When the G20 pledges of $100 billion of SDR recycling materialize, there will no doubt be much self-congratulatory celebration. But the celebrations will be meaningless if the SDRs do not make their way to the low- and middle-income countries (LMICs) that need them. Will we be able to tell if the SDRs are hitting their mark? 
At this stage it seems most of the recycled SDRs will flow through the IMF’s Poverty Reduction and Growth Trust (PRGT) or Resiliency and Sustainability Trust (RST). As we have noted elsewhere, tracking SDRs is akin to tracking Bigfoot—you can see the prints but rarely see the animal. In this note we explain the extent to which we can track recycled SDRs from donors through the IMF and then to LMICs.
The bottom line (because the details can be tortuous)
With a great deal of effort, it’s possible to trace recycled SDRs—and as of today, not one recycled SDR has reached an LMIC! But, so the public don’t all have to turn into Bigfoot sleuths, the IMF could provide clearer and more timely reporting on the processes that transform donor pledges of support into loans to LMICs.
Two monthly tables would help.
The first table should clearly state:
- Donor pledges to the PRGT and RST by donor. As of now, all we have are aggregate numbers for the RST and various press reports that don’t add up.
- For each donor, pledges that have been transformed into agreed commitments and the amounts disbursed. As of now, these can be deduced from quarterly tables published by the IMF (with a three- to five-month lag) or by mining monthly online data, which takes considerable effort.
A second table should show the donor resources used to finance each PRGT and RST loan, so the public can discern how many recycled SDRs are making it to LMICs. This would make tracking SDRs and recycling much easier as now analysts have to guess which donor funds are financing which loans.
The data for these tables exist and various pieces are reported in different publications, often with a long lag. They should be made public in a format that does not require an expedition through the formidable dunes of data on the IMF website.
The tortuous details
What are the mechanics of loans under the two trusts and what are we trying to measure?
The PRGT and RST are built on loan agreements between the IMF and donor countries. The two parties agree that, on demand from the IMF, the donor will make funds available in a specified currency or SDRs for onlending to an LMIC under the terms of the loans associated with each trust.
No SDRs or other currencies are held in the loan accounts of the two trusts. The PRGT and RST consist of several accounts, with activities in the loan accounts being the most interesting for tracking. (Descriptions of the other accounts can be found here and here.) When an LMIC has agreed with the IMF for a loan from one of the trusts, the IMF contacts a donor or donors and asks that the appropriate amount of funds be made available. The donor lends the money to the trust. Then the trust immediately onlends the money to the LMIC. No money sits in the PRGT or RST loan account—the money just passes through the PRGT or RST loan account.
Thus, two measures of recycling of SDRs are appropriate:
- Commitments from donors, which are made in the signed agreements between the IMF and the donors. This would measure the increased “firepower” the IMF has at its disposal in the PRGT and RST.
- Disbursements to LMICs, which are made when a loan is made to an LMIC, with the IMF having activated its agreement with a donor. This would measure how the firepower is used to help LMICs.
The questions for those tracking the SDR allocation is: “How much does recycling increase the firepower of the IMF, and how are the recycled SDRs helping LMICs?”
What information does the IMF make public about PRGT/RST loans—and how can we track recycling?
On the donor side, the IMF reports the date and size of each loan agreement with donor countries and the undrawn balance on a quarterly basis in its Quarterly Report on IMF Finances. For the PRGT, the IMF catalogues the date of each loan arrangement with the donor, the amount promised, and the amount undrawn. If we assume all new donor commitments made after August 23, 2021—the date the new SDR allocation was made—are “recycled” SDRs, then we can track what portion of the pool of funds available for lending under the PRGT is recycled SDRs; that is, how much recycled SDRs have raised the firepower of the trust. But this information is made available with a three-to-five-month lag—as of late August, the report for end-April 2022 is not yet published.
But we also need to measure whether that increased firepower is being used, or how many recycled SDRs are being disbursed to LMICs. This is a bit more complicated. On the one hand, each donor agreement is tracked in the Quarterly Report on IMF Finances, including the undrawn balance, and thus by subtraction the drawn balance. For all the agreements that came into force after August 23, 2021, we could sum the drawn balances and count those as disbursed recycled SDRs. However, if there are undrawn balances from agreements that came into force before August 23, 2021, then the recycled SDRs are not adding anything to the overall disbursements to LMICs. So, an alternative measure of the effective firepower resulting from recycling would be how much total disbursement exceed the disbursements that would have been possible without recycling, that is total donor commitments prior to August 23, 2021. Table 1 should make this distinction clearer.
For the RST, since all commitments and disbursements will take place after the August 23, 2021, SDR allocation, we can assume that all loan agreements with donors are supported by recycled SDRs or the foreign exchange equivalent. If the IMF reports on RST loan agreements with donors as it does with the PRGT, that will give us a measure of the commitment of recycling under the RST. And all RST lending to LMICs can be counted towards disbursements of recycled SDRs reaching LMICs. In addition, when donors make loan commitments to the RST, they are requested to make an upfront deposit equal to 20 percent of the commitment. These deposits are invested by the IMF to build reserves for the RST, thereby lowering risks for lenders to the RST. These deposits should also be counted as recycled SDRs.
Loan disbursements to LMICs under PRGT-supported arrangements and the RST are made in installments—often over a three-year period. Trust resources for future installments will be reserved and thus unusable for other LMICs. These reserved funds are not allocated to specific donors. In its weekly report on finances, the IMF indicates these amounts, aggregated across donors.
Also, the IMF will always hold back from tapping some donor commitments to preserve a buffer for future unforeseen lending and to allow the PRGT to meet possible encashment calls from lenders. Moreover, some commitments from donors specify the lending instrument that they will support, further limiting the IMF’s choice of donor to support a loan disbursement to an LMIC.
What information would help us track SDR recycling?
These complex arrangements will make it difficult to track recycled SDRs, but Table 1 lays out a hypothetical presentation that would give a good idea of commitments and disbursements of recycled SDRs.
Table 1. Hypothetical commitments and disbursements of recycled SDRs
In this table, donors have made PRGT commitments totaling SDR 21 billion, of which SDR 12 billion was made before August 23, 2021, and SDR 9 billion after. Thus, the firepower of the PRGT (commitments) has been increased by SDR 9 billion recycled SDRs.
Of those PRGT commitments, SDR 12.8 billion has been disbursed, of which SDR 8.1 billion was mobilized using the pre-August-2021 agreements with donors, and SDR 4.7 was mobilized using the post-August-2021 agreements. So how many recycled SDRs have been received by LMICs? There are two possible answers:
- SDR 4.7 billion was disbursed drawing on the post-August 2021 donor agreements, so all those SDRs could count as recycled.
- But, before August 23, 2021, the total firepower of the PRGT was SDR 12 billion and SDR 12.8 billion have been lent, so only SDR 0.8 billion was lending that could not have been made absent the recycling commitments.
The first answer is plausible if the IMF’s choice of loan agreement to draw upon was somehow constrained—say by currency preference or maturity—and pre-August-2021 agreements were not useable. The IMF does not publicly explain its choice of which donor loan agreement to use, so it will be hard to know to what extent loan agreement choice is constrained.
A more precise number, somewhere between SDR 0.8 billion and SDR 4.7 billion, could be determined if the IMF published an accounting of how each donor agreement was used, linking donor disbursements to specific loans to LMICs. (Or one could measure receipts of loans on the LMIC side and when those receipts exceeded the amount available from pre-August 2021 agreements, the excess could be attributed to recycling. This is future work, as matching accounting from the donor and recipient side is very difficult absent an explicit link by the IMF.)
For the RST the situation is a bit more straightforward, under the assumption that all contributions are from recycled SDRs. In table 1, donors have committed SDR 3.25 billion to the RST and SDR 0.7 billion have been disbursed. In addition, donors will have contributed SDR 650 million (20 percent of SDR 3.25 billion) to the RST Deposit Account. So, RST-related recycling would amount to SDR 1.35 billion in this hypothetical example.
Overall, the table indicates that recycled SDR commitments in both trusts would be SDR 25.1 billion and disbursements between SDR 1.5 billion and SDR 5.4 billion.
Where do we stand now?
To understand where we stand now, we first need to look at the state of PRGT financing before August 23, 2021—the date of the SDR allocation. The data for July 31, 2021, are in table 2.
The Weekly Financial Report for July 31, 2021, indicates that SDR 4.4 billion of the undrawn donor resources have been committed to LICs through loans from the PRGT, but are as yet undrawn.So, donor resources available for new loans on that date were SDR 15.9 billion.
According to the latest data, some progress has been made in recycling SDRs through the IMF. A summary of the change in PRGT lending capacity and activity is found in table 3. Available PRGT loan resources have increased by SDR 3.8 billion. Lending has increased by SDR 1.4 billion, while donor resources available for new loans increased by SDR 2.4 billion.
Table 2. PRGT resources before August 23
Table 3. Changes in PRGT lending capacity and activity
In Billions of SDRs
In the past year, how many SDRs have been recycled? From a commitments point of view, the SDR 3.8 billion increase in total loan resources could be said to be recycled. But a look at the available details of the donor commitments to the PRGT in table 4, drawn from the IMF Financial Data Query Tool, indicates a larger number. Four countries have increased their commitments in the last year: Canada (SDR 0.5 billion), Italy (SDR 1.0 billion), Japan (SDR 2.8 billion), and Spain (SDR 0.35 billion), for a total increase of SDR 4.65 billion, a larger number by SDR 0.85 billion than that inferred from table 3. It is difficult to reconcile the difference in these numbers, because the IMF does not release the details underlying table 3, in particular the moment at which new loans are included and fully utilized loans are excluded from the accounting. A loan-by-loan accounting would be useful in this regard.
In addition to the donors that have signed agreements with the IMF, many donors have pledged to recycle SDRs through the PRGT. These were catalogued in a paper on the Review of the Adequacy of PRGT Resources in April 2022. We include these unrealized pledges, amounting to another SDR 4.75 billion, in the last column of table 4.
Table 4. PRGT resources before and after August 23rd, 2021
There are also other difficulties in putting a hard number (rather than a range) on the amount of recycled SDRs, as illustrated by the Japan PRGT comment. In a public statement, the Japanese minister of finance commits to SDR 2.8 billion in recycling to the PRGT, of which SDR 1.8 billion had already been pledged. Thus, the IMF records SDR 1.0 billion as post-August-2021 pledges. However, the size of the existing loan agreement with Japan was increased by SDR 2.8 billion in January 2022, thus one could say the full amount was recycling. This example indicates the inherent inability to distinguish “old” SDRs from “new” SDRs and what counts as recycling and what does not. Furthermore, once new drawings are made on the Japanese agreement, it will be impossible to distinguish whether recycled or unrecycled SDRs are being drawn.
No recycled SDRs have been disbursed. Total lending commitments from the PRGT at end-July 2022 amount to SDR 19.8 billion and of that, SDR 15.3 billion have been disbursed, all from agreements made before August 2021. As best we can discern from the online data, the four agreements made with donors post August 23, 2021, have not yet been tapped.
No agreements have been signed to mobilize SDRs for the RST, at least according to publicly available data. While the IMF has announced pledges of SDR 37 billion they have nowhere detailed the donor pledges, although in a IMF press conference the following potential contributors were named: China, Japan, South Korea, Italy, France, Germany, Saudi Arabia, the Netherlands, Canada, the UK, Switzerland, and Spain. No amounts were associated with any potential lender, but later press reports have indicated that at least Spain has pledged SDRs 1.9 billion. No agreements with the IMF have yet been signed for the RST and no disbursements have been made.
Some suggestions for more clarity in IMF reporting
Summary information on PRGT and RST finances should be consolidated in one report, published monthly. Now there are four sources of data: periodic reports on financing of the PRGT; the quarterly report on IMF finances (published with a three-to-five-month lag); a summary weekly financial report (found here under “Financial Activities”); and the IMF financial data query tool, which gives current information. A summary table, like table 1, could be published monthly without much difficulty with loans by donors.
The IMF should be clearer about donor pledges, donor commitments (signed agreements), and disbursements. While we recognize that there is a necessary process for donors and the IMF to move from pledges to disbursements, there is no easy way to track the progress of these processes. As it now stands, the IMF and donors can claim success on recycling SDRs (the red numbers in table 2, for example) before such pledges have become signed agreements and when no money has made its way to LMICs. At the same time, both the IMF and donors could take credit for commitments realized. All four of the signed donor commitments have flown under the public radar. Transparency would quiet critics (like us) who disparage the apparently slow progress that has taken place.
The IMF should consider publishing how the money from each donor agreement is used. Every disbursement of money from the PRGT is in fact two loans: one from the donor to the PRGT and the second from the PRGT to the LMIC. These transactions are nowhere linked, at least in public IMF data. In some instances, they can be inferred. Making the link explicit would add to the transparent tracking of how donor resources are being used and what “credit” donors can take for recycling SDRs.
 For its contribution, Spain augmented an existing loan agreement, while Canada and Italy had new loan agreements. All PRGT disbursements relying on the pre-existing Spanish agreement will be counted as recycled disbursements as virtually all previous SDRs had been drawn.
For those unfamiliar with SDR recycling, this paper might be helpful.
We combine data from the January 2021 Quarterly Report on IMF Finances and end-July data from the IMF Financial Data Query tool and the Weekly Financial report, the using data from end-July 2022 to generate tables 2 to 4.
Some loan resources have been committed but not yet disbursed. As noted above, disbursements to LMICs are made over time and thus some donor resources are, in effect, committed for these future payments. However, these future disbursements are not allocated to specific lenders when commitments are made and thus only an aggregate figure of undrawn lending commitments is available.
We drew the data in the table from the IMF’s Financial Data Query Tool. The summary table published in the Weekly Financial Report shows slightly different numbers for Total Donor Loan Commitments, which we cannot reconcile. But the donor resource availability is the same.
Rights & Permissions
You may use and disseminate CGD’s publications under these conditions.