and to some extent also of economic efficiency, will be operating in the other direction. These forces, consisting essentially of the attraction of available interest earnings on reserve assets without an offsetting cost in terms of security or portfolio imbalance, may be particularly compelling for countries that are substantial borrowers of funds on international markets. Such countries have traditionally limited the net interest burden and exchange risk involved in such borrowings by holding a substantial part of their reserves in money market investments designated in the borrowed currency.

More generally, the absence or insufficiency of interest return on an asset such as SDRs may leave the international monetary system exposed to the risk of polarization in the gold exchange standard diagnosed by Holtrop. Portfolio equilibrium between an interest-bearing and a noninterest-bearing asset17 is dependent largely on the prospect of a compensating change in their respective capital values, which will be the dominating influence determining preference for the one asset compared with the other. For any particular portfolio mix of the two assets, therefore, the system is in equilibrium only at one particular point of uncertainty regarding these prospective capital values; in the absence of uncertainty, the system will be in equilibrium only up to such ratio of the noninterest-bearing asset as is regarded as a minimum on grounds of portfolio spread. The smaller the speculative expectation of changes in relative valuation among reserve assets, the greater will be the pull of relative interest rates.18

This influence of relative interest rates on reserve composition will often exert its impact only over a long timespan, for the reasons touched on above. This will be particularly so for the impact on holdings of SDRs. Under the Fund’s Articles of Agreement, allocations of SDRs are decided on for basic periods of up to five years and can be changed within the chosen basic period only in the event of “unexpected major developments.” Within the basic period, therefore, the amount of SDRs will be predetermined by the earlier decision regarding their incremental supply. The amount of SDRs held by countries other than the United States may admittedly be influenced by these countries’ preference for SDRs instead of dollars, as influenced by the relative interest rate, inter alia, because stronger demand for SDRs may make countries more inclined to request conversions of dollars into SDRs. Such conversions would, of course, be limited by the amount of available SDR holdings of the United States. Such conversions also require the agreement of the U. S. authorities; and the United States, like other participants, can generally use SDRs in transactions with other countries only where it has a balance of payments need or in the light of developments in its reserves.19 Under the rules in force for the first basic period,20 the United States, like other participants, is also required to maintain over time SDR holdings of at least 30 per cent of its net cumulative allocation.21 A more general provision requires participants to pay due regard to the desirability of pursuing over time a balanced relationship between their holdings of SDRs and their other reserves.

The effect of “demand” factors on holdings of SDRs as against U. S. dollars by countries outside the United States will therefore be limited within any basic period. But over a longer period—and it is with long-term influences that this paper is concerned—these demand factors may play a notably larger role, through their effect on the size of SDR allocations. This effect may be transmitted in two ways. First, to the extent that a preference for dollars has made the net transfer of gold and SDRs to the United States larger (or the transfer of these assets from the United States smaller) than it would have been with a lower preference for dollars vis-à-vis gold and SDRs, this will entail correspondingly larger growth in actual reserves in the current basic period; accordingly, an offsetting reduction in the SDR allocation in the subsequent basic period will be necessary if a given target trend for long-term growth in global reserves is to be followed. Second, a preference for other reserve assets over SDRs will tend to make creditor countries more cautious in agreeing to SDR allocations of given size, on a prospective view both of the possible short-term “cost” to them of exchanging preferred for less-preferred reserve assets and of the likely emergence, in these conditions, of reserve growth in forms other than SDRs. If assets other than SDRs are preferred as creditor claims, fewer SDRs are likely to be created.

V. Composition and Distribution of Reserve Increments with SDRs

The amount of dollars accumulated in reserves will, of course, depend not only on the proportion of reserve changes of countries other than the United States that take the form of dollars (α) but also on the “initial” flow of dollars to foreign official holders as determined by the U. S. balance of payments: it will be a combined function of these two influences. If the SDR allocation is set so as to supplement the resulting supply of dollars in the amount necessary to reach a target rate of change in global reserves,22 and if these are the only sources of reserve change,23 then the components of reserve changes and the distribution of such changes among the United States, on the one hand, and other countries, on the other hand, will be determined by three variables: the target increase in reserves, the U. S. payments balance, and the “dollar proportion,” α. The shares of the United States and other countries in the SDR allocations are assumed to remain fixed at one fourth and three fourths, respectively, which is roughly the present position. The system is summarized in the equations set out at the head of Table 1.24

Table 1.

Determination of Reserve Composition and Reserve Distribution Where SDRs Are Allocated to Achieve Target Global Reserve Increase, with Zero Gold Accrual

article image

For any given target increase in global reserves under this system, the SDR allocation will be a combined function of the U. S. payments balance and the dollar proportion. As the U. S. payments balance varies over time, the variability and amount of the SDR allocation will then depend on two influences: (a) the average size of the dollar proportion over time (the lower it is, the less variable and the larger 25 the SDR allocation) and (b) the variations in the dollar proportion in relation to the U. S. payments balance: thus, if the dollar proportion moves inversely with the U. S. deficit, this will tend to stabilize the dollar flow, and therefore also the SDR allocation; by contrast, variations in the dollar proportion in the same direction as the U. S. deficit will exacerbate variations in the dollar flow and in SDR allocations, while reducing fluctuations in U. S. holdings of reserve assets.

Under this system of setting SDR allocations as the estimated residual needed to achieve a target increase in global reserves over the medium term, the dollar proportion will also determine the size of the reserve change of countries other than the United States, for any given U. S. payments balance. Thus, an increase in the dollar proportion will not only raise the proportion of dollars to SDRs and gold in reserve accruals of other countries but also reduce their total reserve gain, reflecting the reduction in the SDR allocation that will be necessary to allow for the additional increment of dollars (Table 1, line 5). Equally, an increase in the U. S. deficit with no change in the dollar proportion will increase reserves of countries other than the United States by less than the increase in their surplus (the inverse of the U.S. deficit), to the extent of their (three fourths) share of the reduction in the SDR allocation that will again be needed to offset the increment in dollars.

The following contrast may therefore be drawn between different relationships maintained by countries under this system of the dollar proportion and the U. S. payments balance.

(a) If the dollar proportion is constant over time, the U. S. balance of payments will determine the amount of SDR allocation, as well as the amount of reserve changes of other countries (which will fall short of the change in their balance with the United States), as well as determining the U. S. reserve position (which will change by less than the change in the U. S. payments balance on a gross basis, but by more than the change in the payments balance on a net basis—debiting increases in dollar balances—also reflecting the effect on SDR allocation). Where the dollar proportion is relatively high, the main impact of shifts in the U. S. payments balance on reserves of other countries will be on their dollar reserves; accruals to their reserves of SDRs and gold will be relatively small and will change relatively little (Table 1, line 7); the fluctuations in SDR allocations in relation to the U. S. payments balance will then influence mainly U. S. holdings of SDRs and the degree to which U. S. dollar liabilities are “backed” by reserve assets (gold plus SDRs).

(b) If, on the other hand, movements in the U. S. payments balance are wholly or partly offset in their impact on dollar accruals by shifts in the dollar proportion, or if the dollar proportion is generally low, the U. S. balance of payments will have a zero or small influence on SDR allocations; and reserve movements between the United States and other countries will be affected by close to the same amount as the change in the U. S. payments balance. The amount of SDR creation will therefore be more stable than under state (a), and holdings of SDRs and gold will be better spread between the United States and other countries as a group. Offsetting shifts between the dollar proportion and the U. S. payments balance, stabilizing the accrual of dollars, would be consistent with a relatively stable demand for dollars on the part of foreign official holders.

Under a policy of creating SDRs so as to make up a target level of global reserve increase, therefore, a change in the dollar proportion will no longer, by assumption, affect the amount of global reserve increases, as would happen in the absence of SDRs. The effect of changes in the dollar proportion will be on the composition of reserve increments, as between SDRs and dollars, and on the distribution of reserve changes (as to both size and composition) as between the United States and other countries (Table 2).26

Table 2.

Effect of SDR Regime on Interrelation Between U. S. Payments Balance, Dollar Proportion, Increments in Global Reserves, and Their Distribution

article image

Before turning to the question of how the dollar proportion may be influenced in a desired direction, some consideration must be given to the effect of SDR creation on the U. S. payments balance itself. This will be broadly as follows, for a given allocation of SDRs.

(a) The allocation to the United States will in itself tend to make U. S. domestic policies more expansionary and policies on external trade and capital outflows less restrictive; this will tend to cause a deterioration in the U. S. balance.

(b) The allocation to other countries will have similar effects on their policies; this will tend to improve the U. S. balance.27

In addition, certain effects of SDR creation on the U. S. payments balance may follow via an impact on the dollar proportion.

(c) Insofar as conversions of dollars into reserve assets held by the United States have hitherto been held back by the reluctance of creditor countries to bring about a reduction in the absolute level of U. S. reserves, a fall in the dollar proportion just sufficient to absorb the addition to the U. S. allocation of SDRs will offset effect (a), above, and thereby, on balance, improve the U. S. payments balance (but see (d), below); in addition, the greater stability of the international monetary system and increased leeway arising from the fact that a reduction in U. S. reserves no longer has a corresponding effect on global reserves may also permit a larger fall than this in the dollar proportion, i.e., involving a net reduction in U. S. reserve assets, and greater pressure for improvement of the payments balance, compared with the situation prevailing before establishment of the SDR scheme.

(d) While SDR allocations should lessen the pressure for countries other than the United States to exercise restraint in reducing their dollar proportion in line with individual portfolio preferences, the existence of the SDR scheme will in itself tend to raise the preferred dollar proportion of reserve increments, on the grounds discussed in Sections III and IV of this paper; if this tendency holds sway, a rise in the dollar proportion (compared with the pre-SDR regime) will reinforce the effect of the U. S. allocation of SDRs in strengthening the U. S. gross reserve position for a given payments balance. This effect might itself, at least for a time, tend to cause a deterioration in the U. S. payments balance.28

VI. Conclusions and Possible Implications

It has been pointed out earlier in this paper that two different paths can be envisaged for the development of the SDR system, on the assumptions made.29 On the first path, SDRs would become a major part of the increment and ultimately of the stock of reserves held in the world as a whole: this might be termed the SDR-gold standard. On the second path, SDRs would provide smaller and less regular supplements to a continued expansion in foreign exchange reserves, the basic function of SDRs then being to provide a sufficient flow of reserve assets to the reserve centers to permit continued expansion in their reserve liabilities without threatening the stability of the system through adverse confidence effects: this might be termed the SDR-gold exchange standard. This paper has also suggested that, as long as SDRs are given mainly gold-like characteristics, natural forces may pull toward the second path rather than the first.

If it were thought desirable to counter such a pull, in order to attain a more universal reserve standard and to avoid future problems following from a further buildup of reserve currency balances, by what means might this be achieved? This question deserves a paper of its own; in brief, the main choice seems to lie between three hypothetical courses of action indicated below, although they are not mutually exclusive. It may be recalled that an important underlying assumption of all three courses is a willingness on the part of the United States and other countries to adjust their payments balances so as to avoid or eliminate excessive accruals of dollars in official holdings. A different assumption would necessarily imply a quite different payments standard.

(1) The most radical solution would be a pooling of all reserve assets, aside from working balances, in a single international account, with provisions for the offsetting of any expansionary or contractionary impetus arising from further accumulation or reduction of official balances in national currencies, after an initial consolidation and with allowance for nationally held working balances.30 Such a solution may well be appropriate in the long run; however, it involves a greater internationalization of reserve holdings than countries yet appear willing to accept. Moreover, the more such a scheme for reserve pooling acted against what were seen as national interests by tying up reserves at interest rates significantly below those attainable in national money markets after allowing for differences in risk, etc., the greater would be the degree of international control that would need to be exercised in order to prevent leakages through nominally private holdings and other loopholes.

(2) A more modest approach would be for the major countries to come to formal or informal agreement relating to holdings of reserve currencies, with a view to limiting or stabilizing increments to global liquidity in this form. Such agreements would presumably need to extend to the main holding countries. Measures taken by the reserve centers alone (e.g., to reduce or eliminate interest on official balances) would be unlikely to be effective; thus, they could easily be sidestepped by the holding of reserves in the external (Euro) markets in these currencies. Another, weaker variant of this approach would be for major countries to continue to override their own portfolio preferences with considerations based on benefits for the system as a whole. However, the pressures for such adaptations seem likely to be notably weaker, and less easily brought to bear, than the pressures arising from a need to prevent a crisis of the gold exchange standard, such as have constrained portfolio preferences of individual countries in the recent past.

(3) The approach requiring the smallest element of international control would be to make the return on SDRs more attractive by comparison with that available on foreign exchange holdings, i.e., to raise the SDR interest rate. A substantial increase in this rate would involve a number of separate considerations, which can be given only summary consideration here. The most familiar and perhaps most feared impact would be an adverse distribution effect for less developed countries, which as a group have so far been net users rather than net additional acquirers of SDRs. However, the reverse is true for many individual less developed countries, and a number of developed countries have been net users. If net resource transfer from developed to less developed countries is desired through the SDR scheme, there would clearly be more direct and more efficient ways of attaining it than through an artificially low interest rate applied across the board.

One other important consequential effect may be noted briefly. If the interest rate on SDRs is to be aligned to maintaining full attractiveness of the asset vis-à-vis foreign exchange, this may leave gold unattractive vis-à-vis SDRs (as well as vis-à-vis reserve positions in the Fund, insofar as remuneration continues to be broadly aligned to interest on SDRs). In practice, this disequilibrium may cause only limited disturbance insofar as (a) some room exists for shifts between gold and reserve positions in the Fund, through sales of gold to the Fund under Article V, Section 6(a), and (b) the entrenched preference in the past of a number of countries for holding at least a minimum amount of gold in reserves may be expected to produce equilibrium in portfolios after only relatively small shifts of this kind.

Among the influences affecting the long-term appetite for SDRs vis-à-vis reserve currency holdings, the rate of return on SDRs is likely to be particularly important. The lack of interest return on gold has been seen as an underlying cause of the buildup of reserve currency balances. The advent of the SDR scheme is itself likely to increase the pull exerted by relative interest rates on reserve assets, to the extent that it diminishes the prospect of major changes in their relative capital values. It is too early in the life of the SDR scheme to predict with any confidence how great or how enduring such a shift in expectations will be, especially in view of the continuing strains on the international monetary system arising from the payments imbalance between the United States and the rest of the world. But it is worth emphasizing that the problems discussed in this paper will be the more pressing, the greater the degree of international monetary stability, and the more entrenched the SDR scheme becomes as the potential source of incremental reserves. In particular, adjustment of the existing imbalance in U. S. external payments might well make the problem a more active one, because the analysis suggests that the internal working of the system would thereafter encourage a renewed buildup in reserve currency liabilities and a corresponding relaxation of external pressures on the U. S. payments balance.


  • Aliber, Robert Z., “Gresham’s Law, Asset Preferences, and the Demand for International Reserves,” The Quarterly Journal of Economics, Vol. LXXXI (1967), pp. 62838.

    • Search Google Scholar
    • Export Citation
  • Despres, Emile, Charles P. Kindleberger, and Walter S. Salant “The Dollar and World Liquidity—A Minority View,” The Ecconomist, February 5, 1966, pp. 52629.

    • Search Google Scholar
    • Export Citation
  • Fleming, J. Marcus, “The Bearing of the Supply of Other Reserves on the Need for Special Drawing Rights,” International Reserves: Needs and Availability, Papers and Proceedings, Seminar at the International Monetary Fund, June 013, 1970 (Washington, 1970), pp. 32149.

    • Search Google Scholar
    • Export Citation
  • Fleming, J. Marcus, “The SDR: Some Problems and Possibilities,” Staff Papers, Vol. XVIII (1971), pp. 2547.

  • Gilbert, Milton, “The Gold-Dollar System: Conditions of Equilibrium and the Price of Gold,” Essays in International Finance, No. 70 (Princeton University, October 1968).

    • Search Google Scholar
    • Export Citation
  • Gilbert, Milton, “The International Monetary Problems of Today and Implications for the 1970s” (speech presented at the International Monetary Conference, North Carolina Bankers Association, Raleigh, North Carolina, October 8, 1970).

    • Search Google Scholar
    • Export Citation
  • Grubel, Herbert G., The International Monetary System: Efficiency and Practical Alternatives (Harmondsworth, England, 1969), pp. 12850.

    • Search Google Scholar
    • Export Citation
  • Hirsch, Fred, “Influences on Gold Production,” Staff Papers, Vol. XV (1968), pp. 40590.

  • Holtrop, M.W., Statement by the Governor for the Netherlands, Summary Proceedings of the Sixteenth Annual Meeting of the Board of Governors, 1961 (International Monetary Fund, Washington, 1961).

    • Search Google Scholar
    • Export Citation
  • Høst-Madsen, Poul, “Gold Outflows from the United States, 1958–63,” Staff Papers, Vol. XI (1964), pp. 24861.

  • Johnson, Harry G., “Theoretical Problems of the International Monetary System,” The Pakistan Development Review, Vol. VII (Spring 1967), pp. 128.

    • Search Google Scholar
    • Export Citation
  • Johnson, Harry G., “Roy Harrod on the Price of Gold,” in Induction, Growth and Trade: Essays in Honor of Sir Roy Harrod, ed. by W.A. Eltis, M.FG. Scott, and J.N. Wolfe (Oxford, 1970), pp. 26693.

    • Search Google Scholar
    • Export Citation
  • Kenen, Peter B., “International Liquidity and the Balance of Payments of a Reserve-Currency Country,” The Quarterly Journal of Economics, Vol. LXXIV (1960), pp. 57286.

    • Search Google Scholar
    • Export Citation
  • Krause, Lawrence B., “A Passive Balance-of-Payments Strategy for the United States,” Brookings Papers on Economic Activity, No. 3, 1970, ed. by Arthur M. Okun and George L. Perry (The Brookings Institution, Washington, 1971), pp. 33968.

    • Search Google Scholar
    • Export Citation
  • Niehans, Jürg, “The Flexibility of the Gold Exchange Standard and Its Limits” (presented at the Conference on Optimum Currency Areas, sponsored by Moneda y Credito and the Harris Foundation, Madrid, Spain, March 1619, 1970).

    • Search Google Scholar
    • Export Citation
  • Officer, Lawrence H., and Thomas D. Willett, “The Interaction of Adjustment and Gold-Conversion Policies in a Reserve-Currency System,” Western Economic Journal, Vol. VIII (1970), pp. 4760.

    • Search Google Scholar
    • Export Citation
  • Officer, Lawrence H., “Reserve-Asset Preferences and the Confidence Problem in the Crisis Zone,” The Quarterly Journal of Economics, Vol. LXXXIII (1969), pp. 68895.

    • Search Google Scholar
    • Export Citation
  • Triffin, Robert, Gold and the Dollar Crisis (Yale University Press, Revised Edition, 1961).

  • Triffin, Robert, “The Evolution of the International Monetary System: Historical Reappraisal and Future Perspectives,” Princeton Studies in International Finance, No. 12 (Princeton University, June 1964).

    • Search Google Scholar
    • Export Citation

Les droits de tirage spéciaux et le fonctionnement du système de l’étalon de change or


La présente communication examine les principales influences qui interviennent dans le rapport entre les DTS et autres formes de réserves dans un contexte de croissance future des réserves et propose certaines conditions générates qui peuvent se révéler essentielles pour que les DTS deviennent une source prédominate de croissance des réserves. On suppose, entre autres, que le volume d’or dont dispose le système demeure constant, que le prix de l’or exprimé en monnaies ne subit aucune modification générale et que, à long terme, les pays ajustent la composition de leur portefeuille d’avoirs en dollars et autres devises suivant leurs préférences, de sorte qu’ils évitent d’accumuler dans leurs réserves, ou qu’ils en éliminent, les quantités de dollars jugées excessives. Ceci, à son tour, sous-entend que les Etats-Unis jouent dans le système le rôle d’un grand pays plutôt que celui de banquier ultime des réserves du système et que, dans leur ensemble, les pays saisissent l’occasion qui leur est fournie par la facilité des DTS—ou tout autre procédé analogue—pour freiner l’expansion accentuée du rôle des monnaies de réserve.

Le présent document aborde l’analyse des effets que peut produire l’existence des DTS sur l’évolution des réserves de change en rappelant d’abord la nature des forces fondamentales qui ont conduit à la croissance des réserves en devises dans le système de l’étalon de change or. L’expansion des réserves en devises se trouve favorisée au premier chef par le fait qu’elles rapportent un intérêt et qu’elles facilitent les transactions. Mais, étant donné que les réserves en devises consistent en engagements de pays qui jouent le rôle de centres de réserve, la croissance de ces engagements au-delà d’un certain taux ou d’un certain multiple des avoirs en or dont disposent les centres de réserve fera douter de leur convertibilité future au taux de change existant et mena-cera par conséquent la stabilité du système. Dans une économie mondiale en expansion, marquée par une tendance à long terme à la hausse des prix, les accumulations d’or monétaire dans le monde pris dans son ensemble auront tendance à diminuer dans le temps; et, en l’absence d’une hausse générale du prix de l’or exprimé en devises, ceci imposera des limites à l’expansion tant des réserves totales de devises que des réserves totales, quelle qu’en soit la forme, pour l’ensemble du monde. C’est ainsi que le système de l’étalon de change or, par lui-même, annonce soit un volume inadéquat de réserves, soit des variations brusques de la valeur de l’or par rapport aux devises.

L’introduction des DTS, dans la mesure où ils sont considérés comme un substitut acceptable de l’or dans les accroissements des réserves, transforme la situation en créant, en fait, une offre potentielle infiniment élastique d’un substitut de l’or au prix fixé. La facilité des DTS lève ou attènue considérablement une grande incertitude qui pesait auparavant sur le système—l’éventualité d’une augmentation importante et générale du prix de l’or exprimé en monnaies—mais, du même coup, elle pourrait aussi éliminer une entrave antérieure à une expansion effrénée du système de l’étalon de change or et, en ce sens, supprimer un “contrôle” dont disposait auparavant le système.

La tendance du système des DTS à évoluer de cette façon dépendra de plusieurs facteurs. Mises à part les hypothèses générales dont il a été question précédemment, notamment la volonté des Etats-Unis et des autres pays d’ajuster leurs balances des paiements de façon à éviter ou éliminer des accumulations excessives de dollars dans leurs réserves officielles, une ou plusieurs des conditions suivantes peuvent s’avérer nécessaires pour que les DTS deviennent une composante essentielle de l’accroissement des réserves:

1) Une mise en commun des avoirs en devises existants dans un seul compte international, s’accompagnant de dispositions correspon-dantes destinées à éviter ou à annuler toute accumulation ultérieure de soldes officiels en monnaies nationales; 2) sur une échelle plus modeste, conclusion d’un accord, officiel ou non, entre principaux pays pour limiter la croissance de réserves additionnelles en devises; 3) un relèvement du taux d’intérêt versé sur les DTS pour les rendre plus attrayants par rapport aux réserves en devises.

Los derechos especiales de giro y el funcionamiento del patrón de cambio oro


En este trabajo se examinan los principales factores que influyen en las relaciones entre los DEG y otros activos de reserva desde el punto de vista del crecimiento futuro de las reservas, y se sugieren ciertas condiciones generales que pueden ser necesarias para que los DEG pasen a constituir una fuente preponderante de crecimiento de las reservas. Se da por sentado, entre otras cosas, que la cantidad de oro del sistema permanece a un nivel fijo, que, en términos generales, el precio del oro en función de las monedas no varía y que, a la larga. los países ajustan sus tenencias en dólares y en otras divisas de conformidad con sus preferencias en cuanto a la composición de sus carteras a efectos de evitar o eliminar de sus reservas lo que consideran como un exceso de dólares. Esto, a su vez, entraña la participación de Estados Unidos en el sistema en su carácter de país grande, pero no como banquero de reservas, y que los países en general aprovechen la oportunidad que les brinda el sistema de DEG, u otro expediente análogo, para limitar la importancia de la función que desempeñan las monedas de reserva.

El efecto que surte la disponibilidad de DEG en la evolución de las reservas de divisas se enfoca en este trabajo recordando, en primer lugar, la naturaleza de las fuerzas básicas que han dado como resultado el crecimiento de las reservas en divisas conforme al régimen del patrón de cambio oro. La expansión de las reservas en divisas obedece principalmente a que éstas devengan intereses y a su conveniencia en materia de transacciones. Pero, en vista de que las reservas en divisas constituyen pasivos para los países que son centros de reserva, el que esos pasivos aumenten por encima de cierta proporción o múltiplo de los activos en oro de dichos centros da pábulo a ciertas dudas acerca de si podrán convertirse más tarde al tipo existente, y, por lo tanto, ello constituye una amenaza para la estabilidad del sistema. En una economía mundial creciente, en que ha existido por espacio de mucho tiempo una tendencia alcista de los precios, las existencias de oro monetario en el mundo en su totalidad tenderán a reducirse con el correr del tiempo; esto impondrá límites a la expansión mundial tanto del total de las reservas en divisas como del monto de las reservas de todas clases si no se aumenta el precio del oro en términos de las divisas. Por lo tanto, el patrón de cambio oro en sí encierra la amenaza de que o bien las reservas lleguen a ser inadecuadas o de que se produzcan variaciones violentas en el valor del oro expresado en divisas.

La introducción del sistema de DEG, en la medida en que éstos se consideren como sustitutos aceptables del oro en los incrementos de las reservas, transforma la situación proporcionando, de hecho, una oferta potencial e infinitamente elástica de un sustituto del oro que tiene un precio fijo. El sistema de DEG elimina o reduce considerablemente una de las mayores incertidumbres de que adolecía el sistema monetario, a saber, la posibilidad de que haya un aumento pronunciado y general del precio del oro expresado en moneda; no obstante, por esta misma razón, podría coartar la anterior expansión desenfrenada del patrón de cambio oro y, en este sentido, eliminar uno de los “controles” del sistema.

La tendencia a que el sistema de DEG evolucione en tal sentido dependerá de varios factores. Aparte de las suposiciones generates a que se ha hecho antes referenda acerca del deseo por parte de Estados Unidos y de otros países de ajustar sus balanzas de pagos a fin de evitar o eliminar las acumulaciones excesivas de dólares en las tenencias oficiales, las condiciones necesarias para la evolución de los DEG como componente principal del crecimiento de las reservas pueden incluir una o más de las siguientes:

1) La aunación de los activos en divisas existentes en una sola cuenta internacional, así como arreglos correspondientes para evitar o neutralizar cualquier futura acumulación de saldos oficiales en moneda nacional; 2) como enfoque más limitado, un acuerdo oficial o extraoficial entre los principales países para limitar el crecimiento de otras reservas en divisas, y 3) un aumento de la tasa de interés que devenguen los DEG a fin de que ofrezcan más incentivos que las reservas en divisas.


Mr. Hirsch, Senior Advisor in the Research Department, is a graduate of the London School of Economics. He was formerly Financial Editor of The Economist and was on the editorial staff of The Banker. He is the author of Money International (London, 1967) and other writings on financial subjects.


The author wishes to acknowledge comments and criticisms on an earlier version of this paper from staff colleagues; from Bela Balassa, Charles P. Kindleberger, Raymond F. Mikesell, Peter Oppenheimer, Robert Triffin; and from participants in the J.B. Say Seminar at the University of Paris-Dauphine on December 7, 1970. He wishes to emphasize that responsibility for the present text is his alone.


Proposal by the Managing Director on Allocation of Special Drawing Rights for the First Basic Period (Washington, 1969).


This assumption is made for analytical convenience; the system would not be significantly changed by allowance for a small accrual of monetary gold that may result from present arrangements for official transactions in gold.


The numbers in square brackets refer to items in the References, which are on pages 247–49.


Cited on pages 248–49.


The change in global reserves, R, equals the change in dollar balances, D, plus newly accruing gold, G(R = D + G). The change in dollar balances will in turn be equal to the proportion held in dollars (α) of the reserve change of countries other than the United States (Ro), which will be equal to the amount of the U. S. payments deficit, Bus (representing the inverse of their surplus before allowing for new gold accrual) plus the net accrual of monetary gold, assuming this to be entirely outside the United States. Thus, Ro = Bus + G, and D = α Ro, or D = α(Bus + G). The change in reserves of the United States will equal the U. S. deficit with negative sign (− Bus) plus increases in dollar balances: Rus = D—Bus. Substitution gives Rus = α(Bus + G)—Bus: the rise in U.S. reserves will be the larger (or reduction in those reserves will be the smaller) the larger «, and the larger G, for any given payments balance Bus. [10] and [5]


Because the relative reduction in the price of gold against other prices will be limited by the extent of devaluations against gold, which will tend to limit the increase in gold absorption in countries that have substantially higher rates of inflation than the United States; and because the impact of reductions in the “real” dollar price of gold on gold production can be dwarfed for substantial periods by technological or “rent” influences in major producing countries. [8]


Without the assumption of a long-run tendency to rising prices, as Niehans has recently claimed [15], the gold exchange standard may be notably more viable than has been suggested in many analyses. However, Niehans’ contention that “There is just about one thing it cannot survive, namely, a long-run inflationary trend of world prices” must surely be seen as a structural weakness of the gold exchange standard, unless it is believed that the nature of the international reserve standard can, at an acceptable cost, force a permanent break in the long-run upward trend of prices that has persisted for almost four decades.


This is represented in the time series in Chart 2, which shows annual accruals of monetary gold, and annual increments in gold reserves outside the United States (derived from Chart 1, on the assumption that the path of reserve composition there shown also represents a time path). The change in U. S. gold reserves is shown as a residual.


An “independent” deterioration in the U. S. balance of payments while global gold accrual was still ample would also be expected to reduce other countries’ appetite for holding reserves in dollars, insofar as this involved expectations of possible depreciation against other currencies. But in these circumstances, involving a possible fundamental disequilibrium of the dollar but not of the gold-dollar system in Gilbert’s terminology [5], some compensating switch in foreign exchange holdings to other, stronger currencies would be likely. For there would then be no corresponding pressure for an increase in the price of gold in terms of all currencies.


The interest rate on SDRs is currently 1½ per cent on, effectively, holdings above net cumulative allocations, with a similar debtor rate charged on shortfalls of holdings below allocations. The Fund may increase or reduce this rate, but increases above 2 per cent or reductions below 1 per cent could not be made unless the rate for the remuneration payable on super gold tranche positions in the General Account is increased beyond 2 per cent or reduced below 1 per cent, as the case may be, which in turn would require a majority of 75 per cent of the total voting power. (Article XXVI, Section 3, and Article V, Section 9.)


The influence of the U.S. payments balance and its financing on total reserve growth, as indicated below, could theoretically be offset by compensating adjustments in SDR allocations and cancellations. This is discussed further in Section V.


Because the existence of the new reserve asset affects the characteristics of existing reserve assets in this way, Gresham’s Law is of more doubtful relevance than is sometimes suggested. In this case, the good new money improves the (formerly) bad.


A minimum for holdings of SDRs over time is currently required under the regulations on reconstitution, which are mentioned on pages 238–39.


Such an effect may be interpreted for 1968–70, beginning with the abandonment of gold pool operations in March 1968 and the declaration of the participating central banks that they no longer felt it necessary to buy gold from the market. This strategy, even as adjusted by the subsequent provision for modest additions to monetary gold holdings under the agreement of December 1969 between South Africa and the Fund, could never have gained credibility in the absence of an assured source of alternative liquidity growth on the lines of the SDR scheme. The central banks in their communique of March 17, 1968 drew specific attention to the SDR connection in the statement: “[Moreover] as the existing stock of monetary gold is sufficient in view of the prospective establishment of the facility for special drawing rights, they no longer feel it necessary to buy gold from the market.”


The role played by uncertainty in the gold exchange standard was succinctly pointed out by Marius Holtrop almost ten years ago. “Paradoxically, one might even say that the system’s stability is based upon uncertainty. For if it were inconceivable that the dollar price of gold might ever be raised, most of the reserve-holding countries might wish to hold only dollars and the gold exchange standard would tend to be transformed into a dollar standard, the United States having to buy almost all the world’s gold. It is very doubtful that this would be a desirable development. If, on the other hand, it were almost sure that the dollar price of gold would be raised, few monetary authorities would wish to continue to hold dollars, and the system would tend to revert to a gold reserve standard, for which no sufficient gold would be available. It is uncertainty, therefore, that presently controls the proper mixture of decisions.” (Statement by the Governor for the Netherlands, Mr. M.W. Holtrop [9], pp. 114–15.) Considerations of portfolio balance are likely, as noted earlier, to qualify these contrasting states, but Holtrop’s essential point remains valid.


Or between a high-yield asset and a lower-yield asset.


The influence exerted by a given interest rate differential may still be affected, on grounds of portfolio balance, by the relative size of accruals of the two assets. This is reflected in the diagram below. The horizontal scale shows the amount of SDRs willingly absorbed; the vertical scale shows the extent to which the interest rate on dollars is above the rate on SDRs. A given relationship is assumed between the stock of dollars and that of SDRs, and the curves I and II represent differences in relative accruals to the given stock. With accruals of SDRs three times as large as accruals of dollars (I), ON of SDRs would be willingly absorbed if the SDR interest rate were within OA of the dollar rate. With dollar accruals three times as large as SDR accruals (II), willing absorption of the same volume of SDRs, ON, is achieved at an interest differential on dollars of OB—i.e., at a lower SDR interest rate for an unchanged rate on dollars.



Citation: IMF Staff Papers 1971, 002; 10.5089/9781451947335.024.A001


Article XXV, Sections 2 and 3, of the Articles of Agreement.


Schedule G of the Articles.


For the United States, the reconstitution of SDR holdings that could be required under this provision would entail (in the same way as additional SDRs acquired through designation) an expansion of SDR holdings along with an expansion of dollar liabilities. To the extent that the countries acquiring the additional dollars in exchange for SDRs do not convert them into gold, the effect of these provisions will be to add to U. S. reserve assets; even if all the additional dollars are converted into gold, U. S. reserve assets will not fall. These provisions may therefore (depending, inter alia, on the preference for gold vis-à-vis SDRs) in certain circumstances have a relaxing effect on the U. S. reserve position.


This assumes either perfect foresight or—more realistically—adjustment of SDR allocation to actual reserve developments after a lag. In practice, increments of reserves of different types are likely to have somewhat different effects, with a generally weaker expansionary impulse for reserve creation that is associated with repayment obligations or an increase in liquid liabilities [Fleming, 3]. Increases in foreign exchange reserves resulting from an outflow of short-term capital from the reserve center will be associated with an increase in liquid liabilities in both the reserve center and the recipient country, and may therefore have an especially weak net expansionary impetus. For these reasons, the appropriate adjustment between SDR allocations and increments in other reserves may not be on the one-for-one basis assumed here for analytical convenience.


Allowance for proportionately small accruals of monetary gold and other foreign exchange would make the analysis somewhat more elaborate without involving any major change in the relationships considered.


This system could theoretically require SDR allocation to be negative in certain circumstances—viz., if the United States incurred a payments deficit that was large in relation to the target increase in reserves, and if the dollar proportion were also large. In such circumstances, the principles of the system could not be followed unless recourse were had to cancellations. It should be emphasized that the critical quantity concerned is not the U. S. payments balance as such but the dollar flow as additionally influenced by the dollar proportion. See also footnote 22.


Up to the point at which a U. S. surplus exceeded other countries’ proportionate share of SDR allocation applied to the target increase in reserves.


The dollar proportion will have no such effect where reserve changes of other countries are zero with SDR allocation, which will be true at the particular point at which a U. S. surplus is identical to the (three fourths) share of other countries in SDR allocation applied to the target reserve increase (Bus=SoSR).


Where the SDR allocation avoids a situation of global reserve tightness, (b) may be a condition of an improvement in the U. S. balance, although not a sufficient condition. This was judged to be so for the allocation of SDRs for the first basic period (Proposal by the Managing Director on Allocation of Special Drawing Rights for the First Basic Period, cited in footnote 2, pp. 9–10).


If this continued tendency were to lead to a persistently large flow of dollars into foreign official holdings, this might itself tend to reduce the dollar proportion on considerations of portfolio balance and of possible depreciation vis-à-vis other currencies. However, this factor, as pointed out above, would tend to increase demand for other foreign exchange assets.


See pages 223–24.


J. Marcus Fleming [4], pp. 43–44.

IMF Staff papers: Volume 18 No. 2
Author: International Monetary Fund. Research Dept.