To conclude the seminar, four panelists—Peter Kenen, Christian Stals, Alexander Swoboda, and Alejandro Végh—sought to pull together the main practical suggestions that had emerged from the two days of discussions. Following their remarks and a final round of general discussion, Stanley Fischer summed up the proceedings.

To conclude the seminar, four panelists—Peter Kenen, Christian Stals, Alexander Swoboda, and Alejandro Végh—sought to pull together the main practical suggestions that had emerged from the two days of discussions. Following their remarks and a final round of general discussion, Stanley Fischer summed up the proceedings.

Peter B. Kenen

My comments fall rather naturally into two groups: first, on the role of the SDR under the present Articles of Agreement of the Fund—the focus of the papers by Michael Mussa, Helmut Hesse, and others; and second, on the role of the SDR as the Fund’s own unit of account in any future reform of the Fund—the focus of the papers by Ariel Buira and Jacques Polak.

There was broad agreement at this seminar that the SDR cannot be expected to become the “principal reserve asset” of the international monetary system, and I heard no regrets on that score. The concerns and conditions that led to the creation of the SDR and the subsequent inclusion of that objective in the Articles of Agreement are not operative today.

I did not even hear concern expressed about the potential instability of a multiple reserve currency system or support for establishing a substitution account to consolidate official foreign exchange holdings. Two developments may account for the lack of concern about instability. First, the huge growth in the volume of foreign exchange trading allows central banks to engage in reserve currency diversification without much affecting reserve currency exchange rates. Second, the development of derivative instruments allows central banks to hedge against exchange rate risk without actually changing the composition of their reserves—a point made by Onno Ruding in a different context. The substitution account was once one of my favorite subjects, but I could find no excuse to trot out that hobbyhorse today. It has gone completely lame.

There was no similar consensus, however, about the usefulness of the SDR as a reserve asset or the strength of the case for issuing more SDRs under the terms and conditions set out in the Articles of Agreement. On the contrary, there was deep disagreement, with Mussa arguing that the notion of “long-term global need” can be interpreted in a manner that would justify a new SDR allocation and Hesse dissenting vigorously. I side with Mussa on this issue, although 1 have some questions about his analysis.

The core of the case against a new allocation is the concern raised by Hesse that it would allow certain governments to pursue irresponsible policies and might complicate monetary management by responsible governments and their central banks. To avoid this possibility, he argued, the IMF should engage exclusively in providing conditional credit.

Those who take this position should be required to explain why they do not carry their own argument to its logical limit. If governments must be prevented from pursuing irresponsible policies, why allow them to hold any reserve assets? Why not make them depend entirely on conditional reserve credit for balance of payments financing? Or should we perhaps infer that some governments can be trusted to pursue responsible policies, while others must be kept on a very short leash?

I am being a bit unfair to Hesse, whose argument is somewhat more subtle. He appears to be saying that market forces can be trusted to establish an equilibrium between the demand for reserves and the supply of reserves. Responsible governments are creditworthy and can therefore obtain reserves by market borrowing at interest rates not much higher than those they can earn on their reserve holdings, whereas those that must pay more to acquire reserves, by high-cost borrowing or by running current account surpluses, are undeserving of less costly access to reserves. There is thus no need for SDR creation to supplement other reserve assets, because the stock of reserves will always be socially optimal when it is regulated by market forces.

But this brings me back to my earlier, blunt question. If we can count on market forces to discipline governments, why do governments need reserves? Why should they not be made to borrow at market interest rates if and when they deem it appropriate to combine financing with adjustment or, in some cases, to opt entirely for financing? The answer is obvious. A government may require financing precisely when markets are unwilling to provide it, and markets are not always right—a view endorsed by many speakers at this seminar. In short, creditworthiness and worthiness are not identical.

I am nevertheless uneasy about Mussa’s analysis, which appears to be based on a somewhat complacent confidence in another sort of equilibrium. He seems to believe that the rather steady growth of global reserves over a long period implies some sort of long-run policy equilibrium. Hence, SDR creation would serve mainly to reduce the cost of meeting the demand for reserves; it would not raise the stock of reserves. Two questions bother me.

First, global reserves have behaved less erratically than one might have expected, given the shifts in exchange rate arrangements and in the structure of trade and payments that have occurred in the 25 years since the collapse of the Bretton Woods system. But can we safely speak of a stable, well-behaved demand for reserves, which would not be affected by changing the way that reserves are created? Individual governments have widely different reasons for intervening in foreign exchange markets and thus gaining or losing reserves. That these have “added up” to stable growth in the stock of reserves is intriguing but not entirely reassuring.

Second, if there is in fact a stable demand for reserves, SDR creation would presumably substitute growth in the stock of low-cost reserve assets for growth in the stock of high-cost reserve assets—although an outward shift in the supply curve could also be expected to raise total reserve holdings. But the ramifications of substitution need to be considered. If Germany and Japan were given extra SDRs, would they accumulate fewer dollars? If so, would that not tend to weaken the dollar? More generally, would SDR creation induce all countries holding currency reserves to accumulate fewer dollars, deutsche mark, and yen, causing all three reserve currencies to depreciate in terms of other currencies or raising U.S., German, and Japanese interest rates relative to other countries’ interest rates?

I turn now to the role of the SDR as the Fund’s own currency and the opportunities to which it might give rise.

In 1979, Jacques Polak published a fascinating paper called “Thoughts on an International Monetary Fund Based Fully on the SDR.” His paper for this seminar (chap. 7) rides that same hobbyhorse, and it is not lame. His argument is no less compelling today than it was in 1979, and I would assign his earlier paper, along with the one for this seminar, as required reading for every member of the Interim Committee.

Although Polak’s new paper presents his proposal as a way to deal with anomalies in the financial structure of the Fund, the proposal is much more ambitious. It would transform the Fund’s operations by having the Fund create SDRs whenever it extends conditional credit, rather than having to draw on a pool of national currencies. The pool of currencies would be liquidated by returning each country’s currency to that country. (If the Fund’s holdings of a country’s currency were smaller than the country’s quota, the Fund would make good the difference by issuing new SDRs to the country; in effect, countries having reserve asset claims on the General Resources Account would receive SDRs instead. If the Fund’s holdings of a country’s currency were larger than the country’s quota, the country would make good the difference by assuming an SDR-denominated obligation to the Fund.) Thereafter, the stock of SDRs would vary with the volume of Fund credit outstanding.

Polak’s paper does not discuss additional ways to create SDRs, but his proposal would not preclude SDR allocations of the type authorized by the present Articles of Agreement. His paper does discuss ways to make the SDR a more attractive reserve asset, to make sure that countries borrowing from the Fund and thus acquiring SDRs could find willing holders for the SDRs when they needed national currencies for foreign exchange intervention. But it might be prudent to preserve the acceptance and designation provisions in the present Articles of Agreement. Polak’s proposal is radical but not irresponsible. The Fund would create SDRs whenever it extended conditional credit, but it could not create unlimited quantities, nor could it engage in open market operations to regulate the stock of SDRs and, in particular, keep it from falling when the stock of Fund credit was shrinking. This was proposed by Robert Triffin, who wanted to convert the Fund into a full-fledged central bank. But that is not Polak’s objective. The Fund would become a bank but not a central bank.

Would quotas play any role under Polak’s proposal? They would continue to govern decision making in the Fund and access to Fund credit, and they could be revised periodically. But it would no longer be necessary to increase quotas in order to enlarge the Fund’s resources, and this change would banish two myths: (1) that all of the national currencies acquired and held by the Fund are equally useful to it, and (2) that decisions by national parliaments to increase Fund quotas are fiscal decisions, resembling, say, a replenishment of the International Development Association, and should therefore be subject to the budgetary process.

Finally, and most important, governments would have firm control over the Fund’s credit creating power. There would be an overall ceiling on the stock of Fund credit, which could not be altered without the approval of the Fund’s members, and the Executive Board would still have to approve the terms and size of every drawing.

There will nevertheless be concern on this score, which is why I have reservations about the proposals of Buira and of Marcello de Cecco and Francesco Giavazzi—under which the Fund would create SDRs to finance large short-term loans to countries facing acute liquidity problems—that the new emergency financing mechanism be converted into an emergency financing facility. The existence of such a facility could pose moral hazard problems. Governments might postpone adjustment; investors might be encouraged to believe that they will be “bailed out” in the event of a crisis. But my main concern here is different. The existence of such a facility could impair confidence in the effectiveness of the constraints on the volume of Fund lending and the size of the stock of SDRs.

Polak’s proposal would require a comprehensive revision of the Articles of Agreement and, therefore, a major effort by member governments to obtain legislative consent. The effort would be worthwhile, however, because it would obviate the need to obtain legislative approval for each and every increase in the Fund’s resources. Quotas would have to be revised periodically, and that would still require legislative approval, but it should be easier to obtain.

The Managing Director noted yesterday that this seminar has 32 participants, and that the same number attended the first meeting of the Bellagio group thirty years ago, when it wrote its report on reform of the monetary system. That earlier meeting, however, was convened by my Princeton predecessor, Fritz Machlup, in response to the assertion by one prominent official that academic economists could not contribute anything useful to the ongoing official discussion of international monetary reform. This meeting, by contrast, was convened by the Fund’s management at the behest of the Interim Committee. I hope that our response to this official invitation will match the quality of the response to the earlier official challenge.

Christian Stals

From the many excellent papers presented at this seminar, a few important facts became clear.

The International Financial System Changed Dramatically

First, the international financial environment of today is very different from the one that existed in the last half of the 1960s when the SDR was created. At that time, the fixed par value system, although very fragile, was still in operation. It was expected of members of the IMF to maintain fixed par values for their currencies as agreed with the IMF. Gold still played an important role as an anchor for the system—it was only in March 1968 that the two-tier system for gold was introduced to initiate the demise of gold in the international monetary system. The U.S. dollar was the dominant international currency used for reserve purposes and in all international transactions and was often referred to as the “fulcrum” of the system. The private financial markets mainly served domestic economies, whereas the multinational institutions, and particularly the IMF and the World Bank, provided liquidity and development finance for their members.

This situation has changed dramatically. Today, gold has been almost effectively eliminated from the international monetary system. The fixed par value system for exchange rates has been abandoned. The world is now on a floating exchange rate basis, although many countries intervene in foreign exchange markets to influence exchange rates. The U.S. dollar still plays a dominant role in the international monetary system, although other currencies, and particularly the deutsche mark and the Japanese yen, are increasingly used in the present multicurrency regime. The financial markets of the world exploded, became globally more integrated, and partly crowded out the multinational institutions from the role they had previously played, also in the creation of international reserves.

Objectives for Creation and Allocation of SDRs Changed

Second, the objectives for the allocation of SDRs today seem to be very different from what they were at the time of the inception of the special drawing right. In one of the papers (Helmut Hesse), the reasons for introducing the SDR were summarized as follows:

  • The main motive was the concern that a worldwide shortage of exchange reserves might occur as and when the U.S. overall balance of payments would go into surplus and that deflationary trends in the world economy could result from this.

  • At the same time, SDRs were supposed to supplement the limited physical gold reserves as a kind of artificial “monetary” gold, thus counteracting the further exchange of dollar reserves into gold and, ultimately, helping to defend the convertibility of the U.S. dollar into gold at the official price.

  • In the light of sharp price swings for gold and the U.S. dollar from the beginning of the 1970s, the objective was later added of replacing these previous cornerstones of the monetary system with the more stable SDR as the central point of reference—the “numeraire” of the system.

  • Finally, it was hoped that the SDR instrument would reduce the so-called asymmetries in reserve creation and distribution.

To meet these objectives, two important principles were established at that time: One was that SDRs would be created only to “meet the long-term global need, as and when it arises, to supplement existing reserve assets.” The other was that allocations of SDRs would be made on a global basis, distributed to all the members of the Fund according to a formula based on quotas. Today, these two matters of principle have become major obstacles in using the SDR for the contemporary needs of the current international monetary system.

Various legitimate and recognized needs were developed in the seminar discussions for a new round of SDR creation in the next basic period (beginning in 1997). They included the following:

  • A special issue must be made to the 38 members that joined the Fund in recent years and did not share in the previous two rounds of SDR allocations;

  • SDRs must be allocated to the group of countries that have little or no access to the international reserve-creating mechanism of private money and capital markets;

  • SDRs represent a low-cost reserve asset and should be issued to the poorer countries of the world to reduce their costs of holding reserves; and

  • The IMF should be given automatic access to a pool of SDRs to enable it to create a global safety net, or to become a lender of last resort, to provide assistance to members in distress.

All these needs, however, require selective issues of SDRs to certain countries only, and cannot easily be justified, or easily tested, against the requirement of a global need for reserve assets.

Despite all the efforts made by a number of participants, no convincing argument was, to my mind, provided to prove that the world at this stage really needs an addition to total international reserves.

(The argument that more SDRs should be created to maintain a fixed ratio between total reserve assets and the SDR is a dangerous one. This ratio may at times decline because of an excessive increase in the other forms of reserve asset, and that would certainly not be the right time to create more SDRs!)

Many Uncertainties Exist About What the SDR Is or Should Be

Another point that came out clearly from the discussions is that many uncertainties still exist on what the SDR really is. In his paper, Mussa stated: “SDRs functioned essentially as ‘unconditional lines of credit’ defined in terms of a basket of national currencies.”

Alec Chrystal made more or less the same point: “[SDRs] are simply the right to incur a liability, that is, to borrow at market interest rates. A right to borrow would not normally be classified as an asset.…” Others, for example, Jacques Polak, referred to the SDR as money: “I have found it convenient to treat the SDR as money.…”

It is important to take a firm decision on what the world wants to make of the SDR—is it going to remain just another form of a special credit facility, to be used mainly among the monetary authorities of the world, or must it eventually develop into global money (as the ECU or euro must eventually become on a regional basis for the members of the European Union)?

It is important at this stage to take this important decision, if possible. The rules for creating money or, alternatively, for providing a credit facility with restricted use will be very different. Once a decision on this basic issue has been taken, more consistent decisions can be taken on issues such as

  • the use of SDRs by private sector financial institutions;

  • the development of clearing arrangements;

  • the nature of the IMF (as a world central bank or as a conventional credit-granting multilateral institution); and

  • market-making operations in the SDR, for example, through open market operations.

Should the IMF Be Seen as an Emerging World Central Bank?

A number of proposals were made for extending the powers of the IMF to act more like a central bank, for example, to provide Lombard facilities and to become the principal issuer of fiat money; that is, SDRs issued should become a liability in the balance sheet of a restructured IMF.

Polak’s proposal for a restructured IMF balance sheet proposes that all existing SDRs should indeed be included as a liability of the Fund. Are they not supposed to be no more than a predetermined right for the holders to obtain credit from other country participants in the scheme, and is the role of the Fund not restricted to that of an intermediary that will facilitate the execution of the “drawing” rights, if the holders so desire?

A realistic warning, however, came from Helmut Hesse:

From a political perspective, it should also be said that many countries would probably not be willing at the moment to transfer such far-reaching powers to the Fund, and thus surrender a large part of their monetary policy autonomy. Any thoughts of making the IMF a kind of world central bank can therefore, under the present circumstances, only be described as “utopian.”

A similar statement was included in the paper of de Cecco and Giavazzi:

Fischer concluded that, at the time of his writing, the difficulties experienced with reaching an international agreement on the further creation of SDRs indicated that the world was not ready for a world money.…

There is little reason to change that conclusion. After 15 years, the world has hardly become more ready for a world money or a world exchange standard.

We cannot have an international currency without an international central bank. We cannot have an international central bank unless at least the major political powers will stand behind it.


In summary, it would seem as if the IMF recognizes the new legitimate needs of its members and sees for itself new challenges and functions to undertake. It is, however, shackled by the rigidities of its Articles of Agreement, and by an SDR system that was designed in a different environment, and for a different purpose. Implicitly or explicitly, many proposals were therefore made for a more pragmatic approach.

The problems of today’s international monetary system should rather be approached from a zero-based basis:

First, clearly define the role and the functions of the Fund in the present international monetary system;

Second, identify the shortcomings, deficiencies, and problems of the current system as it now operates;

Third, find new solutions and perhaps new instruments that will enable the Fund to play a useful role in solving these problems; and Fourth, restate and reconfirm the objectives of the Fund as a multilateral institution in the future international monetary system, taking account of the realities of the present and emerging situation (surveillance, conditional credit facilities, exchange rate guidance).

We may then perhaps find need for a new version of the SDR, possibly even under a new name, that will garner the support of all the members of the Fund. In the terms used by Mussa, the new SDR can be designed to be “modest” and “moderate,” not because of limited issues, but mainly because of its inherent characteristics. The approach may not serve to make the SDR the principal reserve asset of the world but can contribute to finding solutions to the more urgent problems that we have to face right now.

Alexander Swoboda

At the end of an intensive and rich two-day seminar, I have a feeling that there is not much to be added, especially after having heard the preceding panelists. Still, let me begin by taking up Wendy Dobson’s remark that the SDR is an answer in search of a question—and then add a few questions and answers of my own.

Why the SDR?

The question of why the SDR was created in the first place and whether that was a good idea was asked at the very beginning of this seminar—and often thereafter. Put another way, what problem or question was the SDR supposed to solve and did it do so well? We have been told that the SDR was designed as a way of meeting a perceived shortage of international reserves or liquidity, actual or foreseen, and in the bargain to alleviate the Triffin dilemma. True enough, though in the event the shortage turned out to be a glut, and the dilemma went away with the adoption of floating exchange rates. But, I think, there were other, or more fundamental, motivations for inventing the SDR. Thus, the SDR, in the same way as the bancor some twenty-five years earlier, was supposed to provide the beginnings of a world currency, in addition to being a substitute for the dollar and a complement to gold. It was to be an “outside,” collectively created money and not an “inside” (created within the system) money or reserve asset such as a national currency.

A principal and desirable characteristic of such an asset is that it would be much more symmetric in its effects than the accumulation of reserves in the form of national currencies, the dollar more specifically. By governing the overall supply of the asset, the international community could exercise a measure of collective action instead of letting international macroeconomic policy be entirely dominated by the policies of one country as a result of the way in which the dollar-gold system functioned in the late 1960s and early 1970s. Unlike gold, the supply of the asset could be governed by the macroeconomic needs of the world economy rather than by the vagaries of the production and private uses of the metal. Substituting SDRs for gold would yield real resource savings from the replacement of commodity by paper money and raise the problem of how to distribute the seigniorage. Substitution of the asset for dollars would also pose distribution problems. Distribution questions were answered by the decision to allocate SDRs in proportion to quotas and universally to the existing IMF membership. The rules governing the overall size of SDR allocations were laid down in Article XVIII.

It is not surprising that the SDR brought only a very partial answer to the problems it was supposed initially to solve, if only because there was no global government or authority to back this fledgling world currency and to ensure that it would become the world’s principal international reserve asset. The formulation of the provisions of the revised Articles of Agreement governing the rationale for and the size of SDR allocations did not help in this respect. And the move to floating rates profoundly modified the role of international reserve assets. As a result, to some extent, the SDR is a solution in search of a problem; in other ways, however, something like it will likely be needed some day as a collectively created reserve asset, as a means of reaping and distributing the gains from money creation at an international level, and as a means of increasing symmetry in the international monetary system.

Floating and the Need for International Reserves

The most often quoted sentence at this seminar has been the passage of Article XVIII of the Articles of Agreement that defines the criterion for SDR allocations: “… to meet the long-term global need, as and when it arises, to supplement existing reserve assets in such a 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.” That sentence does not offer very clear guidance on when and how many SDRs to allocate even under fixed exchange rates. The concept of need appears theological to the economist, at least in this context. In addition, there is no guarantee that an issue (cancellation) of SDRs would add (subtract) an equal, or at the limit any, amount to international reserves in a world in which the latter are held in the form of national currencies and hence are created inside the international monetary system. Moreover, “supplement” can be given at least two meanings that are not necessarily mutually exclusive: add to (or subtract from) international reserves, on the one hand, and provide an alternative instrument for holding such reserves, on the other. But, under fixed exchange rates, it is at least possible to establish a link between the quantity, or rate of growth, of international reserves and inflation or deflation in the world.

This link is far less obvious under floating rates, after 1973 if you wish. First, one of the distinguishing characteristics of floating exchange rates is to allow a much more significant difference in inflation rates than fixed rates. Inflation or deflation thus becomes a national rather than an international matter. Second, the functions of international reserves themselves have become almost exclusively national; individual countries hold reserves to use in the event of a run on their currencies, or as a confidence-building device to forestall such a run, or as a means to finance rather than adjust a current account disequilibrium when borrowing becomes too costly. There is also the question of what constitutes international reserves in such circumstances and what aggregate to use when measuring the total of international reserves in the world. In other words, the sum of total international reserves is no longer a world aggregate with clear systemic implications. Add to this the possibility of borrowing reserves under high capital mobility and the injunction to allocate in response to a clearly identifiable “need” and in order to influence a common world price level, or the common macroeconomic fate of one and the same world economy, and the allocation criterion as worded in Article XVIII is unlikely to serve as a clear and useful guide to decision making by the Fund.

Would it be bending the constitution to allocate SDRs in today’s circumstances under the existing Article XVIII? Perhaps, because it would be difficult to establish “need” unambiguously. But then it always was. If one was willing, as I would be but others I am sure would not, to interpret “supplement” as referring to providing an alternative useful asset as well as, or instead of, increasing the total quantity of international reserves, a need would be established if enough Fund members deemed that there was a “documented desirability of issuing more SDRs.”

What Role for the SDR Today?

What, then, are the arguments for issuing SDRs today or, if you prefer, for keeping the SDR alive? In this respect I found the arguments of Michael Mussa rather compelling. The basic idea is simply that there is a gain for the international community as a whole in providing access to an unconditional borrowing facility at lower cost than that that would be incurred in securing this facility individually. From that benefit one has to deduct the cost of increased inflation for the world as a whole, but that, I would agree with Mussa, is likely to be near zero for modest allocations. There would thus be a gain in efficiency. It has sometimes been said that such a gain would accrue entirely to the developing and transition economies whose access to borrowing is more limited than that of the richer or industrial countries. It has been said that SDR allocations as an option to borrow have no value for the latter countries, which can borrow anyway at market interest rates and thus, as potential creditors, only incur obligations and costs when allocations are made. This neglects the fact that such an option is valuable even for those that currently enjoy low borrowing costs in an uncertain world in which their currencies may be attacked in the future. The only country for which such an option would be entirely without value would be one in whose currency (and interest rate) the option is not entirely denominated. As long as the SDR is a basket of currencies, no country is entirely in that position. In other words, there is some value in having the SDR even for rich countries, since it reduces the expected cost of illiquidity for them as well as for the poorer, less privileged countries. The fact that any country holds owned international reserves at all shows that it is not sure to be able to borrow large amounts of reserves at the lowest market interest rate in any possible future circumstance.

A second argument for keeping the SDR alive is based on equity considerations. Equality of access to borrowing facilities at the same price at least over a modest range is appealing on solidarity grounds. But equity and universality of access would also argue for allocating SDRs retroactively to new members. To the extent necessary, the Articles of Agreement should be amended to make possible such retroactive distributions of SDRs in proportion to quotas.

Third, it is good to have a unit of account that is not a given national currency. This is not an overwhelming argument for the SDR, but the SDR does serve a useful function in that respect. It is also useful to preserve the only embryo we have of an outside world, rather than a national, or regional, currency.

These arguments do not stop me from agreeing with Eichengreen and Frankel that in their social science fiction future there is unlikely to be a dominant role for the SDR. But it may be the beginning of the basis for a more symmetric international system whose reserve currency is likely to be an entirely different animal from the SDR. Nor do these arguments justify stretching the SDR too far, for instance, by designing an emergency financial facility that would promote use of the SDR.

Finally, many of our discussions have been couched in terms of efficiency and economic analysis. Fundamentally though, we are also dealing with political and equity issues and decisions. This brings me back to trying to provide, in conclusion, my answers to the questions posed at the outset of the seminar by our hosts:

  • Yes, one should allocate on both efficiency and equity grounds; but no, one should not allocate too much. In order to encourage the holding of SDRs and avoid their being used as a means of long-term borrowing, there may be a case for linking the rate of charge to the length of use of SDRs.

  • Yes, the criterion for distribution should be a proportion of quotas.

  • Yes, there should be a retroactive allocation to existing and prospective new members.

  • Yes, measures that would improve the usefulness of the SDR, in particular in private transactions, such as the issue of SDR-denominated instruments by international organizations or by governments, would be welcome.

Alejandro Végh

I will concentrate my remarks on three points: first, some comments on Mussa’s and Hesse’s papers as representing rather extreme positions or very different positions on the very controversial subject that we dealt with yesterday and today. Second, I will refer to Polak’s proposition and some related papers, and here I will deal a little bit with the question of conditionality and surveillance, which is really my main preoccupation. And I will also make a reference to the question of a safety net, on which I have some disagreement with the views of other participants. Finally, I will refer to the matter of timing for some basic reform if and when it is seen necessary or politically feasible. This is almost science fiction, of course.

Let me refer first to Mussa’s paper because it is the most expressive and sophisticated version of the Fund management position. Let me first make some minor methodological observations that do not, in my opinion, invalidate Mussa’s thesis, but I think they are pertinent.

I was rather surprised by the finding, the econometric finding or the statistical finding, of Mussa regarding the stability of the ratio between reserves and imports. And this matter is mentioned in a few other papers. I consider this a rather old-fashioned variable, and I would suggest that perhaps it is more of an accidental than a causal relationship. Today, with the kind of exchange rate systems in the world and the differences between pure float and managed float and all kinds of floats and fixed exchange rates and currency boards and substitution of domestic currency, it is very difficult to explain this kind of relationship. So I will not consider it too important.

On the matter of substitution of domestic currency, let me say that I consider this a very important and beneficial phenomenon of modern economic development, and, of course, it tends to reduce the demand for reserves. I think if it is extended into the future, it will make for a different world. It is a monetary reform performed and decided by the public, and government intervention is usually just to facilitate it—which is a good thing in itself—by some legal changes.

One of the many benefits of substitution—partial or total substitution (in my own country it was 91 or 92 percent) of bank deposits into dollars—is to reduce the incentive for printing money, because the benefit is so small on such a small monetary base in domestic currency. In other words, the benefit is to reduce the need for reserves, for international reserves for that significant proportion of foreign currency that is circulating, because the reserves are held by the Federal Reserve and the Bundesbank.

Incidentally, in the Eichengreen–Frankel paper, there are some figures on the amount of dollars held abroad: 60 percent of the dollars issued by the Federal Reserve are held outside the United States and 30-40 percent of deutsche mark are held outside Germany. This situation, of course, has benefits and dangers for the reserve currency countries. It is not entirely irrational that Germany and Japan, for instance, have been very reluctant to allow such use until recently, and some of the dangers were mentioned by Kathryn Dominguez yesterday.

Another matter I would like to refer to with respect to the projection made by Mussa on the demand for reserves in the future is the very high cost of reserve accumulation, which one of the participants mentioned yesterday. In my experience in Latin America especially, I find that governments sometimes go too far and do not make appropriate use of resources in building up international reserves. Take the case of Brazil, for instance. This issue is not only a matter of relating the rate of return on reserve assets to the cost of borrowing in international markets, but it is also a matter of the cost of sterilization, which is approximately the rate of interest paid on internal debt. For example, in Brazil at this moment, the cost of sterilization is about five times the rate of return on assets.

Still, these minor observations, this uncertainty in the projection of demand for reserves, do not invalidate Mussa’s argument, which I would summarize in a well-known American phrase: small is beautiful.

I must confess that I felt rather uncomfortable after my first reading of Mussa’s paper, perhaps because of my professional distortion as an engineer and mathematician. I asked myself, well, if small is beautiful, why not go to the limit and not do anything, not issue any SDRs at all? But I think the beauty of the argument and the main point, I would say, of Mussa’s paper and of the position of the Fund is that this is precisely what one should do in order to keep the instrument alive for future purposes, if and when. And I think that is correct, and as Mussa argues, and argues well, the fears of such an allocation on a modest basis are greatly exaggerated.

I read with great pleasure Helmut Hesse’s paper and other papers in the same vein. It is a very well-reasoned paper, but I find that the final argument is at fault, because if one carries his very rigorous, logical argument to its conclusion, I think the paper should recommend the abolition of the SDR and the elimination of the provision to make it at some moment the principal asset of the international monetary system. But I gathered the impression from reading this paper that this is not Hesse’s wish or desire, since he quotes approvingly Professor Henry Wallich’s sentence, “Let’s put it on the shelf.” And I think that if one wants to put the SDR on the shelf for future and basic important objectives, then Mussa’s proposition is reasonable, because otherwise immobility—present immobility and the immobility of the last fifteen years—will lead to paralysis and extinction of the instrument.

My second point is just to express my agreement with Polak’s proposal and some of the points raised in the papers by Ariel Buira and by Muhammad Yaqub and his coauthors. Although I quite agree with Governor Stals that Polak’s proposal is unrealistic and Utopian for the time being, it may not be in the future. I support Peter Kenen’s position on this. I enjoy the symmetry, the elegance, and the simplicity in Polak’s proposal. But I would add that I would associate this proposal with a very high preference for conditional lending and for keeping international credit to a minimum. And I was glad to hear from Buira and from Azizali Mohammed that this, I think, would be their position as a practical matter also when their proposals are considered.

In my experience as a former finance minister, on the IMF Board, and as an observer like most of you, I am absolutely confident about the proposition that when you provide a country’s government with conditional liquidity, the benefit accruing to the country derives to a much larger degree from the conditionality than from the liquidity. Indeed, often too much liquidity can help to delay adjustment and be detrimental to the recipient’s long-run interest.

In connection with this matter, I would also like to add a word of caution on the safety net and the role of the IMF as a lender of last resort. I have been worried about some of the propositions I heard yesterday and today, and I must say that I was not entirely satisfied with Buira’s answer to Peter Kenen’s question on this matter yesterday afternoon.

As an observer of all these matters, I am increasingly worried about the cost of banking crises and bailouts at the national level: the savings and loan problem in the United States, the bailout of the Continental Illinois Bank and of Crédit Lyonnais in France, and similar crises in Brazil, Argentina, Japan, and so forth. In my own country, Uruguay, we have spent about $2 billion, in a country in which the GDP is $15 billion, in bailing out some banks. And I think it is very dangerous to extrapolate to the international level. So I would ask some questions, very hard questions on this point.

Can we afford many Mexicos? If we could, even if we could, is it good for the general welfare of the world economy? I really doubt it, and I think sometimes it is better for future discipline not to avoid default. A dimension of unlimited liquidity provided by the decision of the Executive Board or even by the Managing Director is really too much for my thinking.

Finally, let me say a few words about when I consider that, if we keep the SDR alive, we can expect the opportunity for revisiting the issue, along the lines of the Polak proposal or some other fundamental reform. I would expect—and this is, of course, highly subjective and controversial, that the occasion could arise after European monetary unification and the final incorporation of Hong Kong into China.

Yesterday, the Managing Director referred to the Asian currency situation, and I remember listening in 1992 to the Chinese Executive Director at the time speak warmly of the Hong Kong experience with the currency board based upon the U.S. dollar. Of course, we must recognize that the present situation regarding the international monetary system is rather satisfactory. There is no urgency for change. But if we can improve it in a few years, why lose the chance on an instrument that could improve performance and equity and remove some market imperfections?

General Discussion

Yusuke Onitsuka argued in favor of offering low-cost reserves to some of the countries that had difficulty gaining access to capital markets, but not necessarily through the Fund. Domestically, small and medium-sized corporations typically faced higher interest rates, but the central bank did not take care of that issue. In Japan, the Ministry of International Trade and Industry (MITI) offered low-interest-rate loans to these small industries. Unless the equity issue could be separated from the SDR issue, the SDR or even the role of the IMF itself would be jeopardized. If some developing countries needed a net transfer of resources, the World Bank or bilateral aid should be the provider.

John Williamson observed that it was the Bank of Japan that acquired seigniorage and then redistributed it to the Government, which used the money through MITI. No international mechanisms existed to perform that function, and some thought therefore that it would be rather natural to have it performed directly through the Fund. Williamson did not believe, however, that that was about to happen.

Commenting on Kenen’s point that the logical implication of Mussa’s analysis was that the reserve currency countries would either have to pay higher interest rates or face depreciation of their currencies, Williamson agreed that those were the mechanisms by which a redistribution of seigniorage occurred. When countries did not rise to Swo-boda’s challenge—in terms of solidarity and sharing the seigniorage—they allowed that issue to dominate their actions and there was a stalemate over the question of allocation.

Williamson queried Stals on his comment that the Polak proposals would turn the IMF into a world central bank. His own understanding was that these proposals would tidy up the bookkeeping of the Fund and thereby enable it to carry out its functions more efficiently. Williamson agreed with Kenen that that was one of the minimal reforms that ought to be pursued at this stage.

Jacques Polak added that Panel B of his Table 1 was a table of a bank, not of a world central bank. Just as Citibank created Citibank money when it issued a loan, it was very reasonable to assume that the Fund created Fund money, which happened to be expressed in SDRs, when it made a loan.

Christian Stals responded by asking whether the IMF was just a financial intermediary that borrowed from its members in the form of quotas or SDR rights and lent it on to other members, or whether it was a bank or a central bank that could make loans by increasing its liabilities. Creating the possibility for the IMF to make loans by issuing SDRs to borrowers that could then be used in market transactions, without the IMF taking away money from anybody else as a financial intermediary, would change the concept of the IMF.

Polak observed that there had been much debate at this seminar on whether Mussa was right in saying that the Articles of Agreement permitted across-the-board allocations in present conditions. Many people had agreed, and a substantial number had not. He did not feel that was the primary issue for this seminar. A much more important question was whether a general allocation was a good idea. Perhaps the most interesting result of this seminar was the paper by Yaqub, Mohammed, and Zaidi, which implied that if—as Mussa had argued—allocations were useful because they saved money for a particular set of countries, then it would be better not to go beyond those countries. By creating far fewer but more targeted SDRs, one would run much less risk of overloading the SDR circuit and therefore make it much more possible to run the SDR system without resorting to designation and without having to worry so much about acceptance limits.

Felix Kani supported that point by way of illustration. His own country, Zambia, had recently embarked on a program under the enhanced structural adjustment facility (ESAF). The conditionality for this program was very tight and included both a substantial buildup of reserves and the servicing of outstanding external debt. If both benchmarks could not be achieved, the arrangement might be suspended. Kani saw an SDR allocation as a means of obtaining at least temporary relief for countries that had demonstrated that they could undertake to implement austere adjustment measures.

Peter Kenen responded to the point made by Stals on the different private roles of the ECU and the SDR, or what Kenen liked to call the “private life” of the SDR. At the time of the previous Fund conference in 1983, Kenen recalled that the two currency units were both doing quite well on the private market. Since then, the ECU had obtained much more official sponsorship and support, including large-scale borrowing by public entities, whereas the SDR had not. That was a key issue. It was not the expectation of economic and monetary union that had given the ECU its start in the market, because the ECU’s initial success had come much earlier. The SDR, of course, had gone the other way and had almost vanished from the market.

Referring to what had been called by seminar participants “the endangered species” argument for the SDR and for an allocation, Kenen submitted that the SDR issue was no less deserving than the spotted owl and that the real resource cost of preserving the SDR was far smaller than the amount to be devoted to the spotted owl.

Concluding Personal Remarks Stanley Fischer

Let me make just a few concluding comments. We have had a remarkably comprehensive and high-quality set of papers and comments, starting with the history of the SDR and then moving on to the current situation and to the far distant future. Some of the questions that have been posed relate to the role of the SDR under the current Articles. Those are the questions of whether an allocation would be justified under the current Articles and whether it would be wise; whether the use of the SDR could be promoted within the Fund or on the outside, within official institutions on the outside or in private markets; and whether within the current Articles there are ways of getting a more equitable distribution of SDRs, as described in some of the papers. And then we have had a second set of discussions relating to the potential role of the SDR if the Articles were to be amended; for instance, through allocations to new members as mentioned by Alec Chrystal, the possible safety net feature, the international clearing mechanism proposed in Robert Heller’s paper, and the Polak proposals.

Within the current Articles, there has been little discussion of the wisdom of creating more SDRs under the current terms; rather, most of the discussion has dealt with whether the Articles permit an allocation. A relevant point made in this regard is that there very clearly remains a demand for reserves. We started with a paper by Robert Solomon, which said that part of the original impetus for the SDR came from the French observation that the United States enjoyed an “exorbitant privilege.” Today, several countries enjoy the same exorbitant privilege of creating reserves.

The equity of the reserve allocation mechanism is an issue that has come forward with increasing emphasis. But that point is not clearly one that justifies the finding of a global need. The most coherent argument for making an allocation on the basis of global need, that is, within the current Articles, is the point made by Michael Mussa’s paper, which says that there are a sufficiently large number of countries in the system for whom the costs of acquiring reserves are so high that there is a need, in the sense in which he interprets it, for an allocation. But we also had arguments that said that within other interpretations of global need, there is no basis for an allocation. I am not sure that the argument got much farther toward a resolution than it had been before, but it was certainly sharpened and clarified a great deal.

Several discussants raised the question of whether the SDR is likely to become the principal reserve asset, which the current Articles require the Fund and its members to work toward. There was not, however, much belief that that was likely to happen in the foreseeable future.

The question was also discussed whether under the current Articles it would be possible to get a higher proportion of SDRs to countries that would benefit from having them than would result under a global allocation. This point was taken up in particular in the proposal of Deputy Governor Zhu for what is called a postallocation redistribution of SDRs, in which those countries that do not need the allocation make them available for redistribution to other countries through the Fund.

Within the current Articles, we could also envisage a number of actions that would enhance the demand for SDRs both within the public sector and perhaps in the private sector, for instance, by encouraging the World Bank to use the SDR, as proposed by Onno Ruding. Of course, different parts of governments are represented in the Executive Boards of the Fund and the Bank and, as Ruding noted, they do not always think in the same way. We also heard a proposal to encourage the use of the SDR in private markets. The precise importance of that for the issues that now confront the Executive Board and the Interim Committee with regard to the possibility of a global allocation is not quite clear, except in the general sense that more people would get to know what the SDR is.

We got tied up somewhat—but not very much—on the issue of whether the SDR is credit or money. In the current system, it is unambiguously credit. And then the question was raised whether it was wise to create conditional or unconditional credit if the Fund was going to create credit. With that discussion, we began to move over to the second set of issues, relating to the possibilities for the SDR if the Articles were amended.

A set of very interesting alternative allocation mechanisms and alternative questions emerge once one goes beyond the current Articles. The underlying point that has been made repeatedly is that if an intellectual case beyond that now in the current Articles is to be made for allocations of SDRs, it will have to be one that relates to the distribution of the gains from seigniorage, the seigniorage being not the creation of currency but the creation of reserves in the current system. And the proposal by Yaqub, Mohammed, and Zaidi for allocation to countries that need it is very close in many ways to the proposals that were made by the Group of Seven industrial countries at the 1994 Interim Committee meeting in Madrid, in terms of how they would like to see the distribution made. But it is very unclear how one would amend the Articles of Agreement to deal with this issue of the appropriate allocation of the gains from seigniorage without doing it on a basis proportional to quotas or some other measure in the Fund. As this proposal is considered, we would have to look much more closely at the basis on which it would be done and the way in which the amendment would be written.

We had also the proposal by Jacques Polak for making the Fund entirely an SDR-based organization, which is extremely interesting. As somebody who still has not figured out what the General Resources Account and the SDR Department and so forth are about, I hope it works—so that I never have to figure out exactly what they are. I suspect that once discussions began on who would gain and who would lose from such a system, it could, unfortunately, be a bit harder to get it moving than it seems on paper. But it is certainly something that should be looked at.

We also had a discussion of a possible role for the SDR as a basis for emergency lending. There is a subtext there, which was noted by several people, of who should be entrusted with making decisions in an emergency, and of the need for accountability.

As the seminar proceeded, and as the papers and the panelists looked further ahead, we increasingly discussed what was called social science fiction. The point about science fiction that we all know, particularly those of us who are beginning to have gray hair, is that what is fiction one day may happen the next. Many of the science fiction stories about communications or computation facilities of only a few years ago are hard facts today. We therefore have to think very hard about the science fiction stories about the SDR and the international monetary system that were told in this seminar.

By focusing on the SDR, we did not, as Wendy Dobson and Onno Ruding mentioned, focus so much on the overall question of the role of the Fund and on which means of financing of the Fund is preferable. Those issues will have to be discussed, and so will the issues brought up by the panelists in this final session, namely, what is the ultimate role of the Fund and of the SDR? The Articles of Agreement tell us that the goal of the IMF is to promote international monetary cooperation by means of a permanent institution that concerns itself with international monetary problems. The international monetary system is operating moderately well now, but it does have a habit of getting into trouble every few years. The Fund has an obligation to keep rethinking the architecture and operation of the international monetary system, and the role of the SDR in it. We shall do that.

Let me thank all of you and declare the conference closed.

  • de Vries, Margaret Garritsen, The International Monetary Fund 1966–71: The System Under Stress (Washington: International Monetary Fund, 1976).

    • Search Google Scholar
    • Export Citation
  • de Vries, Margaret Garritsen, The International Monetary Fund 1972–78: Cooperation on Trial (Washington: International Monetary Fund, 1985).

    • Search Google Scholar
    • Export Citation
  • de Vries, Margaret Garritsen, The IMF in a Changing World 1945–85 (Washington: International Monetary Fund, 1986).

  • Horsefield, Keith, The International Monetary Fund 1945–65: Twenty Years of International Monetary Cooperation (Washington: International Monetary Fund, 1969).

    • Search Google Scholar
    • Export Citation