Enhancing International Monetary Stability--A Role for the SDR?

The SDR has enjoyed renewed attention lately in the context of debates on international monetary reform. To be sure, the term SDR has been used to refer to three different concepts—(i) a composite reserve asset created in 1969: the “official SDR” as defined in the Fund’s Articles; (ii) a potential new class of reserve assets: tradable SDRdenominated securities issued by the Fund or an investment vehicle backed by a subset of the Fund’s membership; and (iii) a unit of account, which could be used to price internationally traded assets (e.g., sovereign bonds) and goods (e.g., commodities), to peg currencies, and to report balance of payments data. All three are discussed in this paper.


The SDR has enjoyed renewed attention lately in the context of debates on international monetary reform. To be sure, the term SDR has been used to refer to three different concepts—(i) a composite reserve asset created in 1969: the “official SDR” as defined in the Fund’s Articles; (ii) a potential new class of reserve assets: tradable SDRdenominated securities issued by the Fund or an investment vehicle backed by a subset of the Fund’s membership; and (iii) a unit of account, which could be used to price internationally traded assets (e.g., sovereign bonds) and goods (e.g., commodities), to peg currencies, and to report balance of payments data. All three are discussed in this paper.

I. Why Enhance The Role of The SDR?1

1. Why the SDR2? The IMS has demonstrated resilience through the various stages of recent crisis and the role of the U.S. dollar as a safe haven has also been confirmed, suggesting it can be expected to remain the most important global reserve currency for the foreseeable future. However, as emphasized recently by the IMFC and G20, and previously noted by the Executive Board, the current international monetary system (IMS) suffers from a range of problems, which have been detailed in earlier Fund work on reserve accumulation and the stability of the international monetary system (IMF, 2010f); and more recently in the context of examining the Fund’s role regarding cross-border capital flows (IMF, 2010i).3 The SDR may have some potential to improve the functioning of the IMS in three mutually reinforcing ways: as a sui generis composite reserve asset defined by the Fund’s Articles and centrally allocated; as a unit of account; and, as a new reserve-grade security issued by the IMF or a subset of its membership (SDR-bond). Expanding the SDR basket would further support the SDR’s role. More specifically, enhancing the role of the SDR could serve the following objectives:

  • Reduce reserve accumulation/imbalances and strengthen the global safety net: Official SDR allocations are a lower cost alternative to accumulation of international reserves through borrowing or accumulation of current account surpluses. The SDR has the same properties as a swap line in that it allows holders access to foreign currency liquidity. SDRs are especially valuable at times of systemic crisis, as they give confidence to the market that the member can access foreign exchange funding without liquidating assets in financial markets that may be impaired or would be subjected to added stress from simultaneous action by several central banks. In addition to potentially reducing precautionary reserve accumulation, SDR allocations could conceivably (albeit not in the current legal framework) be used as a policy incentive against excess accumulation of reserves for non precautionary purposes.

  • Develop a new reserve asset: Issuance by the Fund (or related investment vehicle) of SDR-denominated securities in sufficient volume could offer a safe haven in the event of disorderly diversification out of the existing stock of assets, as well as offering an alternative mode of Fund borrowing at time of high potential demand for its resources.

  • Reduce impact of exchange rate swings: The SDR unit of account could be used to price global trade, denominate financial assets, peg currencies, and keep accounts and official statistics. The SDR’s basket characteristic provide a less volatile unit of account and store of value than its components when measured in domestic currency terms, thereby helping cope with exchange rate volatility for both the official and private sectors (see Box 1). These benefits are all the greater as the use of the SDR in both goods and asset markets is developed. Such development would allow the SDR to serve as focal point for IMS evolution, a more efficient outcome than several segmented markets in various national currencies.

    SDR Trade Pricing

    Quoting commodity prices (e.g. oil) in SDR has a number of advantages compared with the current system of pricing in U.S. dollar (except for countries using the U.S. dollar as their currency); including:

    Stabilizing export receipts. Assuming that the objective of commodity exporters is to stabilize export receipts in terms of local currency, the ideal pricing strategy for them would be to quote prices directly in domestic currency. However, there are a number of considerations that would make this impracticable or inefficient (e.g. existing pricing practices, where standardized price quotes in a common vehicle currency dominates pricing behavior for homogenous goods, and pricing to market practices). A second best might be SDR pricing for the following reasons:

    • As basket, SDR is less volatile (compared to component currencies) in domestic currency terms so would provide greater certainty over export receipts, especially as trade is based on forward contracts;

    • SDR pricing would eliminate some of the volatility of exchange rate movements, contributing to smoother price evolution. Commodity prices would better reflect underlying market conditions.

    Natural demand. Commodity exporters are importers of other goods and services from different parts of the world, therefore, SDR pricing of exports would better reflect the evolution of their terms of trade and import bill to the extent that a portion of imports in non-dollar denominated. For example, if the euro were to appreciate against the dollar, commodity exporters with a component of imports from Europe would be better hedged if their exports were denominated in SDR, as part of their receipts would then be denominated in Euros even if it does not exactly match the composition of their imports or the invoicing currency.

    Benefit to importers. To the extent that commodity importers have a market-determined exchange rate regime (or simply do not peg to the U.S. dollar), SDR pricing would also stabilize their import bill and, generally, would be a better hedge for the composition of their export receipts.

    By way of illustration, the tables below demonstrate the benefits described above for a number of countries with floating exchange rate regimes. For freely floating currencies, exchange rates are less volatile against SDR, and hence pricing oil in SDR, for example, would reduce the volatility of prices expressed in national currencies.

  • Accommodate a greater role for emerging market currencies in the IMS. The basket could smoothly accommodate a greater role for emerging market currencies in the IMS, notably the RMB: inclusion in the basket would allow holders of SDR-denominated assets to acquire greater exposure to those currencies than otherwise possible, and would create a dynamic conducive to financial deepening and, depending on conditions for joining the basket, capital account liberalization in emerging markets. Generally, as the SDR composition is updated regularly to reflect the relative importance of different currencies in global trade and finance, a system with a liquid market in SDR-denominated assets could flexibly accommodate the evolution of the IMS and mitigate the impact of any sudden shifts in demand for different currencies. However, expanding the basket too much could increase the complexity of the SDR, potentially reducing its attractiveness to reserve managers.

2. Realism. Caution is needed, however, in considering the SDR’s potential:

Regarding official SDRs

  • With their use limited to the official sector, official SDR holdings provide an imperfect reserve asset, as they cannot be used directly for market intervention or liquidity provision.

  • There are tight legal and political constraints to expanding the use of the official SDR, such as the need for an amendment of the Articles to change the way SDRs are allocated or the need for an 85 percent majority of voting power to agree to an allocation of any size.

  • Sharp policy tradeoffs must be borne in mind when considering expanded official SDR allocations, in particular, in balancing conditional and unconditional liquidity provided in crisis times; or balancing risk of “misuse” of SDRs versus their attractiveness as a reserve asset.

Regarding non official SDR

  • There is currently no natural demand or supply for SDR-denominated securities, reflecting the strong network externalities and increasing returns to scale that characterize use of a currency, in its three functions: store of value, means of payment, and unit of account. It is an open question whether the SDR could take on a systemic role without providing all three functions on a large scale.

  • A further question mark concerns the benefits of the basket compared to component currencies. These depend on (i) the frequency with which agents need to rebalance their portfolios; and (ii) the ease of access to the component currencies. For most actors, replicating the current basket is essentially costless given the size, liquidity, and efficiency of global currency markets in the four component currencies; however high frequency rebalancing (e.g. weekly or daily) could involve considerable costs. Inclusion of less easily available currencies would also enhance the basket’s benefits compared to assembling its components.

  • Given these issues, providing impetus for the development of a market in SDR-denominated securities could entail significant upfront costs (e.g., liquidity premium, hedging costs), which the official sector would have to bear, possibly for a long time.

3. Way forward. These obstacles may not be insuperable, but significant political will and consensus across the membership would be needed to address them. That said, most other potential solutions to the problems in the IMS likewise require substantial investment of political will and consensus-building. With this proviso in mind, the rest of this paper outlines some concrete steps that could be taken to enhance the role of the SDR as an alternative reserve asset and unit of account, along with an initial discussion of their benefits, realism, and potential downsides. Clearly, none of these ideas is ready for implementation overnight, but the paper aims to stimulate an open-minded debate that could lead to identifying steps worth further consideration. Needless to say, enhancing the role of the SDR cannot on its own provide a remedy to all the ills of the IMS. But combined with the other efforts intensified in the wake of the crisis—global policy coordination and stronger surveillance; improved global safety nets; orderly capital account liberalization and financial deepening in emerging markets—a greater role for the SDR might help achieve stability as the IMS transitions to a more multipolar equilibrium, and beyond.

4. Outline. The rest of the paper is organized as follows: Section II presents a number of modalities to expand the supply and use of official SDR (as defined in the Fund’s Articles of Agreement) with the aim of reducing reserve accumulation and contributing to resolve global imbalances. Section III discusses steps to develop new SDR-denominated reserve assets. Section IV elaborates on the use of the SDR as a unit of account for global trade and financial instruments to reduce the impact of exchange rate volatility. Section V discusses tradeoffs in expanding the SDR basket. Section VI concludes with issues for discussion.

II. Reducing Reserve Accumulation

A. Expanding the Supply of Official SDRs

5. Why? Expanding the volume of official SDRs is a prerequisite for them to play a more meaningful role as a substitute reserve asset that could thereby contribute to reducing global imbalances. Even after the extraordinary allocation approved in 2009, total outstanding SDRs (204 billion) represent less than 4 percent of global reserves—well under the peak of 8.4 percent reached in the early 1970s. In part reflecting this small share, they are not actively traded. As originally conceived in the late 1960s, SDR allocations were intended to help meet growing global demand for reserve assets in a context where the U.S. dollar’s link to gold constrained the U.S.’s ability to do so. This rationale disappeared in 1971 with the end of gold backing for the U.S. dollar. But it may be relevant again in today’s medium-term context where many countries have been meeting their demand for reserves by accumulating current account surpluses, which have an overall deflationary impact as countries that had current account deficits over the past decade are now trying to narrow their current account deficits. Further, demand for international reserves appears to be growing much faster than traditional sources of supply, with no clear alternatives on the immediate horizon, which could lead to a debasement of the quality of reserve portfolios (see IMF, 2010f). Larger official holdings would also make active management of this component of reserve portfolios more likely, facilitating development of a meaningful market.


Global SDR holdings

In percent

Citation: Policy Papers 2011, 047; 10.5089/9781498339483.007.A001

Source: IFS, WEO

6. How? Regular—say annual—allocations covering a significant proportion of the expected demand for precautionary reserves would meet this objective. An annual allocation of the equivalent of US$200 billion dollars would raise SDRs as a proportion of reserves to a little over 13 percent in the early 2020s.4 Extrapolating from one estimate of precautionary demand for reserves (Obstfeld, Taylor, and Shambaugh, 2008), sums of this magnitude could be sufficient to meet half of the average precautionary demand for reserves over 2000-09. On-going work on designing reserve adequacy metrics could inform the analysis on the appropriate pace for expanding SDR allocations. Such regular allocations could be made if (and only if) Governors representing 85 percent of the voting power of the SDR department5 agree they are necessary to meet a “long-term global need ... to supplement existing reserve assets (...)” while avoiding inflation or deflation.” This provision sets a high bar of political consensus among shareholders, particularly considering the need for legislative approval in some cases.6 The best argument for a long-term global need lies in strong accumulation of reserve assets, which is expected to continue.

7. Precautions. Expanding the volume of SDR allocations would increase the contingent claim on all other participants in the SDR department, who could be required under the designation mechanism to provide freely usable currencies—currently the U.S .dollar, Euro, Sterling, and Yen—in exchange for SDRs up to twice their own allocation of SDRs. This requirement could quickly become burdensome, particularly for members with floating currencies that do not typically hold large official reserves, and could not provide their own currency if not among the freely usable ones. Two complementary solutions could be envisaged. First, expanding the voluntary trading agreements, to avoid recourse to the designation mechanism and concentrate SDR sales on those members most willing to buy. Second, by expanding the current list of freely usable currencies to include a few additional currencies that meet the definition of freely usable currency under the Articles i.e. that are being widely used, in fact, to make payments for international transactions and are widely traded in the principal exchange markets (Article XXX (f)). (Such an expansion in the list of freely usable currency is distinct from and would not necessarily affect the composition of the SDR basket).

B. Limits and Potential Downsides

8. Partial solution. Given its limitations mentioned above, the official SDR may not be perceived as perfect substitute to borrowed or own reserves, reflecting the fact that it cannot be used directly for market intervention or liquidity provision, as well as its low-yield and the relative difficulty to hedge any net positions. In addition, part of the growing demand for reserves is not precautionary, but reflects other objectives such as influencing exchange rate competitiveness. SDR allocations could only influence such intervention if accompanied by policy commitment to reduce reserve accumulation. Indeed in the absence of such commitment, SDR allocations might even encourage further reserve accumulation, by providing a channel for reserve diversification without market impact.

9. Misuse. SDRs are a low cost source of unconditional financing; as their outstanding volume increases, so does the risk of their being used in a way detrimental to macroeconomic stability. Examples include procyclical fiscal financing in inflationary environments, use of SDR as a substitute to adjustment, or use of SDRs that contributes over time to an unsustainable debt burden (e.g., as the SDR rate rises). In such cases, not only may macroeconomic stability be compromised, but also potentially the member’s ability to pay charges on their allocations or eventually reconstitute their holdings.

10. Inflationary impact. The potential impact on money creation of the large one-off allocation in 2009 was considered likely to be small and easily absorbed, and any expansionary impact on global demand to be beneficial. Large, regular allocations are similarly unlikely to prove inflationary, including at times of limited slack in the global economy, as argued by Richard Cooper in Supplement 1 to this paper. No money is created unless countries sell their SDR holdings to issuers of freely usable currencies (and even then, any money creation could be sterilized if the relevant central banks so wished), and no additional demand for real goods and services is created unless the proceeds are spent and there is no countervailing policy response. Essentially, the global inflationary impact of SDR allocations is expected to be limited even assuming large cumulative allocations, so long as central banks issuing freely usable currencies credibly stick to their inflation targets. An additional safeguard is provided by the discretion to not make allocations—or even cancel existing allocations—at times of strong global demand and inflation concerns.

C. Options for Enhancing the Official SDRs

11. Basic improvements. A number of marginal measures could be taken to improve the SDR in its current form. Providing greater certainty over the basket composition through more objective valuation rules could improve the attractiveness of the SDR. Similarly, daily interest rate setting, rather than weekly, would allow reserve managers fair valuation of SDR assets on a continuous basis and facilitate hedging operations (see Annex 2). However, the lack of appropriate daily instruments in underlying currencies remains a constraint (see IMF, 2010h). Other small changes identified in earlier discussions include clarifying and expanding the scope of permissible operations by moving from a positive to a negative list, eliminating the mandatory use of the official exchange rate in SDR operations, and simplifying the reporting requirements to the Fund to record transfers among members.

12. Reconstitution and other safeguards. Reinstating a reconstitution requirement— whereby members would be required to restore their SDR holdings to the same amount as their overall allocation over a given timeframe—would help improve liquidity in the voluntary market by raising transaction volumes and ensuring demand for two-way transactions. It would also reduce the scope for misuse of SDRs as open-ended cash transfers, which can be particularly problematic in cases of fragile debt sustainability. To further reduce the risk of misuse, options could include requiring a discussion of any planned use of SDR allocations in Article IV consultations, and where relevant, an ex post assessment in Article IV reports, as well as in any UFR report. At a minimum, use of SDR allocations should be taken into account in debt sustainability analysis. While such safeguards may not guarantee “sensible use” of allocated SDRs, restraining their use more tightly would go against the principle embedded in the Articles of Agreement that such use is unconditional. It would also attenuate SDRs’ substitutability with own reserves.

13. Advisory group. In order to give comfort that decisions on SDR allocations are not dominated by political considerations, an advisory board of eminent experts—possibly including central bankers issuing freely usable currencies—could be established to provide an independent opinion on matters concerning the provision of global liquidity to guide the Managing Director’s proposals and Board of Governors’ decisions on the need and frequency of SDR allocations. The group could also make recommendations on the cancellation of SDR allocations when relevant (see e.g., Subacchi and Driffill, 2010). While its views could not be binding in the current governance structure of the Fund, such a group could nevertheless help ensure a robust decision-making process.

14. SDR lending/ pooling. Countries willing to do so could—on their own or as a pool— on-lend their SDRs to the ones in need of external financing. The latter would pay the SDR interest rate (possibly with a mark-up to compensate the pool lenders for credit risk), and exchange these SDRs for freely-usable currency to meet their balance of payments needs. This mechanism would make use of the available SDR resources (after the 2009 allocation, G7 and euro area countries could mobilize a total of SDR 111.7 billion, for example). Even if the SDR rate were to rise over the next years, it is unlikely SDR-based financing would become more expensive than alternatives given how it is determined (i.e., as a weighted average of representative short-term rates in the component currencies). Further, there is no rollover risk associated with the use of the SDR (although borrowers and lenders could agree a timeline for the SDRs to be returned (or additional compensation for extended-period loans).

15. More radical reforms. While the aforementioned steps could be taken within the current Articles, they are unlikely to go far enough to allow a much greater role for the official SDR in the IMS. The steps discussed in the remainder of this section may fulfill that objective; however their implementation would require an amendment of the Articles.

16. Contingent allocations (Escrow). To reduce the risks from unconditional availability of significant amounts of financing, a requirement could be instigated to hold all or part of the SDR allocations in escrow for use only in case of shocks, with a decision by the Executive Board or the Board of Governors determining when the SDRs would be released for use (membership-wide, or targeted to countries affected by the shock)—e.g., at times of systemic stress. Such a mechanism would have the downside of reducing the substitutability of own reserves for SDRs and would therefore likely be less effective than regular allocations in containing the demand for precautionary reserves. However, the decision process could be designed to reduce uncertainty over the likelihood of ad hoc SDR allocations at times of financial distress. And keeping the allocations in escrow may help overcome resistance to large regular SDR allocations at other times, particularly when liquidity appears abundant.

17. Targeting. As SDR allocations are distributed in line with quotas, the bulk of the allocations go to countries that are not accumulating reserves, may have little need for additional reserves, or are unlikely to have an active need to participate in SDR transactions. While there is no significant cost from this “misdistribution,” it complicates political decision making on the size of the allocation, as the headline amount is much larger than the actually relevant one. Changing the way in which allocations are distributed could alleviate this problem. Thus, allocations could be focused on countries considered to have low levels of precautionary reserves compared to broadly accepted reserve adequacy benchmarks (ongoing work on reserve adequacy metrics would provide some guidance), or as noted above, on countries affected by a systemic shock.

18. Incentives. One could also envision SDR allocations to be conditional on policies being deemed appropriate according to benchmarks to be determined, or on a forward-looking policy commitment to adhere to a certain level of reserve accumulation. Another model has been proposed by Truman (2010b) under which the membership would agree on a set of policy norms intended to help assess adherence to a member’s obligations; findings of non adherence could lead to freezing the member’s access to SDR holdings or denial of participation in future SDR allocation. A downside of such a proposal would be to imply much lighter pressure on reserve issuing countries, which essentially have no needs for SDRs, than on all other countries. This could also be combined with the escrow approach.

19. Implications for Fund lending. Any move in the direction of conditional and/or targeted SDR allocation would blur the distinction between SDR allocations and IMF lending. Thus, one might ask why go through a cumbersome process of amending SDR allocation rules to achieve something that could be done through IMF lending. One reason would be the objective of increasing the role of the SDR in the system, through an increase in the volume of SDRs in circulation; another would be views on the desirable size of the Fund: SDR allocations (escrowed or not) offer an additional safety valve to deal with systemic shocks that might otherwise overwhelm the Fund’s commitment capacity, unless its resources were kept permanently far above foreseeable needs.

20. Private use of SDR. Allowing private sector holding (and trading) of official SDRs would enhance their reserve asset quality as central banks could use them directly to intervene or extend liquidity to the market, instead of having to go through the Fund’s voluntary market or designation mechanism to exchange their SDR holdings for useable currency, a transaction that can take several days to complete. It could also help spur development of a market in other SDR-denominated assets and contribute to the development of other reserve currencies. Private sector institutions could be interested in holding official SDR based on a calculation that the market may develop further and for the potential benefits of SDR as a source of liquidity. Allowing the private sector to use SDR holdings as collateral to obtain freely useable currencies at times of globalized liquidity squeeze or financial distress could increase the incentives for holding SDRs. This would require consent from key central banks to provide a market for private participants, or opening the Fund’s voluntary market to them, which may require capping the amounts of SDRs eligible for such transactions. Private holdings of SDRs could thus directly contribute to alleviating stress at times of crisis and would mimic the global provision of foreign currency swap lines.

III. Developing New Reserve Assets

21. A less constrained route. Official SDR allocations would clearly contribute to the objective of expanding the supply of safe reserve assets. However, given their limitations and the very high degree of political consensus needed to enhance their role, it is worth exploring whether there would be other ways of creating new reserve assets that could be denominated in SDRs (not to be confused with existing types of reserve assets, that could be redenominated in SDRs; these are discussed in Section IV). The focus in this section is on steps that could be taken by the Fund or a subset of its membership acting together. These include: i) Fund-issued SDR bonds; and ii) other SDR-denominated reserve-backed assets that could facilitate the diversification of reserve holdings and spur market development for other SDR securities.

A. Fund-issued SDR Assets

22. SDR-denominated bonds. The Fund could build on its issuance of notes denominated in SDR in 2009.7 Expanding the issuance of SDR notes would achieve two objectives: directly contribute to increasing the supply and diversity of global reserve assets, and help establish a benchmark yield curve to price SDR-denominated assets issued by others. The Fund would be a low-credit risk issuer since its liabilities are backed by available quota resources, which are claims on the entire membership rather than a single country, and by other GRA assets such as gold. Fund-issued notes could be held both by the official and private sectors to create market depth and liquidity. A regular program of modest issuance would also build conditions that would allow the Fund to tap markets in very large amounts if needed during a global systemic crisis. As these are precisely the times when there is a flight to quality, this would be an automatic mechanism to expand the market in safe assets when the demand is highest. So long as such issuances were undertaken strictly to supplement the Fund’s resources (i.e., for purposes of meeting UFR demand from members), this move would be possible under the current Articles. However, issuance in private markets would require the resolution of questions such as the appropriate maturity structure of the bonds in light of the repurchase schedule of Fund financing, the implications for Fund operations of regulatory requirements in the jurisdictions where bonds would be issued, and the design of appropriate safeguards against perceived conflicts of interest related to the Fund’s surveillance and financing functions. It is also unlikely that it would be possible within the constraint of the current Articles to issue enough paper to establish and maintain at all times a benchmark yield curve.

23. New financial structure. By contrast, expanding the issuance of notes in regular intervals irrespective of a need for supplementary resources would entail a large shift in the role and financial structure of the Fund, requiring an amendment to the Articles of Agreement. Similar design issues would need to be resolved as noted above for issuances to supplement the Fund’s resources. In addition, such a new structure would require a scaling up in the Fund’s investment operations and strategy. Should the Fund face a shortfall in the amount needed to cover payments, there would either have to be resources diverted from other operations (not allowed under the existing Articles), or members would have to commit new funds. Careful consideration would also have to be given to the operational structure of such borrowing operations, including protection to avoid additional conflicts of interest arising from the Fund’s surveillance and financing role and its new role as global borrower and investor of borrowed resources.

24. A longer-term vision. After substantial market depth has been established, which would require very substantial amounts to be issued, and if the perceived safety of other global assets were to decline, Fund-issued SDR securities could be a potential anchor off which to price risk throughout the system (assuming the SDR valuation method ensured the SDR were representative of economic weights in the IMS). In the even longer run, if there were political willingness to do so, these securities could constitute an embryo of global currency. Of course, this longer-term vision is highly hypothetical and support for the nearer term steps need not imply commitment to the longer-term ones.

B. Reserve-backed SDR Assets

25. Substitution account. The idea of a substitution account to exchange members’ foreign currency reserve assets for SDR-denominated claims was considered in the 1970s and eventually rejected for lack of agreement on who would bear the exchange rate risk (see discussion of substitution account in IMF, 2010f). The idea was that an off-market reserve pool managed by the Fund would provide an opportunity for large reserve holders to diversify their reserve assets while limiting the risk of market disruption, particularly a sharp depreciation of the dollar. Use of the account could therefore smooth the process of reserve currency diversification that would accompany a shift in the IMS toward a multi-currency regime and would also increase the supply of SDR securities. However, issuing SDR-denominated claims against assets comprising U.S. dollar or other currencies exposes the account to exchange rate and potentially interest rate risk. Unless participants were not allowed to redeem their claims (which would challenge the liquidity of the securities issued by the account), a risk-sharing mechanism would need to be agreed, and it is not clear why this problem could be more easily surmounted today than in the 1970s.

26. Alternatives. To mitigate exchange rate risks, the account could be structured to exchange SDR-denominated claims for foreign currencies with the same composition and proportions as the SDR basket, and retain a similarly-structured portfolio. However, this would essentially eliminate the currency diversification benefits to members, while not necessarily providing obvious benefits in terms of global diversification of reserve assets (unless the account were to invest in official SDRs—or SDR-denominated securities issued by others than the traditional reserve issuers; this however would expose the account to credit and liquidity risk that would need to be covered by members). Another alternative would involve having a willing subset of the Fund’s membership pool (part of) their official reserves into a Trust managed independently (within or outside the Fund), and agreeing among themselves on risk-sharing rules. This could achieve both reserve diversification and supply of a new kind of SDR-denominated security (issued by the Trust), and would therefore seem the most promising route if members with large enough reserves were interested in pursuing it.

IV. Mitigating Exchange Rate Volatility

27. How? Use of the SDR as a unit of account could mitigate the impact of exchange rate volatility through: i) use of SDR to price international trade, report data on international transactions, and as an exchange rate peg; and ii) denomination of assets in SDR. The latter would require the development of a private market for SDR-denominated assets with impetus from the official sector and support to build the market infrastructure. The rest of this section discusses these issues. SDR valuation is clearly key to determine the optimality for any country or economic agent of the SDR basket as a hedge against exchange rate volatility. It is discussed in the next section. In all cases, an important distinction is between the denomination of certain transactions or securities and their settlement. Use of the SDR for denomination (what is at issue here) leaves full discretion to the parties to choose any currency for the actual settlement of their trade.

A. Pricing, Accounting, and Pegging

28. Foreign trade pricing. While it might be most advantageous for countries to trade in their own currencies to reduce uncertainty over export (import) receipts (payments), in practice, this would not be a globally efficient outcome. As such, use of SDR invoicing and pricing could be advantageous as it would stabilize export receipts (in local currency) compared to single currency pricing, and more generally would avoid exchange rate volatility amplifying commodity price swings (see Box 1). Clearly, the argument does not apply for countries whose foreign trade is predominantly denominated in their own currency, or even for those that trade overwhelmingly in one foreign currency. Generalized use of SDR as a unit of account for global trade invoicing would also create demand for SDR-denominated assets as both the public and private sectors develop exposure to SDR-denominated flows. Liquidity of such instruments would likely be a problem initially however. More generally, premature switching to invoicing global trade in SDR could be disruptive, raising transaction costs as final settlement would have to be undertaken in the underlying currencies. However, a clearance/payments system could be established to settle SDR claims directly (see Annex 2).

29. Data reporting and accounting. To reduce the valuation effects that usually impact statistical reporting, and facilitate cross-country comparisons over time, the IMF could disseminate balance of payment statistics in SDRs (perhaps in addition to US$ initially). The transition to a new reporting system could be costly for some users and would have large administrative costs for the Fund, with few benefits beyond signaling, until SDR usage became more common. Data on international reserves might be an exception where SDR reporting would have immediate benefits however. Further, a number of international and regional institutions already use the SDR as a unit of account for their accounting (e.g., the BIS, which finds that it stabilizes the value of its balance sheet). Some central banks (e.g. RBNZ) use it as a unit of account for internal monitoring of foreign reserves. Similarly, global corporations might benefit from the stability of the SDR. However, global accounting standards currently require that the currency of the country where most of the operations are based be used for official accounting purposes.

30. Pegging. A number of countries peg their exchange rates to a basket of currencies. This may be preferable to single currency peg for countries with diversified trade and investment partners as it better reflects the composition of their balance of payments flows. While the SDR may not be any country’s optimal basket (except perhaps for commodity exporters in a world where commodities were priced in SDR), it has the merit of pre-existing and being publicly quoted, which has both transparency and practical benefits. Moreover, to the extent that pegging to another currency implies importing its monetary policy, an SDR peg would allow a monetary policy that is representative of global conditions and less dependent on the policy stance of any single economy.

B. Asset Denomination

31. Critical elements. The development of a market for SDR-denominated assets would facilitate acceptance of a greater role of the official SDR as a reserve asset and unit of account, by enhancing liquidity and facilitating hedging. Significant efforts by the official sector would likely be necessary to provide initial impetus for market development. A set of policy commitments by the official sector would likely include: i) creating the market by announcing official sector commitment to issue and invest in SDR-denominated assets; and ii) support to market infrastructure development.

32. Role of sovereigns and IFIs. Aside from the Fund, other IFIs and sovereigns could play a critical role in supporting the development of a private market for SDR-denominated assets. Specifically, the commitment of a few highly-rated issuers to a regular program of issuance of SDR-denominated assets could kick-start market development. Crucially, it would provide the market with benchmark instruments to enable the pricing of other SDR-denominated securities that could be issued by the private sector. A public statement of interest in such assets by large reserve holders would mitigate the liquidity premium likely to be demanded initially by investors. Denominating existing reserve assets in SDR will not in itself create new reserve assets. Yet, it could contribute to a more orderly process of global asset diversification, and reduce the volatility associated with exchange rate movement in domestic currency. Although the basket composition may not be optimal for any given country, it is potentially a better hedge against exchange rate volatility than foreign debt in a single currency, and over time greater liquidity might be obtained by issuing in one denomination—the SDR—than in each of the underlying currencies to replicate the basket. Similarly, a number of IFIs periodically raise funds through market bond issuance, including the World Bank and regional development banks. A switch to denominating those bonds in SDR could provide impetus for market development. To reduce the need for hedging such borrowing, these institutions could also denominate part of their lending in SDR. SDR bonds could also be swapped back in usable currencies with the intermediation of private-sector banks, thus engaging the private sector and encouraging further market development.

33. Costs. SDR-denominated assets would operate in a shallow market at first and therefore would likely carry a liquidity premium. This is estimated initially at around 80-100 basis points, which could render it too costly for any individual country or IFI to take the first step and provide the impetus for an SDR-bond market, particularly in a context of fiscal consolidation pressures (see Annex 2 for details on the liquidity premium estimate). More generally, bundled goods are typically priced at a discount over their components if all are readily available, and the same might be expected of SDR-denominated instruments unless they included less readily available components (see next section). The experience of the ECU—where several countries and European institutions primed the market together— showed that markets developed quickly to the point where arbitrage pushed the interest rate on an ECU instrument close to the weighted average of the comparable national interest rates. But they also offered easy diversification at a time where bond markets were less globally developed, and in particular access to German government securities by non residents was constrained. A functioning interbank market and clearing mechanism also increased the liquidity of ECU instruments and ensured the appropriate pricing of ECU instruments relative to component currencies (see Box 2). A similar process might be expected as a market for SDR assets develops.

Development of the ECU Market

The market for ECU-denominated assets, including derivatives, developed rapidly with little official support beyond regular issuance of ECU-denominated debt by sovereigns and EC institutions. The private sector took the lead in developing the market, including market making and establishing a settlement mechanism. By 1990, the ECU ranked the sixth most used currency in terms of outstanding international bonds. At the outset, ECU-denominated assets were attractive substitutes to its component currencies and allowed market participants to circumvent some regulatory restrictions.

The European Currency Unit (ECU) was created as a basket currency in 1979 in connection with the European Monetary System (EMS), comprising nine European currencies and was later expanded to include the currencies of the twelve members of the European Community (EC). The ECU served as the basis for the EMS exchange rate mechanism (ERM) and for defining the parity grid among component currencies. The official ECU was created as a liability of the European Monetary Cooperation Fund (EMCF) by swapping such ECUs for 20 percent of the gold and gross U.S. dollar reserves held by central banks participating in the EMS. Official ECU was used only in transactions with EC central banks and a limited number of designated holders. Private ECU, on the other hand, was ECU-denominated liabilities of the banking sector. The proceeds from issuing obligations in the form ECU-denominated liabilities were payable in ECU-denominated deposit liabilities of designated banks. A number of European sovereigns, including France, Italy, and the U.K. issued ECU-denominated domestic bonds (Mehnert-Meland, 1994; Levich and Sommariva, 1989).

The ECU Banking Association created a clearing system centered around 45 ECU clearing banks, with the BIS acting as an agent of the private ECU clearing and settlement system. The use of the private ECU grew rapidly, including retail banking, especially in Italy and the U.K. By the first quarter of 1991, ECU-denominated bonds accounted for more than 5 percent of total international bonds outstanding and for over 15 percent of total secondary market turnover in the international bond market (McCauley, 2006). Although there was never an official mechanism to maintain the parity between the private and public ECU, until 1988 the value was fixed in terms of the official basket by a group of major European banks who stood ready to convert private ECUs into the basket at par. A growing concern in these institutions about their increasing net ECU exposure led them to discontinue fixing the private/public ECU exchange rate. Prior to the ERM crisis in 1992/93, the private ECU came to command a significant exchange rate premium over the official basket, reaching a peak of a 100 bp in 1991. The rules set by the BIS as a clearing house maintained the link between the conditions prevailing in the official ECU market and the term structure of interest rates charged in the private ECU market, thereby maintaining the relation between official/private ECU exchange rate as determined by an interest rate parity condition (Folkerts-Landau, 1992). Overtime, as the private ECU market expanded, European central banks started holding private ECU as part of reserves to use for market intervention.

Although the prospects for a common currency were initially limited, ECU-denominated assets offered the benefits of pre-packaged diversification for bond holders and smaller currency exposure to bond issuers at a time when currency swaps did not exist. Issuing ECU-denominated bonds also facilitated regulatory arbitrage and reflected a ‘search for yield’ on the presumption of the basket stability. The success of the private ECU was in part due to the restrictions on the use of the Deutsch Mark (DM) to denominate bonds by nonresidents; foreign banks were able to issue ECU-denominated assets as a proxy for issuing DM bonds, given its weight in the ECU basket. ECU-bond holders were able to obtain excess returns over DM-denominated assets while bearing little exchange rate risk, as the weights and cross-exchange rates were set within the official basket. In the aftermath of the ERM crisis in 1992/93 and the financial liberalization required under the single market, the ECU issuance declined and the stock of outstanding bonds shrank sharply from a peak of close to ECU 90 billion in 1992 to less than 60 billion in 1997. Issuance only recovered as the introduction of the euro approached and reflected the EC’s commitment to convert one ECU for one euro when the common currency was introduced (McCauley, 2006). The experience with the ECU market points to the link between public and private sector interaction to develop a market, as well as profit-making opportunities that encourage the private sector to take the lead in market-making.

34. Mitigating costs. To enhance initial market liquidity and reduce the premium faced by first movers, it may be useful to have a ‘group’ issuance where a number of countries issue jointly, thus expanding the volume issued, reducing fixed cost to individual issuers and the liquidity premium. Coordination should be aimed at establishing relatively quickly liquid benchmark instruments throughout the maturity spectrum. Another approach, possibly used in parallel, would be to issue SDR-denominated assets for private placement with interested sovereigns and institutional investors, as they are likely to hold those assets to maturity, thus may not require a high liquidity premium to invest in the newly-created assets (see Annex 2). This would mitigate the concern over lack of liquidity initially, though of course it would not enhance the liquidity of the SDR market overall. Another key part of the equation would be to ensure there is adequate demand for those assets (see below).

35. Potential investors. Some of the features of the SDR as a basket and the potential for diversifying credit risk could attract large reserve holders and long-term institutional investors. For most other classes of investors, demand for SDR-denominated assets would essentially have to follow the lead of public sector institutions, and early sensitization of these actors by the international official community would be critical. Sovereigns and large institutional investors would have to take on the role of developing the market and enhancing its liquidity. Potential investors could be classified as follows:

  • Large reserve holders and sovereign wealth funds could lead market demand with the objective of diversifying their reserve holdings (therefore focusing their interest on new SDR assets more than re-denomination), and a lesser interest in liquidity than other classes of investors. Other long-term and ‘buy to hold’ institutional investors could follow if they have appropriate investment mandates to diversify their portfolios into highly-rated SDR-denominated assets. Such investment mandates typically change with inertia however, and place a high premium on market depth.

  • Availability of liquid investment and hedging instruments to cash managers would encourage demand by large global corporations (assuming greater use of SDR in global trade). Demand by fund managers and retail investors are likely to be the last step in building demand for SDR assets. Fund managers would have an incentive to include SDR-denominated assets in their portfolios if those instruments were included in major bond indices, which are often passively tracked by fund managers. Retail investors could also be attracted by the features of SDR-denominated assets as a prepackaged diversification strategy.

36. Market infrastructure. A functioning market for SDR-denominated assets would require the public sector to provide liquidity support for SDR-denominated financial instruments and facilitate trading; frequent interest rate setting and benchmark pricing; and facilitate the establishment of a settlement and clearing system. Liquidity support could be provided through an “SDR window” or Repo facility to discount SDR-denominated assets and provide daily interest rate setting. A settlement and clearing system could be developed among private sector financial institutions who would act as market makers and clearing house, similar to the experience of the ECU clearing system. While these would need to be created from scratch, experience with the ECU suggests the process may be relatively fast, and need not involve significant public support beyond the initial phase (See Box 2 and Annex 2 for detailed technical background).

V. SDR Basket Composition

37. Objective. An SDR basket composition that maximizes the SDR’s attractiveness in all the roles discussed above is one that makes it, to the greatest extent possible, representative of global economic and financial weights, liquid, and simple to use and hedge. Unfortunately, there are trade-offs between these objectives. At a minimum, there will be a trade-off between the necessary evolution of the SDR composition (if it is to remain representative) and predictability of its future value; although a rules-based valuation process could go a long way toward addressing this problem. A work program on reviewing the SDR valuation methodology and considerations for expanding the basket is expected during 2011 (see IMF, 2010h).Without prejudging its conclusions, the remainder of this sections sets out a few considerations to bear in mind in the process.

38. Emerging market currencies. Including emerging market currencies in the basket as their shares in global trade and finance increase could enhance the attractiveness of the SDR by increasing its diversity and representativeness. Moreover, such process could facilitate a greater role for emerging markets in the IMS as well as support financial deepening in those markets. However going beyond a few additional currencies could mean adding currencies with low weight, increasing complexity and transaction costs for those tracking the SDR basket or needing to hedge their exposure. A side effect of such changes might therefore be that some central banks may not be willing to increase their SDR exposure. This would have implications for their willingness to participate in the voluntary trading arrangements that support SDR liquidity. If a more than a few of the smaller arrangements lapsed, the whole system could be undermined. On balance, adding a large number of emerging market currencies thus seems undesirable, but there may be a case for the ones with the largest weights in global trade and economic growth.

39. Not-fully convertible currencies. Under the Executive Board-approved SDR valuation method in place since 2000, all the currencies in the SDR basket must be “freely usable currencies,” as determined by the Fund (currently Dollar, Euro, Pound, and Yen). From the entry into effect of the Second Amendment, when the concept of freely usable currency was established, to 1982, the SDR basket included a number of currencies that not only were not on the list of freely usable currencies, but also had restrictions on both current and capital account transactions. Adding a not-fully convertible currency in the basket could actually increase demand by the private sector for SDR-denominated assets, as it would make it easier to acquire exposure to this currency than possible at present (with that advantage disappearing as the capital account becomes more open). At the same time, a nonconvertible currency in the SDR basket could aggravate some of the challenges facing greater use of the SDR. In particular, hedging costs and risk management would be much more complicated, as SDR holders/issuers may not be able to hedge fully against currency fluctuations of a nonconvertible currency. Similarly, central banks with a duty to put in place adequate risk management policies may find it difficult to hold an SDR that implies an exposure which cannot be managed owing to incomplete convertibility. Thus, prematurely adding not-fully convertible currencies would reduce the attractiveness of SDR, and indeed could lead the current SDR system to collapse (as noted above) but as illustrated by the example of the Chinese RMB (see below), it is not clear that currencies should await full convertibility to be allowed into the basket. A change in the SDR valuation criteria to allow the addition of non fully convertible currencies to the SDR basket (see also next paragraph) would not imply that they would also be added to the list of “freely usable currencies” as only currencies that meet the test for freely usable currencies under the Articles can be added to that list; the ability to use the currencies received in exchange for SDRs for balance of payments purposes and the SDR’s reserve asset character would therefore remain protected.

40. The RMB question. The recent review of the SDR valuation concluded that in spite of China’s prominent share of global exports, the RMB should not be included in the SDR basket, as it did not meet the criteria to be determined a freely usable currency ((http://www.imf.org/external/np/sec/pn/2010/pn10149.htm). It is an open question however whether this criterion should be retained as part of the SDR valuation method. Recent reforms that allow nonresidents, including central banks, to hold RMB-denominated deposits and the gradual development of RMB derivatives in Hong Kong could contribute, overtime, to resolving some of the technical difficulties in hedging RMB exposure. These could perhaps be supplemented by additional convertibility agreements between the PBC and other SDR designated holders. Such steps might be sufficient to preserve the SDR’s attractiveness with traditional reserve managers, particularly if explicitly part of a broader plan including a credible public commitment to internationalize use of the RMB and to liberalize capital flows. Other issues might need to be considered however, in particular the fact that the RMB’s value is tied to the dollar and managed by the authorities. Thus, adding the RMB to the SDR basket would de facto increase the weight of the U.S. dollar in it, and allow the exchange rate policy of one country to impact the value of the SDR in a discretionary way.

VI. Conclusion and Issues for Discussion

41. Risks to the stability of the IMS should neither be exaggerated nor readily dismissed. The SDR, if used more in its various guises—official composite reserve asset, unit of account, and possibly new class of reserve assets—could potentially contribute to the long-term stability of the system, alongside other reforms that have their limitations too. The paper presented a number of options to this end, many coming with question marks as to their effectiveness or feasibility (see Annex 1).

42. The options discussed above could be thought of as a menu whose elements can be experimented with incrementally, with some requiring a large degree of consensus in the international community that may take years or decades to build (e.g., most of those requiring amending the Articles of Agreement), but others implementable in a meaningful way by a subset of like-minded countries within a relatively short-time frame (e.g., use of SDR as a unit of account for some international trade and bond issuance). There is no obvious sequencing to consider between the different sets of options presented here. In principle, all can be pursued independently. However, it is clear that they are mutually supportive (for example, managing large SDR allocations would be much easier if there were liquid markets in SDR-denominated securities or if SDRs were used to invoice international trade). Further, unless a critical mass of steps is taken, the SDR is unlikely to make a visible difference in the functioning of the system. But over time their combination could result in the SDR playing a significant role in the system and enhancing its stability. This being a process measured in decades rather than years, a key issue is how to establish and retain momentum over time.

43. At this point, Directors’ views are sought on the following questions:

  • Do directors agree that enhancing the role of the SDR might contribute to the stability of the IMS over the medium to long term?

  • Among the options presented in the paper to expand the supply of official SDR, develop new SDR-denominated assets, and encourage use of SDR as a unit of account, which ones appear promising in the near term and should be considered further, including through public consultations and with potential SDR users in the private sector? Which ones might be considered for the longer run? And which should be discarded?