The goal of this session was to examine how the composition of official reserve holdings affects systemic stability in today’s world of multiple reserve assets, and whether substitution of SDRs for currencies in official reserves might contribute to stability. The principal paper was by Christian de Boissieu, followed by discussion of the issues by Paul De Grauwe, Kathryn Dominguez, and Makoto Utsumi.

The goal of this session was to examine how the composition of official reserve holdings affects systemic stability in today’s world of multiple reserve assets, and whether substitution of SDRs for currencies in official reserves might contribute to stability. The principal paper was by Christian de Boissieu, followed by discussion of the issues by Paul De Grauwe, Kathryn Dominguez, and Makoto Utsumi.

The SDR in the Light of the Multicurrency System

Christian de Boissieu

Since the creation of the SDR, the international monetary system has experienced dramatic changes. These changes are well known and do not need to be analyzed in this paper. The purpose of this paper is to analyze the rather disappointing record of the SDR in the light of the evolving international monetary system. The development of a multicurrency system at the world level has seen a major change since the 1970s, even if it originated during the 1960s and played some role in the collapse of the Bretton Woods system during 1971-73. Three aspects will be covered: (1) the evidence and the arguments relating to the role of the SDR; (2) the assessment of this role in the light of currency competition taking place within the multicurrency system; and (3) some expectations about the role of the SDR.

Evidence and Arguments

Official SDR Operations Are Reaching a Threshold

The SDR was introduced in a period of stable exchange rates (the Bretton Woods system), before the dramatic acceleration in financial innovation and at a time of expected liquidity shortage. It was seen, at the end of the 1960s, as the means to provide the additional liquidity required for a sustained growth in world trade; reduce the dependency of the international monetary system on production and distribution of gold; and alleviate the dependency of that system on the dollar and reduce (not remove) some structural asymmetries inherent in the functioning of the gold exchange standard (see the convergences, from a very different theoretical background, between Robert Triffin and Jacques Rueff).

More than 35 years after the introduction of the SDR, its record is rather disappointing. Money is defined by its functions. To use the terminology proposed by Hicks (1967), the SDR is a highly “partial” money. Only 3 countries (including Libya and Seychelles) presently peg their exchange rate to the SDR (15 did in 1980). It acts as a unit of account (the numeraire) for IMF operations and as a medium of exchange for very specific operations (resource transfers between participating member countries of the IMF, thoroughly analyzed by Coats, Fursten-berg, and Isard (1990)). The Fund’s transactions are denominated in SDRs, even though actual payments and repayments can be made in a variety of currencies as well as in SDRs. As a store of value, its role is also marginal. In April 1995, SDR holdings amounted to 2.4 percent of overall official reserves (excluding gold). For the developing countries, the figure was still lower (1.1 percent). The arithmetic behind these figures is uncompromising: given the fact that the second and last allocation of SDRs ended in 1981, leading to overall outstanding holdings of SDR 21.4 billion, the proportion of the SDR in total nongold reserves could only decrease (from 5.4 percent for all countries and 2.6 percent for developing countries in 1982). Contrary to the goal of the Second Amendment of the Articles of Agreement, the SDR has not become “the principal reserve asset in the international monetary system.”

When looking at the official circuit and role of the SDR, the stock dimension has to be complemented with a flow approach, namely, some information about the rate of turnover of the stock and the degree of activation of SDR holdings. The official circuit of the SDR includes three categories of participants: (1) the Fund through the General Resources Account; (2) the participants in the SDR scheme (in effect, all IMF member countries); and (3) some other official entities designated by the Fund. Even if it does not correspond to a really significant velocity concept, for a steady global amount of SDRs, annual transactions in SDRs appear to have increased dramatically in the 1980s. This is particularly true for the “transactions by agreement” (whereby participants exchange SDRs for other currencies), compared with “transactions by designation” (whereby the buyers of SDRs are designated by the Fund because of their strong reserve and balance of payments positions). With the increased flexibility in the holding and use of SDRs (removal of the reconstitution requirement in 1981, etc.), the possible gap between actual and desired holdings must have been reduced substantially. To document this statement, the demand for SDRs by participants in the scheme in the light of a portfolio demand for money would have to be estimated. The demand for SDRs by participants is a function of a scale variable, the interest rate on the basket of currencies that constitute the SDR and interest rates on other reserves, exchange rate risk considerations, etc.

Private SDR Operations Remain Negligible

The use of the SDR outside its official circuit has remained quite limited until now. On this matter, information is scarce and piecemeal. Moreover, when a financial operation is denominated in SDRs, the settlement and the payment are very often in a variety of other currencies.

The private SDR market developed quite impressively at the end of the 1970s and in 1980–81. Since 1982 it has been on a downward trend. For example, according to Bundesbank sources, there are no SDR-denominated deposits (either demand or time deposits) with German commercial banks. The same applies to France. The SDR-denominated liabilities of U.K. and Belgian banks have dropped dramatically during the 1980s (see Table 1).

Table 1.

SDR-Denominated Liabilities of Banks

(In millions of SDRs)

article image
Sources: National Bank of Belgium and Bank of England.

Piecemeal information on bonds and loans denominated in SDRs suggests that their absolute amount was much smaller than bank deposits in SDRs (International Monetary Fund, 1987). For instance, no SDR-denominated syndicated loans have been granted since 1982.

Both the SDR and the ECU have in common that they are defined as a basket of currencies. But they differ in many other respects. Whereas the private SDR has not emerged, the private ECU is a reality (Table 2).

Table 2.

Market Shares of the ECU

(In percent)

article image
Sources: Organization for Economic Cooperation and Development; International Monetary Fund.

As regards the ECU, a dialectic between its two uses—the official and the private—has taken place, which confirms that the two aspects cannot and must not be decoupled. Even for the ECU, however, the integration of the two compartments (the official and the private) remains imperfect.

Clearly, it is too easy and somewhat excessive to compare the relative success of the private ECU to the failure of the private SDR. The experience with the private ECU leads to an interesting conclusion. A significant drop in the market shares of the private ECU occurred after the two crises in the European Monetary System in 1992–93 and the consequent loss of credibility. The recovery has been slow since then. Market operators were negatively impressed over a certain period by the lack of both economic convergence in Europe and political credibility of European integration. Therefore, the premium, that is, the gap between the theoretical value of the ECU basket and the actual value of the ECU, has increased dramatically and is still quite high.

Reasons for the SDR’s Poor Record

When the SDR was introduced, it was the answer to a collective demand for greater flexibility (or elasticity) in the international monetary system. When I refer to the concept of elasticity, I mean the capacity of the financial sector (financial intermediaries, capital markets, monetary authorities, etc.) to satisfy in an adequate manner and without much lag and distortion the financial needs of the real sector (investment, growth, etc.). The great theoretician of monetary and financial elasticity was Wicksell (1898), with his analysis of the transition from a cash to an organized credit economy.

In this role of giving more flexibility and elasticity to the system, the SDR soon after its creation was circumvented by the conjunction of several structural changes: the suspension of gold convertibility of the U.S. dollar, the transition to floating exchange rates, and the transition from a situation of actual or perceived dollar shortage to a configuration of effective or potential dollar overhang owing to structural imbalances in the U.S. balance of payments. Is there a correlation between the poor record of the SDR in quantitative terms and the fact that the necessary elasticity was introduced by other means? If we look back at the period 1968–72, it is evident that the creation and first allocation of SDRs was a necessary but insufficient answer to the manifold challenges that the international monetary system was facing. As Rhomberg (1991, p. 155) rightly summarizes, “When the first allocation period ended in 1972, it was clear that there was no need for further SDR allocations to alleviate a threatening shortage of reserves.” Therefore, very quickly, it appeared that the SDR would play a marginal role in a better adjustment between the supply of and the demand for international liquidity. The transition to floating exchange rates reduced, ceteris paribus, the demand for official reserves. The supply of new liquidity was satisfied mainly through the dollar overhang and the privatization of the international monetary system (recycling the petrodollars in the 1970s, development of Euro-markets, and access to international credit and capital markets for many sovereign borrowers). An “official” currency such as the SDR had to stay apart from and could not benefit from the privatization of the international monetary system. It was exposed to all its implications. In addition, political pressure aggravated by the 85 percent voting rule has blocked, with ad hoc and changing coalitions (among some Group of Seven countries, between some developed and developing countries, and among developing countries), a further allocation of SDRs since the second one (1979–81).

Figure 1 represents the evolution of the reserve ratio during 1970–95. Over the period the ratio has fluctuated between 0.3 and 0.55 and had about the same value in 1970 and 1995 but in a totally different institutional configuration. But this ratio and its inverse—a velocity concept whatever the theoretical and empirical difficulties of measuring the velocity of money at the international level—cannot be decoupled from what they do not take into consideration, namely, the nonofficial sources of international liquidity.

Figure 1.
Figure 1.

Ratio of World Reserves to World Imports, 1970–941

(In percent)

Source: IMF, International Financial Statistics.1 Total reserves of all countries, including gold holdings at market prices, divided by world imports.

The reference to the privatization of the international monetary system leads to another consideration. The SDR was not only exposed to the implications of political decisions and institutional changes; it was also a victim of the acceleration in financial innovation that gave the required elasticity to the system. In the Anglo-Saxon countries (the United States, Canada, and the United Kingdom), financial innovation surged during 1972–74, i.e., just after the end of the first SDR allocation, for reasons that are well known (rise in inflation rates and nominal interest rates under the Regulation-Q system in the United States). The SDR, which was a financial innovation when it was created, became quickly outmoded by more sophisticated financial instruments such as derivatives markets. When looking at the competitiveness of the SDR (that is, its ability to face currency competition), it is important to keep in mind that the conditions of this competitiveness have changed dramatically with the acceleration and propagation of financial innovation. The international monetary authorities have been trying to adjust to the new competitive conditions (regarding financial instruments) by increasing the attractiveness of the SDR: new definition of the basket (since 1981), regular changes in the respective weights given to the currencies included in the basket, and calculation of the interest rates on SDRs. But it must be acknowledged that, as regards competitive conditions, there is no true “level playing field” between the SDR and some of the other monetary and financial instruments.

The reasons for the virtual absence of the private SDR are manifold and well known. I do not want to make an exhaustive study here but would rather emphasize several aspects of the problem. Some are more technical; others are more political.

The Attitude of the Monetary Authorities. The private use of a currency is not limited to its use by private agents such as firms and households. It can also refer to some transactions by central banks and some other official bodies (for example, the private ECU transactions initiated by some European central banks). To be more specific, it is clear that most of the participants in the SDR scheme, with some diversity of judgment among them, were not so keen to promote the private SDR.

The Attitude of Market Operators. Markets do not like to initiate and trade SDR-denominated operations for at least two reasons. First, it is perceived, rightly or wrongly, as a rather complex instrument (though it is much less complex than most derivatives instruments such as futures and options, it does not offer the same range of services that they offer). Second, market operators prefer to deal with currencies that are backed by identifiable economic policies and policymakers.

The Existence of Superior Financial Instruments. In the framework of a very simple risk-return model, it could be shown that from the point of view of portfolio management and also microeconomic policy vis-à-vis precautionary cash balances, the SDR is most likely to be dominated by some other financial instruments. Let us take two examples:

  • In many cases, the current basket of currencies that defines the SDR represents a second best, a third best, or even an nth best. This could also be true for the ECU and it has to do with the possible trade-off between tailored and standardized financial instruments. For example, many multinational firms keen to better manage their exchange rate and/or interest rate risk could be led to resort to an optimal basket different from the effective SDR basket (or the ECU basket) in terms of the currencies being considered and/or the respective weights attached to them. The best “ad hoc” baskets from an individual viewpoint do not coincide in general with the official basket.

  • As regards the hedging of market positions (interest rate, exchange rate risk, etc.), the SDR basket could be dominated by derivatives instruments that allow a “fine-tuning” strategy.

The Absence of Clearing Procedures. This argument is widely cited to explain the lack of the private SDR, and many proposals have been made to fill the gap (Coats, 1982; Kenen, 1983). Here there is clearly a “chicken and egg problem.” The correlation between the apparent absence of the private SDR and the lack of a clearinghouse runs both ways. A clearing facility does not have to be introduced until operations on the private market reach a certain threshold. We had this kind of experience with the ECU in the early 1980s when Kredietbank took the lead and initiated the first clearing procedure for operations in private ECUs. This initiative was justified by the emerging private market for ECUs. Conversely, the clearing facility has fed the market. A token, among others, of the present very limited status of the SDR is given by the fact that it has no autonomy vis-à-vis its basket definition. This differs from the situation of the ECU: as already indicated, in the short run, the value of the private ECU as determined by demand and supply conditions could diverge from its theoretical basket definition.

Assessing the SDR in the Light of the Multicurrency System

Stability Considerations in the Multicurrency System

Owing to its basket definition, the SDR conveys an obvious stability. But the discussion must be broadened. When applied to international monetary questions, the concept of stability is ambiguous as it could refer to several interrelated aspects: (1) the efficiency of adjustment mechanisms at the international level, concerning the correction of external imbalances, the removal of possible exchange rate misalignments, etc.; (2) the degree of empirical volatility of exchange rates (nominal and/or real), with crucial issues such as the formation and role of expectations, overshooting phenomena and bubbles, etc.; and (3) the intertemporal stability of a given configuration for the international monetary system. These three approaches to monetary stability sometimes lead to an uneasy trade-off. For instance, a certain degree of volatility in nominal and/or real exchange rates is required to absorb shocks and to accelerate the adjustment process.

The multicurrency system refers to the long-run decline in the international role of the dollar and the associated surge of the deutsche mark and the yen as major reserve currencies. With the long-term changes in the respective market shares of the three major currencies, the international monetary system has become less asymmetrical, with many implications. Does the multicurrency system, compared with the more asymmetrical monetary regime of the pure dollar standard, facilitate the adjustment process? There is no general answer to this question, as there are very few general results derived by industrial economics in the case of an oligopoly. Intuitively, it is likely that up to a certain point currency diversification could strengthen the stability of the system by better absorbing shocks and distributing risks more widely. But this kind of statement needs to be supported by both theoretical models and empirical estimates. The transition from the monetary monopoly of the dollar to an oligopoly—the so-called monetary triad—is gradually changing the geographic distribution of seigniorage. As regards systemic stability, the existence of a hegemon could be required, as assumed in the theory of hegemonic stability (see Keohane, 1984). The analogy with market forms could be enlightening. A Stackelberg duopoly extended to the oligopoly case could facilitate the adjustment compared to the Cournot solution. At least this is somewhat suggested by the experience of the Bretton Woods regime and the working of the EMS. In both cases it was (for the Bretton Woods regime but also regarding the implementation of the 1987 Louvre agreement) and is (for the EMS) necessary to distinguish at least two dimensions in the issue of symmetry: (1) the degree of symmetry between the reserve currency countries and the others, between surplus and deficit countries (clearly the two criteria do not lead to the same taxonomy); and (2) the degree of symmetry in central banks’ interventions on the foreign exchange market.

Since the end of the 1960s, that is, before the breakdown of the fixed exchange rate system, the dollar has been challenged without being replaced. This more symmetrical situation could explain a part of the ongoing financial instability. This instability could be the price to be paid for this (long) monetary transition. Therefore, we cannot infer that more symmetry necessarily means less instability.

Inertia in the International Monetary System

When applied to international monetary questions, inertia refers to some aspects of public and private agents’ behavior and the relative rigidity, at least in the short run, of the market shares of the various reserve currencies. Table 3 gives the market shares of the U.S. dollar, the deutsche mark, and the yen for nongold official reserves, denomination of external bank loans, and external bond issues during 1980–94.

Table 3.

Shares of Reserve Currencies

(In percent)

article image
Sources: International Monetary Fund; Organization for Economic Cooperation and Development.

Figures for the ECU were given in Table 2. We have information about the market share of the SDR in official reserves, but since there is no reliable global information about external bank loans and bond issues, its market share could be nil or negligible.

Over the period, the market share of the dollar has seen a downward trend, with a low slope for the share in official reserves and a more pronounced slope for the denomination of external bond issues. The deutsche mark and the yen have been the “winners,” and they have increased their market shares significantly at the expense of the dollar. The record for the denomination of external bank loans is somewhat different, as in 1993–94 when the dollar posted a market share of about 80 percent, close to that of the early 1980s.

Coming back to nongold official reserves, an estimation of the simple linear trend over the period 1976–94 leads to the following conclusion: an extrapolation of past and current trends indicates that, ceteris paribus, the market shares of the dollar and the deutsche mark in nongold official reserves would be about the same in about 2023. For the dollar and the yen, they could be equal in about 2028!1 This is not a scientific proof of inertia in market shares of reserve currencies. It is just a way of underlining that even if the dollar is more and more contested, it is not going to be fully replaced by another reserve currency in the short and medium term. The monetary triad is going to stay quite asymmetrical for some time. The simplest linear trend is certainly not the best way to calculate inertia phenomena and the ceteris paribus assumption is certainly not valid. Looking back at the history of the international monetary system, we know that (1) a transition from a monetary pattern to another one could be long owing to inertia and other related phenomena; and (2) history is not linear and it shows many examples of discontinuities, institutional changes, etc., leading to major breaks.

Other things will not stay equal if, for instance, Europe is in a position to comply with the Maastricht deadline and to introduce the single European currency (the euro) as of 1999. One of the aims of the founding fathers of the euro is to better balance the international monetary system. If the euro quickly achieves substantial credibility because the markets consider it the apparent successor of the deutsche mark and appreciate the economic policies and the degree of political union within the euro zone (this is just a scenario!), the drop in the market share of the dollar (for official reserves and bond issues) will accelerate. In Figure 2, the intersection would then take place well before 2023!

Figure 2.
Figure 2.

Proportion Rules on Development of the SDR

Many uncertainties still remain about the transition to the single European currency and its conditions, as many uncertainties also remain about the future of the dollar zone and the yen (the concept of the yen zone is rather ambiguous as many Asian countries peg their currencies to the U.S. dollar, which remains the dominant currency in the Asia-Pacific region). Nevertheless, the introduction of a credible euro could have a major impact on the market shares of the reserve currencies. According to the scenario envisaged by the European Commission and the European Monetary Institute, the euro will become a fully fledged currency by early 2002 at the latest. One short-term goal for the euro could be to achieve a market share at the world level that is at least equal or possibly superior to the “sum” of the market shares of participant currencies (after deducting any possible double counting). In both the short and the long term, market shares of the euro will be conditioned by the strength of the European economy, the degree of political integration and stability in the zone, and the monetary-fiscal mix.

What are the relationships between the market share of each reserve currency on the one hand and the level and volatility of its exchange rate on the other? Empirically, there is a dramatic contrast between the relative inertia of market shares of reserve currencies and the overshooting in the adjustment of nominal and real exchange rates. For both official and private holders of dollars, it could be expected that the exchange rate risk is managed, possibly hedged. When the dollar is dramatically overvalued (as in 1981–85, for example), investors can be induced to increase their exposure in dollars, particularly at the start of the appreciation period, if they can make rational expectations about the timing and magnitude of this period. Conversely, when the dollar is undervalued (as it has been vis-à-vis some European currencies and the yen over the past three to four years), holders can be led first to reduce their exposure in dollars and to increase it later when they believe that the undervaluation phase is going to be reversed.

Table 3 suggests two conclusions. First, as regards nongold official reserves, the share of the dollar has been rather inelastic to exchange rate considerations. It increased slightly in the early 1990s despite the undervaluation referred to above. For the period of the bubble on the dollar, its share rebounded slightly during 1981–84. But other explanatory variables would also have to be considered. In particular, the recent rebound in the share of the dollar in nongold official reserves must be related to the huge intervention on the foreign exchange market by several major central banks (the Bank of Japan, the Deutsche Bundesbank, etc.) during 1994 and 1995.

Second, there is an apparent correlation between the use of the dollar for the denomination of external bonds and its external value. It is likely that the lasting undervaluation phase (at least vis-à-vis some European currencies and the yen) could have deterred some investors. Other variables and further empirical work are needed. In order to test any relationships between the market shares and the external value of reserve currencies, intra-annual data and lags in effect would have to be considered.

The distinction between the market share and the external value of any reserve currency has to be recalled. How should a “hard” currency be defined? Depending on the criterion that is privileged by the analysis, a currency could be said to be “hard” if at least one of the following conditions is fulfilled: (1) it bears a low real interest rate (it generates trust in the minds of the investors); (2) its external value is high (also because its internal purchasing power is stable); or (3) its market shares in international transactions are high. Very often, criteria (1) and (2) are closely connected. Criterion (3) refers more to the international role of a reserve currency. At certain periods, the three criteria have converged for the pound (when it was the key reserve currency) and for the dollar. They could be mutually consistent, but in many historical circumstances they have to be considered separately.

The inertia assumption relates not only to international financial transactions but also to commercial transactions. Information about the invoicing of exports and imports is piecemeal and national. Tavlas (1991) estimated the currency denominations of world exports for 1980 and 1987. In order to aggregate information available at the national level, he combined two sets of data: the currency compositions of exports for the six largest industrial countries and for members of the Organization of Petroleum Exporting Countries (OPEC), on the one hand, and the share of each country (or zone) in world exports on the other. The market share of the dollar went down from 34.5 percent in 1980 to 24.8 percent in 1987, while the market share of the deutsche mark and the yen rose from 10.2 percent to 12.4 percent and from 2.0 percent to 3.5 percent, respectively. It is likely that an updating of Tavlas’s work would confirm the downward trend of the dollar taking place with short-run fluctuations but also with much inertia.

Tables 4 and 5 give some information about the invoicing of French foreign trade and German exports. The French case is certainly not representative of what has occurred for world exports and imports, but it can lead to partial observations.

In the French case (Table 4), the relative stability of the share of the deutsche mark contrasts with some opposite moves in the shares of the dollar and the franc. Since 1980, the share of the dollar in the invoicing of French exports has increased, at the expense of the franc. The opposite applies to French imports. The figures in Tables 4 and 5 suggest a large inertia in invoicing practices.

Table 4.

Currency Invoicing of French Foreign Trade

(In percent)

article image
Source: Banque de France.

Although inertia is a word, it also corresponds to facts, but it does not constitute a theory in itself. How should the relative stability in the market shares of the three major currencies be understood? Several arguments have to be emphasized.

After the dollar, still the dollar? Namely, given the economic position and the political role of the United States, the substitutability between the dollar, on the one hand, and the deutsche mark and the yen, on the other, remains imperfect. These latter two currencies, particularly the deutsche mark, periodically play the role of a safe haven, but they cannot count on the same deep financial markets and political credibility as does the dollar. It must also be recalled that until recently both the German and Japanese monetary authorities were reluctant to promote the international use of their currencies. Both countries have been keen not to expose too much the external value of their currencies to speculative movements, and the Bundesbank was and still is reluctant to expose its monetary policy to deep and volatile international influences.

Table 5.

Currency Invoicing of German Exports

(In percent of total exports)

article image
Source: Deutsche Bundesbank.

The market for reserve currencies is not very contestable (the expression refers to the concept of a perfectly contestable market, as defined in industrial economics), and this evidence creates some gap vis-à-vis the Hayekian model of competing currencies (Hayek, 1976). Barriers to entry are numerous (existence of critical size for market shares, existence or absence of deep and resilient markets to support any major currency, and the economies of scale argument that gives a premium [lower transaction costs, etc.] to the key currency).

Inertia in monetary behavior, which is consistent in the short run with overshooting phenomena and large portfolio shifts, also has to do with long-term trade relationships and the existence of “preferred habitats” for many users.

Competition Among Currencies and Asset Substitution

The perceived quality of a currency determines both its value and, in the long run, its market shares. How to envisage the quality concept? According to Hayek (1976), two “not wholly unrelated dimensions” have to be taken into consideration: (1) the expected behavior of the value of the currency (stability or variability) and (2) its acceptability (or degree of liquidity; in Hayek’s analysis the two words are seen as equivalent).

The first criterion assesses any money by its capacity to have a purchasing power approximately constant in time. This criterion is emphasized in various proposals for the construction of a “hard” SDR, which would preserve its purchasing power in terms of goods and services (see below). In the competitive environment, economic agents (official and private) are going to prefer currencies with low inflation variability and low inflation costs.

The second criterion refers to the medium of exchange function but also to liquidity considerations. As underlined by Tavlas (1991), a reserve currency must rely on domestic capital markets and financial intermediaries that are deregulated (for prices and quantities), deep and resilient (for the markets), and sophisticated. For example, the Bundesbank was reluctant about financial innovation until the end of the 1980s. It had to accept the changes and in some cases accelerate them in 1988–90 in order to maintain and improve both the external competitiveness of German banks and Frankfurt’s position as a financial center.

Behind the acceptability and stability criteria, there is the trust that is also conditioned by political factors.

Asset substitution is conditioned by both economic and political factors, which determine the relative quality of the currencies. Among a long list of determinants, interest rates and exchange rates interfere with various categories of risk premiums, which depend on the degree of asset substitutability. The tension between currency concentration and diversification is as permanent internationally as it is domestically. Concentration is justified by economies of scale (in a Baumol-Tobin type of model) and the reduction in transaction costs. It could also be the consequence of lexicographic preferences, derived from “safe haven” considerations, etc. Diversification is rational for risk averters. Whatever the definition of the SDR, the current or another one, it is and will continue to be exposed to the implications of the concentration/ diversification trade-off.

The SDR is not fully in the market for key currencies for the numerous reasons given above. Moreover, it is facing two serious drawbacks. First is the lack of political credibility. Behind the basket of the five major currencies that define the SDR, what is the current cohesion and credibility of the Group of Five? It is very limited indeed. I do not want to suggest that it would be a good idea to add the lira and the Canadian dollar to the current basket because the Group of Seven is more institutionalized than the Group of Five. Certainly, such an addition would not lead to a “harder” SDR. But the lack of institutional and political backing is one of the major drawbacks that the SDR faces. The SDR is backed by the IMF and some participants in the SDR scheme, which is necessary but not sufficient. It lacks political support and a vision.

The second drawback (a technical one) is the basket definition. A basket currency could be competitive as a unit of account and as a store of value in a world of high financial volatility. As a medium of exchange it is generally outclassed by its components or by other currencies (de Boissieu, 1988). A basket of currencies is exposed to higher transaction costs. Even if these costs are decreasing as innovations are implemented within the banks and in capital markets, the competitive disadvantage still remains. This applies also to the ECU under its current basket definition, although the ECU has gradually gained some autonomy vis-à-vis its basket definition. The failure of the ECU as a medium of exchange (see its very low market share for invoicing EU member countries’ foreign trade, as illustrated in Tables 4 and 5) can be explained by many factors (it is not given the status of a legal tender, etc.), among which the argument of transaction costs is not negligible.

What to Expect?

Elements of a Prospective Overview

Can we figure out the general conditions of the equilibrium between the demand for and the supply of nongold official reserves in the future and the role that the SDR can take in the general adjustment process? Certainly not. Nevertheless, four aspects deserve some attention.

  • The external position (calculated on the basis of the current account and the official settlements account) of the reserve currency countries. The correction of current imbalances is going to take much time. For example, the effective or potential dollar overhang will stay for some time; the United States will keep posting current account deficits; and Japan will continue to count on a large current surplus for many years. The balance between the demand for and the supply of reserves has to be assessed in the light of the pattern of national savings ratios and external imbalances and the degree of sensitivity of these imbalances to exchange rate adjustments.

  • The evolution of the creditworthiness of deficit countries. In international markets, some rationing will remain in effect, being implemented through price and nonprice channels. There is no major reason for financial innovation to slow down. On the contrary, process innovations are currently accelerating and there is a significant potential for product innovation. The privatization of the international monetary system is largely irreversible within the next ten years (beyond that, we must not fully discard the possibility of a change as the privatization of the system in the long run relates to a trend but also to long-term fluctuations).

  • The exchange rate regime. In the short run, the scenario of a target zone system at the world level, comparable to the one implemented after the Louvre accord, is very unlikely.

  • The implications of the transition to a single currency in Europe. A direct consequence is going to be a new definition of the basket. The SDR will be defined by a Group of Three or a Group of Four basket. Taking into account the weights in effect since January 1996 and extrapolating them in the future (just for simplification), it appears that the share of the euro zone could vary between 32 percent (with Germany and France belonging to the euro zone, but not the United Kingdom) and 43 percent (with the United Kingdom also participating in the euro zone). More important, as suggested above, the single European currency, provided that it gets sufficient credibility, could speed up the decline in the U.S. dollar market share and seigniorage. Many uncertainties still surround the transition to a single currency (the timetable, the interpretation of the convergence criteria that will condition the future of the multi-speed Europe, the monetary relations between the “ins” and the “outs,” etc.). The way Europe faces these challenges will have lasting implications for the competitiveness of the euro and for the working of the multicurrency reserve system.

I have analyzed elsewhere two versions of the multicurrency system (de Boissieu, 1988). In the weak form of this system, one “complete” currency is gradually supplemented with (and alleviated by) the development of other reserve currencies that remain “partial” (according to Hicks’s definition). According to the strong form of the multicurrency system, competition is more acute as it takes place among “complete” monies. Whether the weak or the strong version of this system prevails has many implications: for the overall level and geographic distribution of seigniorage, the forms and the effectiveness of adjustment processes, the way lexicographic preferences (as exemplified by Gresham’s law or its reverse formulation) can operate, etc. The weak form of the system is currently effective, and I believe that, owing to the inertia phenomena cited above and the imperfect asset substitution between the dollar and the other reserve currencies, this weak form will stay in place for many years. As long as asset sub-stitutability remains imperfect, a limit for exchange rate fluctuations exists.

In any multicurrency system, sound and credible national economic policies—particularly the monetary-fiscal mix—in reserve currency countries are a condition for systemic stability. But, the multicurrency dimension of the system calls for more effective multilateral surveillance. With competing currencies, how can reserve countries be induced to accept an ex ante coordination of their policies? I think that we will continue to live in a world in which the degree of policy coordination is going to lag much behind the degree of financial integration. The resulting gap is somewhat unavoidable, and it has to be considered when analyzing international systemic stability. The coordination of national prudential policies (through the exercises taking place in Basle and Brussels) has been much more credible and effective so far than the coordination of monetary policies.

Asset price variations and portfolio shifts are the token and the channel of the adjustment process. When analyzing volatility in a muticurrency system, the issue of spillover effects is particularly crucial. The spillover of price volatility from foreign exchange to capital markets (and vice versa) constitutes one important dimension. The spillover of volatility from one currency area to another area is another relevant dimension. For instance, the removal of any exchange risk within the euro zone could, under certain conditions (in certain models), increase the volatility of the exchange rate between the euro and some “out” countries (such as the lira and the pound, provided that these currencies do not participate at the start in the single currency), between the euro, on the one hand, and the dollar and the yen, on the other. Under other conditions, the opposite conclusion would apply.

A “Niche” for the SDR?

In this world of competing currencies, the SDR is going to remain in the short and the long run a partial money. Any policy designed to improve the status and the role of the SDR must consider what I call “proportion rules.” First, the official use of a currency cannot develop over a certain threshold if it is not sustained in the medium and long term by developing the private use of the same currency; conversely, it is clear that the private use needs to be supported by a sufficient official use. Second, concerning the private use of any currency, the financial use cannot develop over a certain limit if it is not sustained in the medium and long term by the use of the same currency in commercial transactions (and vice versa).

These proportion rules based on several historical experiences (including the case of the ECU) state that the functions of money are separable up to a certain limit and not over this limit. Panels A and B of Figure 2 depict this phenomenon, suggesting that for a currency to become a complete money it is necessary for it to develop within the hatched cone in each panel.

Until now the SDR has (modestly) developed alongside the vertical axis in each panel. The proportion rules presented above indicate that alongside the vertical axis further development of the SDR could be limited by rather low ceilings (CC and C’C’).

A relevant application of the proportion rules concerns the following question: can and must the SDR be used in market interventions by national and/or international monetary authorities? By definition, the scope for official interventions in SDRs is limited by the apparent absence of the private SDR.

To reverse the trend in the market share of the SDR in reserves, a new allocation is desirable. It is clear that allocation cannot be decoupled from two other dimensions: distribution and substitution.

Allocation and Distribution. The rule according to which SDR allocation must be proportional to quotas is open to discussion. The (unjustified) image of the SDR as a development financing tool has certainly not improved its competitiveness vis-à-vis the major key currencies. I think it is possible to introduce some limited selectivity in the allocation process without blurring the image of the SDR.

Allocation and Substitution. Even if it has to be closely monitored, inflation is not currently the main threat compared with the two periods of SDR allocation. Nevertheless, if the very idea of a new SDR allocation generates a significant threat of inflation for some major countries, a gap between the gross and the net monetary impulse caused by partial monetary sterilization could be obtained. For many political and technical reasons (the burden of exchange rate risk, etc.), the creation of a substitution account will not soon return to the agenda. Nevertheless, some incentives could encourage a partial and limited substitution of official SDRs for other reserve currencies.

Technically, it is conceivable to improve the attractiveness of the SDR. Let us consider two dimensions.

The Creation of a Clearing Mechanism for Operations Denominated in SDRs. As already suggested, this mechanism cannot be envisaged as a deus ex machina. It would facilitate the development of the private SDR but it would not be implemented until this development had reached a certain size. The experience with the ECU leads to several conclusions. Some financial institutions have to take the lead. Decentralized initiatives and lobbying activities could be successful (see the role of the ECU Banking Association concerning the clearing system organized with the technical support of the Bank for International Settlements).

Creation of a “Hard” SDR. We are not short of proposals, more or less inspired by the example of the ECU and the European debate in the early 1990s (for a clear overview, see Thakur, 1994). Over the years the share of the dollar in the SDR basket has slightly dropped. After the last revision, the weight of the dollar (39 percent) remains higher than the share of the deutsche mark in the ECU basket (close to 30 percent). It must be recalled that the U.K. “hard ECU” proposal was very different from, for example, the Spanish proposal to “harden” the ECU basket. The U.K. proposal was to introduce a parallel currency (the (n+l)th) in the system and to leave market forces to determine its role and market shares. As far as the SDR is concerned, four points have to be emphasized.

  • The SDR is already a (limited) common currency.

  • Given the economic and political circumstances, the SDR will keep its basket definition for some time.

  • The attractiveness of the SDR basket (and therefore its competitiveness) could be improved by preserving its value in terms of other currencies (from the technical viewpoint, this is more difficult at the world level than within the EMS owing to floating exchange rates) or in terms of goods and services. Thakur (1994) presents interesting simulations in the case where the “hard” SDR would preserve its purchasing power over goods and services. This line of reasoning raises two major issues: (1) like “hard” ECU proposals, the “hard” SDR would generate both economic and political difficulties (the attitude of several major countries vis-à-vis indexation procedures and the potential systemic instability generated by increased currency competition); and (2) any insurance concerning the value of the SDR (either in terms of other currencies or goods and services) would have to be offset by an interest rate lower than the current one.

  • With its basket definition, the current one or a “harder” concept, the role of the SDR as a unit of account and a store of value could be extended. Either nothing or a very marginal outcome has to be expected regarding the medium of exchange function.

The drawbacks of the SDR in the multicurrency system are less technical than political. The political economy of the SDR will remain the clue. As long as the SDR is not backed by an identifiable central bank, by a monetary policy (or, more broadly, a monetary-fiscal mix), and by a “lender of last resort,” it will have difficulty achieving credibility and marketability. Many crucial issues are at stake: the degree of coordination within the Group of Seven and later the Group of Three (with the euro), the balance between rules and discretion, and the forms and effectiveness of multilateral surveillance, including the role of the IMF.

Kathryn M. Dominguez

The SDR was initially introduced into the Bretton Woods par value system to solve the Triffin dilemma. The dilemma arose in the 1960s during a period in which the U.S. balance of payments was in chronic deficit. If the United States continued to run balance of payments deficits, its ability to maintain gold convertibility would be undermined, but if it eliminated its deficits, this would lead to a global reserve shortage. SDRs were created to provide liquidity, to reduce dependency of the monetary system on gold, and, most important for this discussion, to alleviate global dependency on the dollar as the primary reserve asset. When the Bretton Woods system collapsed, however, so too did at least the first two rationales for the SDR. After the introduction of the SDR in 1969, an excess supply rather than an excess demand for dollars existed, the gold standard was suspended in 1971, and countries allowed the values of their currencies to be determined by market forces in 1973, thereby, in principle, mitigating the need for foreign exchange reserves. The purpose of this seminar is to explore the potential roles of the SDR in the current international monetary system.

The question of whether a role for the SDR continues to exist hinges on at least three issues. First is the issue of global liquidity: does the international capital market currently provide countries adequate liquidity? The liquidity question is examined in a number of the other papers at this seminar, so I will comment only briefly on this issue. Second is the issue of the intrinsic nature of the SDR as a form of money. This is the issue on which de Boissieu mainly focuses in his paper. And third is the issue of the role and currency denomination of foreign currency reserves. I will mainly focus my remarks on this question. I will begin, however, with the first two issues, liquidity and “moneyness.”

The original purposes of SDRs were clearly overwhelmed by economic events, and in the 1970s it was the international capital market and not the SDR that became the primary provider of global liquidity. Few dispute that the SDR is largely a redundant resource for industrial countries with access to international capital markets. For many developing countries, however, access to capital markets is limited and comes with significant risk premiums. If financial markets are efficient, this would reflect not a liquidity problem, but simply that these countries are not creditworthy. However, the recent events in Mexico suggest that markets may not be perfectly efficient. International capital markets initially underestimated the economic and political risks in Mexico; once this misjudgment was corrected, investors rapidly withdrew capital from Mexico and other developing countries in numbers and amounts that are hard to understand given the data. Evidence suggests that access to credit deteriorated for many countries in the aftermath of the Mexican crisis, even though nothing intrinsic happened in those economies to change their creditworthiness.

If financial markets are not efficient, there is a case for public provision of liquidity to those countries whose credit is rationed by the market. It does not necessarily follow, however, that the SDR is the best source of public liquidity. In their current incarnation, SDRs provide unconditional liquidity. The IMF is mainly in the business of providing conditional liquidity. It remains an open question whether new allocations of SDRs would undermine IMF conditionality programs in certain countries. And, more generally, it is not clear whether SDRs would replace, or complement, current developing country borrowing.

A final issue worth considering is whether new allocations of SDRs would actually improve matters for countries that are credit constrained. In the past, SDR allocations have been distributed in proportion to IMF members’ quotas. Consequently, the countries that are least in need of SDRs will receive the largest share of new SDR allocations. If the rationale for a new SDR allocation is to address market failures in the private capital market, it makes little sense to provide the bulk of the allocation to countries unaffected by the market failures. Moreover, if all countries receive SDR allocations, the market is likely to discount the significance of the new liquidity source, and the market perception of countries that were credit constrained previously is unlikely to change.

Although numerous unresolved questions remain about the potential role of SDRs in providing liquidity to countries that are otherwise credit constrained, the modesty of a new SDR allocation (Camdessus proposes SDR 36 billion) implies that the potential positive and negative effects are also likely to be extremely modest. For this reason, one of the arguments made by proponents of a new SDR allocation, that in times of crisis SDRs can provide needed reserves that would otherwise be unavailable from the private market, is unpersuasive. For example, Mexico would not have been able to forestall its crisis with an SDR allocation of the size proposed by Camdessus.

The second issue is the design of SDRs. Is the SDR optimally designed? The paper by de Boissieu examines why the SDR has failed to be competitive with the other major reserve currencies, and whether it has the potential to be competitive in the future. The paper documents the poor performance of the SDR as a major reserve currency in the current system. It correctly argues that several intrinsic features of the SDR make it an unlikely winner in the competition among reserve assets. The SDR is a “partial” money in that it acts as a unit of account only for the IMF, it is a medium of exchange only for limited operations, it cannot be used in foreign exchange intervention operations, and no private markets or clearing mechanisms exist for it. Perhaps most important, there is little political support for the SDR. The five countries whose currencies make up the SDR basket (the United States, Germany, Japan, France, and the United Kingdom) do not advocate a larger role for the SDR and have not been in favor of new allocations of SDRs since the second allocation in 1981. De Boissieu correctly concludes that although the SDR can technically be made more attractive by encouraging its private use, its minimal role in the current system is unlikely to change without a fundamental change in political attitudes.

The optimal design of the SDR is also examined in Wolf’s paper. Specifically, Wolf examines the optimality of the current SDR basket weights that reflect the relative shares of countries in exports of goods and services and the relative shares of the five currencies in official reserve holdings. He finds strong evidence that the SDR weighting scheme is not optimal. However, because countries can augment the SDR weights by holding additional currencies in their reserve portfolios, this design flaw is not very important.

The third potential rationale for SDR allocations is to alleviate the global dependency on the dollar as the primary reserve asset. Many economists predicted that this rationale for SDRs would become irrelevant when countries allowed their currencies to float. The experience of recent years, however, is that countries continue to hold substantial foreign currency reserves, and the dollar continues to be the dominant reserve currency.

The current system is a multiple reserve asset system, but the dollar makes up 63 percent of central bank reserve accounts. Reliance on the deutsche mark and the yen has increased only very slowly over time. Is it optimal for the dollar to remain the dominant reserve currency? The answer to this question in the late 1960s was no, and this, again, provided one of the rationales for the creation of the SDR. From the perspective of the United States, the continued popularity of the dollar has benefits and costs. On the positive side, the United States gains seigniorage when other countries hold dollar reserves. On the negative side, large holdings of dollars outside the U.S. monetary system may increase the vulnerability of the dollar to circumstances outside the control of the U.S. Government. If large numbers of foreign central banks decide to sell dollars at the same moment, this could dramatically influence the value of the dollar. Large portfolio shifts away from the dollar may arise with the introduction of the euro or the formation of an Asian currency zone, and such shifts in foreign preferences could have large effects on the dollar exchange rate.

Why has the dollar remained the dominant currency in official reserve holdings? Academic research does not provide a fully convincing explanation. A similar puzzle arises in explaining why U.S. investors currently hold more dollar-denominated assets than is optimal. The most popular hypothesis is that there is a strong status quo bias in central bank (and private) portfolio allocations. In the early 1960s the United States was the largest economy in the world, and the dollar was the optimal reserve asset. Central banks and private investors have been slow to adjust their portfolios as the relative position of the U.S. economy has eroded over the last twenty years.

Does it make sense to reduce the role of the dollar in central bank portfolios? And can the SDR serve a role in this process? Modern portfolio theory generally argues in favor of diversification. And it is difficult to imagine a reasonable model that would suggest that the optimal weight on the dollar is currently over 60 percent for most central bank portfolios. As I mentioned earlier, a sudden shift away from dollar holdings has the potential to destabilize the value of the dollar. This is where the SDR might serve an important function. A new allocation of SDRs could provide a means by which the international system could make a one-time switch away from dollars without adversely influencing the value of the dollar. This idea of using SDRs as a means of moving the dollar reserve system to a multicurrency system is not a new one. Indeed, Peter Kenen and Jacques Polak had this diversification process in mind when they developed the concept of a substitution account.

In conclusion, the role of the SDR in the current international monetary system is extremely limited. De Boissieu’s paper does an excellent job of documenting the numerous technical and political reasons for the SDR’s failure to compete against the dollar, deutsche mark, and yen as the principal reserve asset. De Boissieu and Wolf provide suggestions for improvements in the marketability and basket structure of the SDR, but they (and I) are not optimistic that these improvements will enhance the position of the SDR in the current system. A new SDR allocation will provide currently credit-constrained countries with additional liquidity, but there are many arguments made in other papers at this seminar that suggest that other methods of providing liquidity may be preferable. The one area in which the SDR may actually serve a useful future role is in helping to diversify central bank portfolios. However, the idea of a substitution account was brought up years ago and dismissed. Given that there seems to be little enthusiasm for a substitution account even among current proponents of the SDR, it seems that even this potential role for the SDR is unlikely to become important in the near future.

Makoto Utsumi

Being stimulated by Christian de Boissieu and other earlier speakers, I would like to begin by posing three questions. First, as has long been discussed since the collapse of the fixed exchange rate system, has the need for official reserves declined? My answer is no. In the past fifteen years, we have learned that the market can be misled and can move apart from so-called economic fundamentals for quite a long period. The first example was the first half of the 1980s, when the dollar was overvalued despite the shift of the U.S. current account into deficit. The second example was the period after the start of the Clinton Administration in 1993, which was characterized by the excessive strength of the dollar and a weakening of the yen that lasted for more than two years, and which finally turned toward more normal levels in about mid-1995. In such a floating exchange rate system, the pure and simple market approach cannot be justified. Each country should be prepared to intervene, to confront the misled market trend.

On the other hand, it has been pointed out that the development of an international financial market has alleviated the need for official reserves. On this argument, I would draw your attention to the experiences of the debt crisis in the 1980s and the introduction by the Bank for International Settlements of capital requirements for banks that have clearly been reducing the international financial intermediation by the global banking sector. In addition, as was seen in the Mexican peso crisis in January 1995, a relatively small misalignment of policy may cause huge and serious market turbulence, which may eventually spill over to other currencies. Each country feels the need to prepare for such a situation. Therefore, the need for official reserves—contrary to the general expectations at the outset of the floating exchange rate system—has not been reduced and is not likely to be reduced in the foreseeable future.

Second, is the multiple reserve asset system the source of instability of the functioning of the international monetary system? My answer, to a certain extent, is yes, but not so much as was thought, and it can be quite a positive factor for the more stable functioning of the system. For example, between 1993 and 1994, the large portfolio shift of some Asian countries caused instability in exchange markets. The central banks of some Asian countries tend to be very eager to deal in the foreign exchange market. But the merits of the multiple currency reserve system compared with the single currency reserve system certainly exceed its demerits. In this respect, I have a slightly different view from that of de Boissieu, because I think we can expect competition among the reserve currency countries to lead to the pursuit of sound economic policy. A multiple currency system can be much more stable than a system that depends totally on the United States.

In this connection, de Boissieu mentioned that the share of the U.S. dollar in reserves is inelastic with respect to changes in exchange rates. On this point, I want to remind you that in the process of the weakening of the dollar for long periods, a huge amount of intervention was exercised to purchase dollars, which resulted, ironically, in a rapid increase in the official reserves of the weakening currencies. Anyway, the multiple currency system is a fact of life. It is an undeniable trend, so we should deal with this trend.

Third, what should be the role of the SDR in the multiple reserve currency system? That is the fundamental and central question of this session. As many previous speakers have pointed out, the reality is that the SDR is far from being the principal reserve asset in the international monetary system. Still, even if the rationale of the SDR has diminished under the multiple reserve currency system, I think the SDR has a legitimate right to occupy its proper place in the system. By holding SDRs, the monetary authorities can diversify exchange risks. If the share of the SDR in total reserve assets increases, the instability caused by portfolio shifts can be diminished. But with the present very limited share of SDRs, how can we expect this benefit to work? I think the most distinguishing characteristic of the SDR as a reserve asset is that through its allocation and its cancellation the control of international liquidity becomes a possibility.

Unfortunately, we have not yet succeeded in finding objective criteria to measure international liquidity. Total official reserves cannot be equated with international liquidity when the global financial markets have developed to such a magnitude. But when a number of countries, while pursuing fundamentally sound economic policies, still encounter a sudden and temporary foreign exchange crisis, we can say that there is a global liquidity problem. The peso crisis and the way it affected some emerging economies may be an example. Fortunately, in the Mexican case, the IMF and the United States were the suppliers of liquidity. In the case of Argentina, Japan and some European countries joined in these efforts. This example suggests that if the allocation of SDRs could be more flexibly realized, the world economy would be better equipped to prepare for storms in the international monetary system.

At the same time, we should not ignore the possibility that the excessive liquidity caused by the allocation of SDRs was the cause of global inflation. That is why I emphasize the importance of the flexible cancellation of SDRs. Allocation and cancellation should be approached symmetrically. The Fund has allocated SDRs, but it has never canceled them. There are many technical questions to be studied regarding this approach. For example, in the case of cancellation, what should be done for those countries that do not hold enough SDRs to be canceled? One approach would be for such countries to go to the IMF for a stand-by arrangement. In this way, SDRs can be transformed into conditional liquidity.

The condition for allocation of SDRs in the Articles of Agreement, which requires a finding of long-term global need, is totally out of date. But, as a matter of fact, the interpretation of this phrase was already transformed in the debate on the second allocation in 1979. The Articles of Agreement also presuppose the possibility of flexible allocation and cancellation regardless of the basic period of consideration. So the flexible allocation and the cancellation or the control of international liquidity by the Fund through these actions might be initiated early before the Articles of Agreement are amended, which should anyway be done in the future.

Finally, I would agree with de Boissieu on the merits of the SDR as an instrument of market intervention. Before directly touching upon this subject, I want to let you know that the SDR has already had a certain role in this field in the past. In January 1988, the United States and Japan engaged in joint intervention to stop the decline of the dollar. Facing a shortage of yen in U.S. reserves to sell in the market, the U.S. Treasury and the Finance Ministry of Japan agreed to a transaction involving SDRs and Japanese yen. That is to say, Japan committed to supply yen to the U.S. Treasury in exchange for the SDRs in U.S. reserves.

Then what would be the merits if the SDR could be used for intervention? Example one: in January 1995, the United States sold a huge amount of dollars to support the Mexican peso in the market. As a side effect, this transaction caused the U.S. dollar to weaken vis-à-vis the yen and the European currencies. Example two: intervention among European countries often takes place with the U.S. dollar used as an intermediary. As a result, a European currency problem sometimes affects the yen-dollar rate. Example three: even when the deutsche mark is suffering instability vis-à-vis not only the dollar but also the yen and other currencies, the Bundesbank does not like the intervention between the deutsche mark and the yen to be undertaken by itself or even by the Bank of Japan. In such a case, the SDR could function flexibly as an intermediary.

The question, a fundamental one, is how we can make it possible for the SDR to be used as a means of intervention. As de Boissieu has explained, it is necessary for the SDR to be widely used in the private sector. I do not want to repeat what he said regarding the limited use of SDRs in the market, but I would like to emphasize that the SDR, if only because of its name, is very difficult to understand. For example, when members of the Finance Committee of the Japanese Parliament visited the Bank of Japan, one member asked a Bank official, “I have heard that the IMF has issued a new paper money; could you show it to me?” Also, when then Prime Minister Takeshita made a joint announcement with President Reagan on the U.S.-Japan agreement to swap between the SDR and the yen, the journalists accompanying Mr. Takeshita did not understand what the SDR was. One asked, “Isn’t it a mistake, isn’t it not SDR but SDI, the Strategic Defense Initiative?”

I think the SDR, to be popular, should have a new name. The ECU has a very historical tradition. Everybody knows that. The euro may become very familiar. So I think that, for the better future of the SDR, a more sociological or psychological perspective should be taken toward popularizing it.

Paul De Grauwe

I can say to Mr. Utsumi that in the Belgian Parliament there is at least one member who knows the difference between the SDR and the SDI!

Now, let me discuss the paper of Christian de Boissieu, which I liked very much. He develops many interesting ideas about the question at issue, namely, the scope and the future of the SDR in international private markets. I want to organize my thoughts as follows. I first want to develop a historic argument whereby I want to reinterpret history. This will enable us to understand a little better why the SDR had so many difficulties in getting off the ground. And then I will develop a second argument that is related to the basics of money.

Let me begin with the historic argument of why the SDR had a handicap right from the start. Everybody refers to the famous Triffin dilemma: it seems to be the starting point for every discussion about the origin of the SDR. This dilemma led to the view that a substitute for gold was needed. There is today a consensus that it was the right diagnosis of how the international monetary system worked during the 1960s.

I would like to take issue with this view and to argue that the Triffin diagnosis was flawed. Let me, therefore, spell out the dilemma again. As you know, the first leg of the dilemma was that if the United States did not accommodate the increased demand for international liquidity there would be a scramble for liquidity, which, by itself, would lead to worldwide deflation. The second leg was that if the United States did accommodate the increased worldwide demand for international liquidity, this would lead to a confidence problem. That is, the United States would be perceived to be increasingly incapable of converting dollars into gold.

In Triffin’s view, the latter would also lead to deflation. He described this process as being like a classic run on the bank: when confidence in the bank vanishes, people run to the bank to convert deposits into hard cash. And that was how Triffin also saw the operation of the gold dollar standard, which led to the conclusion that whatever happened, the gold exchange standard would lead to deflation. The historic example was, of course, the 1930s, when this had indeed happened. And, therefore, in order to resolve this situation, one needed the SDR as a means of avoiding worldwide inflation.

The striking fact of the postwar period was that there was no deflation but there was a problem of inflation. This became very clear during the 1960s. Paradoxically, at the time economists were discussing reform of the international monetary system, based on this Triffin diagnosis, something quite different was going on, and that was not deflation but inflation.

So what went wrong with the Triffin analysis? I think one can say that during the 1960s the gold exchange standard tended to disappear and to switch into the pure dollar standard system, whereby the control that convertibility could have exerted on the United States disappeared. As a result, the United States was left free to follow expansionary policies. The question arises: why did these other countries not convert their dollar holdings into gold? The answer is that most of the important industrial countries were under political pressure not to do so. There were only a few countries, in particular France, that made the conversion, but they did it in small amounts and on a symbolic scale.

For all practical purposes, therefore, the system that economists wanted to reform in the 1960s had already ceased to exist. In other words, it had evolved into something quite different from what was thought at the time, and, as a result, the need for SDRs to prevent the world from falling into a depression never arose in the 1960s.

As a result, the Triffin dilemma did not really exist during the 1960s. The Triffin dilemma may have been a problem in the interwar period but was certainly not one in the 1960s. Therefore, one can also say that the SDR was a solution invented for a problem that had long ago ceased to be one. And it is therefore not surprising that the SDR never really got off the ground.

My analysis here is a little different from the traditional analysis that says that the SDR did not get off the ground because when it emerged the world had already moved into a flexible exchange rate regime so that the world did not really need it any more. My point is that even in the 1960s, under fixed exchange rates, one did not really need the SDR in the way it was then thought necessary.

Let me note the following: the fact that SDRs were never needed does not mean that they have to be abolished. Sometimes things that have been constructed needlessly can still be used for good purposes. In this connection I was very much impressed by Mussa’s argument that the SDR can be used to correct for a market failure that makes it difficult for developing countries to access capital markets.

Let me take the second point that relates to the issue of why the SDR struggles with the difficulty of emerging in private markets. Here I would like to refer to some basics of money and, in particular, to the collective good nature of money. Some people have also called this the network externality aspect of money. What do we mean? The utility of money for an individual is based on the fact that many other people use the same money. Its utility therefore has everything to do with the fact that many people use the same money and very little to do with the intrinsic qualities of that money.

The implication of the collective good nature of money is that a newcomer will have difficulties finding a niche in the market because it has to be attractive enough to compensate for the advantage that other currencies have by the fact that many people use that money. And even if that new currency is in some sense intrinsically superior, it may be very difficult for that money to emerge.

Let me give you the example that is often mentioned in this connection. The Qwerty keyboard has become the standard almost everywhere, and yet we know that there are better keyboards. The market does not automatically switch to the better one. There is a problem here that is essentially the same as with money. One has to overcome the large numbers and the fact that those who are in the market have a natural advantage that is very difficult to beat.

In this sense, I very much agree with what Christian de Boissieu wrote in his paper: that the ECU, contrary to what is often said, is no more successful than the SDR. Of course, one often refers to the ECU as being more successful as an investment currency. And, indeed, the ECU has found quite a significant part in the market as an investment currency, but this has little to do with the functions of money as a unit of account and as a medium of exchange. And here we observe that the ECU has not been more successful than the SDR.

In order to overcome this handicap, therefore, one cannot rely on a smooth substitution driven by market forces. Much of the analysis about how the SDR could substitute for national currencies is based on the view that the SDR is like a private commodity, very much like the substitution between margarine and butter. When the price of margarine declines a little, consumers buy a little more margarine and a little less butter. This is the model that many people have in mind for how the SDR could find a niche in the market.

The view has led to proposals, like de Boissieu’s, of tinkering a little bit with the SDR, making it a little bit attractive here and there, to make it more popular in the market. I do not think this is really the way currencies are being substituted for one another. There is a collective-action problem that has to be overcome that has to do with the collective good nature of money. To introduce the SDR into the market, some political decision will have to be made. The switchover from one currency to another requires collective action.

Let me give you another example. In Sweden, many years ago, they decided to switch from left-hand-side to right-hand-side driving. Consider two possible models of this change. One would be a slow transition from one side of the road to the other—the market solution, so to speak. And the other would be a collective action in which the Government says, on this day, at that hour, we will all together switch to the new side of the road. In a certain sense we have the same problem with money. To overcome this collective good nature of money, some collective action will be necessary.

In fact, that is what we are trying to do in Europe today with the ECU. We have discovered that the slow substitution of national currencies for the ECU has just not worked. I mean in terms of the ECU as a unit of account and a medium of exchange. The only way it can work is by means of some political decision to start on a particular date, for example, the first day of January 1999, with a new money. This is the only way that it can be done. I think this is also the way that will be necessary if we ever want to make the SDR a currency alongside the others. It would mean, for example, that governments are willing to accept the SDR as legal tender in their own countries.

Let me end with the ECU and the euro. Is it likely that in 1999 we will have the euro? What better way to answer this question than to look at the market and at what the market believes today. So I looked at the forward interest rate with the settlement date five years in the future, that is, in 2001. If the market is convinced that the euro will be in existence, these forward interest rates on the ECU (which in 1999 will be transformed into euro at the conversion rate of one to one) should be equal to the deutsche mark forward interest rates.

Today, this is not the case. There are differences of 50–60 basis points. The interesting thing though is that if you now compute the spreads with the mark of the theoretical ECU, where you use the market forward rates of all the constituent currencies in the basket, you find approximately the same number, which means that the market is really forecasting that the future ECU—the one that will be in existence in the year 2001—will be just like today’s ECU. In other words, the market is skeptical about the existence of the euro because the euro, as you know, would be one in which the basket definition will have disappeared.

The market therefore does not seem to believe that the euro will exist in the year 2001. But, as we have heard from the Managing Director, the market is not always right.

General Discussion

Robert Solomon offered a brief defense of Robert Triffin in response to Paul De Grauwe’s assertion that the Triffin dilemma was “flawed.” It was true that the anticipated deflation did not materialize in the 1960s as Triffin predicted. But the major industrial countries at that time were concerned that the growth of their reserves appeared to be slowing. This concern prompted them to agree to the creation of the SDR.

Makoto Utsumi, responding to Kathryn Dominguez’s observation that the Mexican peso crisis could not have been averted by a modest-sized SDR allocation, referred to the “Hashimoto plan,” proposed in 1990 by Ryutaro Hashimoto, then the Japanese Finance Minister. Under this plan, the major industrial countries would voluntarily open a line of credit to the IMF funded by the SDRs they received in a general allocation. With this line of credit, the IMF would establish a “liquidity facility.” The facility would be used to rescue countries pursuing fundamentally sound economic policy but experiencing a temporary liquidity crisis. This proposal would, in effect, transform unconditional liquidity—that is, an SDR allocation—into conditional liquidity.

John Williamson, seconding Solomon’s defense of the Triffin dilemma, said the unforeseen emergence of inflation in the 1960s did not take away from the fact that the monetary system embodied a dilemma that potentially could be resolved by an additional primary reserve asset.

Williamson agreed with De Grauwe that collective action was necessary for making big monetary transitions, but he disagreed on the form that such action should take. Because the natural role for the SDR was as an international rather than a domestic money, making the SDR an internal legal tender was not appropriate. In any case, establishing the SDR alongside domestic money would aggravate the currency substitution problem. An analogous collective action to establish the SDR in the international domain was at least intellectually conceivable: all countries could outlaw the legal enforceability of contracts written in any currency other than their national currency or the SDR.

Montek Singh Ahluwalia observed a tendency in the discussion to view the SDR as a new money rather than a line of credit. But the SDR exhibited none of the characteristics of money: it was not hard money or a form of reserve money that led to a multiplier, nor did it introduce an exchange rate into the monetary system. Because it was not money, there was therefore a difference between presenting the SDR as a new and different asset that was somehow going to introduce more stability into an otherwise multiple-reserve currency system and simply looking at it as a line of credit that added to liquidity but had little effect on the other elements of the system.

Christian Stals concurred with Ahluwalia’s view. The SDR was an unconditional credit facility and a low-powered reserve asset. Further, the IMF was not a world central bank that created new money to be distributed globally. Why, then, attempt to make the SDR more widely used in the private markets? Was there really a need for an international clearinghouse for the SDR?

Peter Kenen mentioned an additional barrier to the SDR’s wide acceptance. It was essentially a short-term asset bearing a short-term interest rate. Many other reserve assets held by central banks today were longer term and bore a higher interest rate.

Turning to de Boissieu’s point that monetary union in Europe could lead to greater external instability, Kenen said he did not believe there was a fixed quantity of instability in the system that would have to manifest itself one way or the other. European monetary union (EMU) would remove one major source of instability—uncertainty about the exchange rates of European Union (EU) members, as well as the exchange market turbulence to which uncertainty can give rise. He did not think some other source of instability had to crop up in its stead.

Christian de Boissieu replied that there was a risk—although not a certainty—that removing some portion of volatility (that is, unifying the exchange rate of EU members) would lead to increased exchange rate volatility vis-à-vis the rest of the world. He cited the example of monetary relations between the “ins” (members of EMU) and the “outs” (European nonmembers—particularly those wanting to join EMU but not yet eligible). Potentially, the risk of competitive devaluation could increase as of January 1999 (the deadline for the startup of EMU) in the absence of some form of monetary cooperation or exchange rate mechanism between the two parties.

Referring to the discussion about whether the official role of the SDR should be increased, which was related to the issue of whether it should be considered money, de Boissieu drew a comparison between the SDR and the ECU. Both were basket currencies, but the SDR, unlike the ECU, had no autonomy vis-à-vis its basket definition. Its presence in the private market was virtually nil compared with the significant market share commanded by the ECU. Unlike the ECU, there was no clearing mechanism for the SDR on the private market, and central banks did not hold private SDRs. This “decoupling” of the official and the private SDR was, in his opinion, the result of the SDR’s lack of autonomy and its identity as a “partial” money, as well as the uncertainty about its status as money or credit. As for the future, in the absence of an emerging private market for the SDR, it would be difficult to increase its official role. Finally, de Boissieu said, there was no known historical example of a currency defined as a basket becoming a complete currency for a geographical region—a problem with which European policymakers were still contending.

Rudolf Rhomberg referred to the analogy drawn in a few of the seminar papers between the SDR and the artificial language Esperanto. Without a literature and schoolbooks and instruction in that language, Esperanto was dead. Without clearing mechanisms and the development of private SDR transactions, the SDR was dead.

De Grauwe argued that it was important not to confuse the substitution between investment currencies in the private market with the introduction of a new currency in place of another (or a collection of others). On the market, if a particular currency became more attractive as an investment medium, the market—in a series of private decisions—would switch from one to the other smoothly and continuously. In this way, the ECU had gained some market acceptance. But this process of substitution did not work with money in its function as a unit of account and a medium of exchange. The strategy of starting with the ECU in the hopes that it would slowly take over the market had not worked, and the ECU, as a money unit of account or a medium of exchange, was, like the SDR, nowhere in Europe or the world. Moving a currency into the front row of the major currencies was a collective action problem—not a series of private decisions made by market participants—and required a political decision.

Dominguez said that only if it could be shown that the current system was not working could the case be made for the SDR as a major player. If it could be argued that there was not enough liquidity in the international monetary system or that the U.S. dollar made up too large a portion of current reserves, then a case might be made for improving the marketability and, hence, competitiveness, of the SDR, and it would thus assume a larger role in the reserves of developing countries.

De Boissieu noted that it would be as difficult politically to sell the idea of a substitution account as a means of enhancing the role of the SDR as it had been fifteen years ago when that proposal was first introduced. The SDR would gain prominence only through the substitution effect, initiated by market forces and incentives, and not through a substitution account.


  • de Boissieu, Christian, “Concurrence entre monnaies et polycentrisme monétaire,” in International Monetary and Financial Integration: The European Dimension, ed. by D.E. Fair and Christian de Boissieu (Dordrecht: Kluwer Academic, 1988).

    • Search Google Scholar
    • Export Citation
  • Coats, Warren L., Jr., “The SDR as a Means of Payment,” Staff Papers, International Monetary Fund, Vol. 29 (September 1982), pp. 42236.

    • Search Google Scholar
    • Export Citation
  • Coats, Warren L., Jr., Reinhard W. Furstenberg, and Peter Isard, “The SDR System and The Issue of Resource Transfers,” Essays in International Finance, No. 180 (Princeton, New Jersey: Princeton University, Department of Economics, International Finance Section, 1990).

    • Search Google Scholar
    • Export Citation
  • Hayek, F.A., Denationalisation of Money (London: Institute of Economic Affairs, 1976).

  • Hicks, John, Critical Essays in Monetary Theory (Oxford: Clarendon Press, 1967).

  • International Monetary Fund, The Role of the SDR in the International Monetary System, IMF Occasional Paper No. 51 (Washington: International Monetary Fund, March 1987).

    • Search Google Scholar
    • Export Citation
  • Kenen, Peter, “The SDR as a Means of Payment: A Comment on Coats,” Staff Papers, International Monetary Fund, Vol. 30 (September 1983), pp. 65661.

    • Search Google Scholar
    • Export Citation
  • Keohane, Robert O., After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, New Jersey: Princeton University Press, 1984).

    • Search Google Scholar
    • Export Citation
  • Lelart, Michel, Les operations du Fonds Monétaire International (Paris: Economica, 2nd ed., 1988).

  • Rhomberg, Rudolf R., “Failings of the SDR: Lessons from Three Decades,” in International Financial Policy: Essays in Honor of Jacques J. Polak, ed. by Jacob Frenkel and Morris Goldstein (Washington: International Monetary Fund and De Nederlandsche Bank, 1991), pp. 15069.

    • Search Google Scholar
    • Export Citation
  • Tavlas, George S., “On the International Use of Currencies: The Case of the Deutsche Mark,” Essays in International Finance, No. 181 (Princeton, New Jersey: Princeton University, Department of Economics, International Finance Section, 1991).

    • Search Google Scholar
    • Export Citation
  • Tavlas, George S., and Yuzuri Ozeki, The Internalization of Currencies: An Appraisal of the Japanese Yen, IMF Occasional Paper No. 90 (Washington: International Monetary Fund, January 1992).

    • Search Google Scholar
    • Export Citation
  • Thakur, Subhash, “The ‘Hard’ SDR,” Staff Papers, International Monetary Fund, Vol. 41 (September 1994), pp. 46087.

  • Wicksell, K., Interest and Prices, translated from the German (1898) by R.F. Kahn, with an introduction by Professor Bertil Ohlin (London: Macmillan, 1936).

    • Search Google Scholar
    • Export Citation
The trend has been estimated for 1976–94 (19 observations). For the U.S. dollar, we get

where y is the market share of the dollar in nongold official reserves; T is the calendar year (1976, 1977,… 1994); and the t-statistic is in parentheses.

The equivalent equation for the deutsche mark is the following:
and for the yen:
  • 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