The reallocation of special drawing rights (SDRs) to the IMF’s Poverty Reduction and Growth Trust (PRGT) is a well-established channel for using SDRs allocated to advanced countries to support low-income countries (LICs). It is again likely to be a preferred option for re-channeling SDRs.
The modalities used in the PRGT, which meet the call by the G20 for reallocating SDRs in ways that preserve their reserve asset characteristics, could also be used for a new trust in the IMF that we discussed in an earlier note in this series. This new trust, which we dubbed the Global Resilience Trust (GRT), could support vulnerable middle-income countries (MICs) as well as LICs as they recover from the pandemic and seek to transition to a more resilient, sustainable, and equitable future.
In this note we consider the technical challenges of channeling SDRs to institutions other than the IMF that have already been approved as ‘prescribed holders’ of SDRs. This innovation, operating in addition to the options described above, would have the potential to significantly increase the volume of SDRs channeled to support LICs and MICs. A further strong rationale would be to allow multilateral development banks (MDBs) to establish lending windows that could operate alongside the proposed GRT as part of a concerted multilateral effort.
Basic design elements
Contributions of SDRs would be used to support the regular operations of an MDB or other prescribed holder (such as the Latin America Reserve Fund, FLAR) or to allow the recipient institution to establish a new special purpose fund akin to the proposed GRT. There are two possible and non-mutually exclusive approaches:
SDRs lent to a prescribed holder could simply be on-lent to LICs and MICs. This approach could broadly mimic the technical arrangements used for on-lending SDRs to LICs by the IMF’s PRGT and could in principle be initiated with minimal delay. 
Capital contributions in SDRs could be made to a prescribed holder. This approach would provide greater flexibility than simply on-lending SDRs and, via leverage, could allow a MDB to scale up its lending capacity by at least 3-4 times the increase in capital. But as we discuss in detail below this approach entails technical challenges for both the providers and recipients of SDRs on which further work is needed.
Under both possible approaches the boards of the recipient financial organizations would determine the purpose, terms, country coverage, and conditionality framework for new lending supported by contributions of SDRs. In principle this could also include MDB on-lending to regional or national development banks; a key consideration would be ensuring that liquidity and credit risks for SDR contributors are adequately addressed.
1. Would the envisaged support for MDBs maintain the reserve asset nature of SDRs?
The reserve asset status of SDRs lent for direct on-lending (the first approach), would need to rely on an encashment regime and the ability of the recipient organization to mitigate credit risks:
An encashment regime based on the long experience of channeling SDRs to the PRGT could maintain the liquidity of SDR claims. In the PRGT, roughly 20 percent of available loans from creditors is held in reserve to meet encashment calls by PRGT creditors if they have a balance of payments need. To limit the scope for unnecessary withdrawals from the PRGT, the IMF can adjudicate if the balance of payments need of the encashing country is questioned. The design of a similar encashment regime for lending SDRs to an MDBs would need to to specify which institution would adjudicate if an encashing country’s balance of payments need were to be questioned.
In the PRGT, credit risks to lenders are mitigated by lending policies and a reserve account. Depending on the adequacy of existing buffers, hard currency resources would likely be required by the recipient MDB to fund a similar reserve account that could ensure repayment to lenders in the event of delayed repayments by borrowers.
Creditor countries would be paid the SDR interest rate on amounts lent to the MDB. If the MDB wanted to make subsidized loans to all or some of its borrowing countries it would also need establish a subsidy account, also funded with hard currency resources.
SDRs could be donated, lent, or pledged to provide capital (the second approach). Each one of these modalities poses challenges to maintaining the reserve asset status of the SDRs contributed:
SDRs donated to a prescribed holder would lose their reserve asset status from the perspective of the donating country. If these donated SDRs are then intended to be held indefinitely to strengthen the recipient’s capital base—becoming essentially illiquid and able to absorb losses—it is not clear that they could continue to be considered reserve assets from the perspective of the recipient organization.
SDRs lent to a prescribed holder would not immediately lose their reserve asset status from the perspective of the country providing the SDRs. An encashment regime could be considered to provide liquidity to the SDRs lent to strengthen the capital base, but there is then a question of whether they would be considered as capital (see below). In addition, there remains a tension between the purpose of the loans—namely to provide a capital buffer that would allow the organization to take on some additional credit risks in leveraging greater lending—and the need to preserve reserve asset status which would require the near-absence of credit risk for the lent SDRs.
SDRs pledged to a prescribed holder to strengthen the organization’s capital base could remain on the balance sheet of the pledging country. As such, in the first instance, the pledged SDRs could perhaps be seen as maintaining their reserve asset status and there would be no loss of interest on these SDR holdings. But given the possibility that the pledges could be called, the same tension noted above between the purpose of the contribution and the avoidance of credit risks also arises.
As noted above, further work is needed by the IMF, the MDBs and contributing countries, to clarify and, if possible, resolve these issues.
2. Does it meet legal constraints?
From the perspective of the IMF’s rules governing the use of SDRs
Although the envisaged recipients of SDRs have already been approved as prescribed holders by the IMF their role would expand significantly. These financial institutions can receive SDRs in transactions with SDR participants (currently all IMF member countries) but typically their holdings of SDRs have been a very small share of the total cumulative allocations. Their role in SDR operations would expand beyond providing convenience to member countries, which was the rationale for their being approved as prescribed holders. Questions could possibly arise as to whether this expanded role was consistent with approvals so far granted.
Similarly, questions could arise as to whether using SDRs to provide capital was covered by existing approvals. In 1979-80 the IMF approved several types of “prescribed operations” between SDR participants and prescribed holders that went beyond simple spot exchanges for freely usable currencies. These included settling financial obligations to another participant, donations, loans, swaps, forward operations, and the use of SDRs to secure the performance of a financial obligation to another participant or prescribed holder. The latter includes pledges in which the SDRs would remain as part of the holdings of the pledgor. However, these approved operations do not explicitly include exchanging SDRs (in the form of donations, loans, or pledges) for equity stakes. Given this ambiguity and the requirement that parties to these transactions confirm to the IMF that their intended use of SDRs is in accordance with the relevant prescribed operation, it would be important to confirm that such exchanges are allowed. If an explicit authorization is required to exchange SDRs for equity, this would require a 70 percent IMF Executive Board majority and all prescribed operations must also be consistent with the “proper use of SDRs.”
From the perspective of the contributing countries
Every creditor country has its own set of rules governing the use of SDRs.,  Most countries view SDRs as reserve assets, and as explained above, retaining this status when lending them to an MDB or another prescribed holder such as FLAR for direct on-lending would require risk mitigation to ensure the assets are of reserve quality, and some ability to liquidate the loan and get the SDRs back if needed. The creditor country’s rules may also entail a maximum length of the loan to recipient countries. For the PRGT, the loan term is 10 years, but longer terms might be needed to allow lending financed with SDR loans to be more consistent with the terms typically offered by MDBs to meet longer-term objectives. Whether this is possible would be determined by the willingness of advanced countries to go along with longer periods and their perception of the associated risk, which in turn would depend on the conditions of lending.
These same considerations on the use on SDRs arise for possible capital contributions. However, unlike direct on-lending of SDRs, their use in capital contributions has not been tested against individual countries’ rules governing SDR usage. There are therefore considerable uncertainties as to what is possible and further work is needed. However, since it may not be feasible to meet all the constraints at once, it is also important for contributors to determine whether they would be prepared to accept some liquidity or credit risks and thus some loss of reserve serve asset status on their SDR contributions. In this vein, it may be possible to develop arrangements at a national level to support the liquidity of claims. For example, depending on the contributing country’s laws and regulations, a government could consider providing a liquidity guarantee to its national central bank that would cover SDRs lent to an MBD; this would mimic the current practice in some countries where central bank lending to the PRGT requires a corresponding credit guarantee from the government. Alternatively, the option remains for countries to make use of the additional reserve buffer provided by the SDR allocation to provide contributions in the form of non-SDR assets and thereby avoid the additional complications that may arise when using SDRs.
From the perspective of an MDB or other prescribed holders
Each recipient organization would need to determine, in conformity with the institutions’ own rules and governance structures, the purpose, terms, country coverage, conditionality framework, and other risk mitigation processes for new lending supported by contributions of SDRs.
Since on-lending SDR loan resources outside the IMF would be a new departure, the feasibility of this approach would need be carefully assessed by each recipient institution. An important consideration would be the scope for lending in SDRs. On operational and possibly legal grounds it may be necessary to conduct loan transactions in dollars or other currencies rather than SDRs and therefore establish standard procedures for exchanging SDRs lent to the institution. Since such exchanges of SDRs are likely to involve some of the same advanced countries that would be providing SDRs, this would again raise the issue of whether contributions could simply be made in non-SDR assets.
The possible provision of capital in the form of SDRs raises questions which would need to be addressed at the level of each recipient. Key among these is whether the MDB’s accounting and legal structure would allow SDR loans or pledges to be considered capital, possibly some form of tier-two capital, and whether this would meet the requirements of rating agencies so that the capital injection can achieved a desired degree of leverage. There are likely to be significant differences in this area across MDBs. Moreover, reaching a solution that best meets the requirements of the MDB and possible SDR contributors is likely to require joint consideration of both parties’ specific constraints and possible tradeoffs as noted above.
3. Does it ensure that the use of the funds is transparent?
The governance structures of designated holders are well established but additional safeguards may be needed particularly if a new special purpose fund is created. For example, additional monitoring and reporting would be needed to provide confidence that a new fund aimed at supporting adaptation to climate change is delivering on key policy objectives. This would be important for the countries providing financial backing through SDRs and to support collaboration in meeting multilateral objectives.
4. Does it address priority policy areas?
There would be greater flexibility in establishing the purpose and policy focus for a fund at an MDB than if SDRs were lent to the IMF (as was discussed in the first note in this series on an IMF Global Resilience Trust) where operations require some link to balance of payments needs. The receiving institution would need to establish, in consultation with perspective contributors and users, the specific purpose and policy for the SDR-supported lending, either reinforcing its existing operations or designing new ones. Each institution would have its own constraints in doing so.
5. Does it cover a targeted group of countries?
Each institution would need to establish eligibility criteria for support from the SDR-supported fund, subject to its own internal rules and processes. This may in some cases entail constraints on country coverage. For example, the coverage of a new IDA-based lending operation facilitated by an infusion of SDR resources would need to be assessed taking account of the possible implications of the IDA allocation process which could constrain total lending to some countries. If the operations to be supported by MDBs are intended to integrate with the operations of the proposed GRT for the IMF, it would also be important to take account of differences in country coverage across institutions.
The range of countries that could be targeted would also in part depend on the choice of institutions to be supported. Of the current designated holders, only the International Fund for Agricultural Development (IFAD) and the two constituent parts of the World Bank—the International Development Association (IDA) and the International Bank for Reconstruction and Development (IBRD)— have global coverage of developing countries. FLAR, for example, could provide much needed liquidity support to MICs that would not benefit from an expansion of IMF support under the PRGT, but this would be limited to its Latin American membership. If the objective was to support a regionally diverse group of countries through expanding the operations of MDBs in, for example, working alongside the proposed IMF Global Resilience Trust it may be necessary to approve the IDB and the European Bank for Reconstruction and Development (EBRD) as prescribed holders.
6. Does it link well with/enhance other aspects of international cooperation architecture?
The degree to which the new lending integrates with other aspects of international cooperation will depend on the ultimate use of the SDRs. Simply augmenting existing facilities would deepen the financial reach of those facilities. For example, strengthening IDA would be an important contribution to the international architecture and FLAR has a long track record of providing liquidity support in anticipation of, or coordination with, IMF financial support in the region.
If new activities were to be financed using SDRs, these should be seen in the context of a related overall global financing plan. For example, if SDRs were to be used to establish a green investment or resilience fund at the World Bank and other MDBs, the effectiveness of the fund would be enhanced if its use was synchronized with operations of a possible Global Resilience Fund at the IMF, country pledges under the UN Framework Convention on Climate Change (UNFCCC) structure for nationally determined contributions (NDCs) for decarbonization and national adaptation plans (NAPs) and other international commitments.
7. Does it leverage/catalyze other resources?
As a matter of policy, the operations funded using SDRs should provide additional resources rather than substituting for existing sources of funding. Direct on-lending of SDRs would almost certainly provide additional lending resources. However, concerns have been expressed that to this lending could be scored as Official Development Assistance (ODA) so that some countries might opt to substitute this lending for other elements of their aid budgets. In the case of capital contributions, similar concerns have been expressed that the use of SDRs for this purpose might substitute for other capital contributions. Thus, for example, if part of an IDA replenishment were met using SDRs, this may not represent mobilization of additional resources and similar concerns may arise in relation to capital contributions to other MDBs. However, aside from this caveat, if SDR-based capital contributions were possible, the resulting leverage to secure private financing would directly catalyze additional resources.
The experience and repute of MDBs, including expertise essential for the design and implementation of, for example, policies to build resilience to climate change, would be important in catalyzing additional resources, including resources for project finance. Given the large scale of the total financing needs for a transition to a green and equitable economic structure, efforts to catalyze additional resources would need to be coordinated with other multilateral bodies. Country operations may also need to be accompanied by IMF macroeconomic policy signaling to catalyze broader support.
8. Does it ensure funds will be made available and used?
Availability of funds. Channeling SDRs to institutions other than the IMF would break new ground. In addition to the technical challenges that would need to be confronted to satisfy contributors’ requirements, including (in the case of loans) interest compensation, and more generally liquidity and credit mitigation, the objectives and design of the financed operations would also need to be sufficiently attractive to potential contributors of SDRs. In this vein, there may also be some reluctance on the part of potential contributors to participate at the outset. To build confidence in this innovation, it may be necessary to start operations on a smaller scale than ultimately desired and, as experience is gained, broaden the initiative to encompass more creditor participants and perhaps additional MDBs.
Frequency of use. Given the nature of the technical obstacles to be overcome to facilitate SDR contributions to MDBs, success in resolving these issues would best be seen as a precursor to longstanding arrangements to channel SDRs outside the IMF. The timescale of economic challenges to be faced in transitioning to a green and resilient future also suggest that facilities to be set up with SDR contributions should have a long time horizon which anticipates repeated use, as is now common for LICs borrowing from the PRGT.
Lack of stigma. Reluctance to borrow from the IMF because of the associated stigma has been a longstanding concern for the institution. In contrast, countries are generally less averse to using the funds of the designated holders such as MDBs or FLAR than they would be to going to the IMF. Nonetheless, while conditionality will be needed to provide assurances to potential creditors, it would important not to burden operations with excessive conditionality that could deter country participation
Financing operations outside the IMF with SDRs would be an important innovation, having the potential to significantly increase the volume of SDRs channeled to support LICs and MICs. But there are significant obstacles that both providers and recipients of SDRs would have to confront.
For most institutions, the more tractable way forward would be to lend SDRs that are simply on-lent by prescribed holders such as MBDs or FLAR. Financing of the PRGT with SDR loans provides a well-tested model but replicating this to provide comparable assurances of liquidity and risk mitigation to creditors will require some further work to allow operations to begin promptly.
Providing capital contributions to MDBs would be of more value, because of the leverage of private finance that this would bring. But the technical challenges for MDBs and potential contributors are also greater. At the very least it will take some time to better understand the possible solutions for a variety of these institutions.
Even if technical challenges can be overcome, sufficient confidence in newly created operations to unlock large SDR support is only likely to emerge over time. There may therefore be a need to start operations at a more modest scale and build these up as experience grows.
 African Development Bank, African Development Fund, Arab Monetary Fund, Asian Development Bank, Bank for International Settlements, Bank of Central African States, Central Bank of West African States, Eastern Caribbean Central Bank, European Central Bank, International Bank for Reconstruction and Development, International Development Association, International Fund for Agricultural Development, Islamic Development Bank, Latin American Reserve Fund, and Nordic Investment Bank
 For simplicity we use the term MDB to mean an MDB that is a prescribed holder of SDRs since most MDBs have been approved as prescribed holders. The Inter-American Development Bank (IDB) and the European Bank for Reconstruction and Development (EBRD) are notable exceptions.
 An alternative approach that has been suggested would be for a new trust in the IMF to on-lend to MDBs. To date, the IMF has only lent to countries and it not clear that its Articles of Agreement would support lending to an MDB. And unless the trust were to take on the credit and liquidity risks for these on-lent SDRs—which seem does not seem practical—the challenge of ensuring the reserve asset status of the on-lent SDRs would remain.
 The liquidity of some loans to the PRGT is also supported by the flexibility to trade them with other IMF members, but only countries whose external positions are judged to be strong enough by the IMF for them to be drawn upon to finance lending operations under the IMF’s non-concessional General Resources Account
 PRGT credit risks are also mitigated by the IMF’s ‘preferred creditor status’ under which official bilateral and private creditors have broadly accepted, for the most part as a policy rather than a legal issue, that their claims are subordinate to the those of the IMF. In general, MDB claims and those of FLAR benefit from a similar status which is also important in supporting these institutions’ access to private markets. A movement away from this norm of preferred credit status for lending supported by SDR contributions would complicate the task of adequately mitigating credit risks.
 Donations pose other challenges for contributors. Since a donation of SDRs removes an asset (SDR holdings) from a country’s balance sheet while retaining a liability (the SDR allocation which cannot be transferred) it is likely to require some form of budgetary approval. The latter would also need to cover the continuing fiscal cost of interest paid on the allocation that would, after a donation, not be matched by the interest earned SDR holdings.
 Exchanging SDRs for an equity stake in a Green Fund was proposed over a decade ago but the idea has not been put to the test. See ‘Financing the Response to Climate Change’, Bredenkamp and Pattillo. While SDRs are routinely lent to the PRGT, they have not so far been lent to other countries or entities and there have been no donations or pledges of SDRs.
 The ONE Campaign has compiled country-by-country rules governing SDRs for most G20 countries. The compendium is available upon request.
 For the national central banks (NCBs) of the Eurosystem there is an additional constraint. Article 123 of the Treaty on the Functioning of the European Union prohibits the provision of monetary financing to public authorities. However, this prohibition is subject to certain exceptions. In particular, NCB lending to the IMF has been permitted on the grounds that such lending “results in foreign claims which have all the characteristics of reserve assets.” Lending SDRs to institutions other than the IMF would presumably need to be justified on the same basis. (See the Opinion of the European Central Bank of December 28, 2020, on NCB participation in IMF borrowing arrangements)
 Approval of prescribed holders requires an 85 percent majority in the IMF Executive Board. This is a high bar but if the approval is intended merely to allow additional MDBs to perform functions considered appropriate for MDBs that are already prescribed holders it should not be too difficult a hurdle.
Rights & Permissions
You may use and disseminate CGD’s publications under these conditions.