An SDR Liquidity Line for MDBs

A $50 billion liquidity line from central banks to multilateral development banks, backed by SDRs, could enable billions in additional development lending. 

In the wake of the pandemic, world leaders agreed to a US$650bn allocation of Special Drawing Rights (SDRs) to bolster global liquidity. Countries across the world benefitted, but the lion-share of SDRs are allocated to larger economies in line with their IMF quotas. The G20 and other advanced economies received around $500bn. 

The G20 agreed to rechannel $100bn of SDRs to developing countries. But to date, recycling mechanisms remain limited, comprising principally of the IMF’s Resilience and Sustainability (RST) and Poverty Reduction and Growth (PRGT) trusts. More recently, the African Development Bank and the Inter-American Development Bank proposed the use of SDRs for multilateral development bank (MDB) hybrid capital and a modest amount of SDRs may be deployed for that purpose.  But a substantial amount of SDRs remain unused on advanced and large emerging economy central bank balance sheets.  

At the same time, the G20 has called on the MDBs to do more to help developing countries attain the UN’s Sustainable Development Goals (SDGs). MDBs are improving the efficiency of their own balance sheets to lend more and pursuing policies to mobilize more financing from third parties.  In a new CGD paper, I propose a new liquidity line backed by non-utilized SDRs on leading central bank balance sheets to complement these reforms.  

MDBs hold substantial amounts of liquidity. The International Bank for Reconstruction and Development (IBRD), an arm of the World Bank, and the four main regional development banks—the African Development Bank (AfDB), the Asian Development Bank (ADB), the European Bank for Reconstruction and Development (EBRD), and the Inter-American Development Bank (IDB)—together hold more than $200bn of high-credit-quality liquid assets, principally from advanced economies. That sum is equal to about 34 percent of their debt outstanding and 36 percent of their development assets. In general, MDB internal rules require enough liquidity to cover 12 months of estimated cash-flow needs, but the banks typically hold considerably more than that amount.  

An SDR line could reduce MDB costs, boost capital and allow increased leverage 

Holding such large amounts of liquidity over many years does not come for free, and a new liquidity line for MDBs could reduce costs. The savings could then be used to finance grants, knowledge, technical assistance, or concessional lending, or be plowed back into capital to then be leveraged to finance increased future lending.  

Rating agency methodologies also suggest a tradeoff between liquidity and capital. Depending on each MDB’s starting point, holding greater amounts of liquidity could allow an MDB to boost lending, and maintain a somewhat lower capital ratio. Considering Standard and Poor’s (S&P) quantitative assessments, a tradeoff exists for the IBRD, the IDB, and the ADB. In other words, a new liquidity line could boost the assessment of liquidity and allow for greater lending and a reduced capital ratio, while maintaining these MDBs’ AAA ratings. 

To reduce the costs of holding liquidity, MDBs invest in a range of (relatively high quality) assets in many currencies to seek higher returns, taking price and credit risks. MDBs also use derivatives to hedge currency and interest rate risks related to their investment portfolios, introducing counterparty risks. These price, credit, and counterparty risks use up capital. A new liquidity line to substitute a portion of the existing liquidity portfolio could then release capital which could be leveraged for higher development lending.  

How it would work 

How could the liquidity line be structured? The line could be a contingent repo (repurchase) facility. Repo contracts are standard in financial markets and the repo market amounts to around $4.6 trillion in the United States alone.  Each MDB would have contracts with several participating central banks, and the facility could be managed through a central agent. This would be a step toward an integrated MDB system, a recommendation that has been made by various G20-sponsored committees. The IBRD or the Bank of International Settlements (BIS) could act in this role. The specific contracts could be refined depending on the precise preferences of each of the participants.  

The facility would be contingent, so unless the line was drawn down, the SDRs would remain on the central bank balance sheets. An escape clause could be added, allowing an individual central bank to withdraw from the facility in the event of a balance-of-payments crisis. As several central banks would have contracts with each MDB, the risk that more than one participating central bank would suffer a balance-of-payments crisis at the same time would be very low. 

Given this structure, the SDRs should continue to count as reserve assets and so be counted in central bank international reserves. Furthermore, the SDRs would not be used for financing (MDBs would continue to finance loans from their capital and debt), which should allay fears that central banks are providing any “monetary financing” and the contingent arrangement should not be inflationary nor threaten economic nor financial stability.1 Still, in the end, the decision as to what constitutes a reserve asset and what constitutes monetary financing is up to each central bank and subject to its particular legal operating framework.  The MDBs would pay a small commitment fee. This would provide participating central banks with a modest income on a relatively small portion of their non-utilized SDRs. 

In the very unlikely event that the line were triggered, the central banks would transfer SDRs to the MDBs. The MDBs could retain the SDRs as liquidity, swap them for hard currency in the IMF voluntary market, or potentially disburse them to member countries that are able to receive SDRs. MDBs would provide collateral to the central banks which could consist of existing liquid securities or synthetic sovereign loans subject to a credit quality threshold.2 The MDBs would pay an implicit interest rate on the repo transaction, in excess of the usual SDR interest rate. The higher rate would reduce the incentive to trigger the line except under truly turbulent market conditions. 

In the absence of such a liquidity line, if MDBs did face a liquidity crunch, they would be forced to sell their considerable holdings of financial assets, exacerbating the market turmoil that prompted the crunch. MDBs are already significant players in global financial markets. The liquidity line would integrate them into the global financial architecture in a way that promotes global financial stability. 

The MDBs would be required to return the SDRs to the central banks within a pre-agreed timeframe, perhaps 6 months. The line itself would have a maturity of 3–5 years and the contract could include a provision to roll over every 6 months unless there were objections from either party, to underline the idea that this would be a permanent feature of the international financial architecture. 

The scale of costs savings from the SDR liquidity line 

Why is holding liquidity costly? Considering the decade before the pandemic, the spread between a 3-month US Treasury bill and a 10-year US bond was on average 1.83 percent. Borrowing at 5 years and investing at 3 months, the spread was about 1.07 percent. This is then the cost of issuing debt at a 5- or 10-year maturity and holding a short-term asset, abstracting from any credit spread. These are then the purest estimates of the cost of holding liquidity in normal times. But we do not live in normal times: Thanks to the global financial crisis, followed by the pandemic, central banks have injected trillions of dollars of liquidity into financial markets. More recently, the Russian invasion of Ukraine sparked an inflationary surge which proved more persistent than anticipated, prompting central banks to hike short-term interest rates and inverting the yield curve. Still, short-term rates are now expected to fall, and central banks are selling long maturity assets from their balance sheets. Most analysts predict markets will normalize and the usual upward sloping yield curve will return. 

Assuming the costs of holding liquidity will return to the levels in the decade before the pandemic, a liquidity line of $50bn would save MDBs between $411mn and $792mn per annum.3 A liquidity line of $50bn would be 24.4 percent of current MDB liquid assets and 7.7 percent of the top 20 central banks’ SDR holdings. In the following year, assuming those savings were retained, and the additional capital leveraged, MDBs could lend an additional $1.2bn to $2.4bn. In subsequent years, this amount would grow. Over 10 years, the liquidity line could enable total additional cumulative lending of between $12.4bn and $23.8bn and over 20 years, between $31.5 and $60.4bn. 

This proposal would allow the G20 to comply with its commitments for SDR recycling, it would allow central banks to deploy a portion of their non-utilized SDRs, it would promote global financial stability, it would release capital currently used to back MDB treasury portfolios, and over time it would increase the net income of MDBs, boosting their capital that could then be leveraged to increase their development assets helping countries attain the Sustainable Development Goals. The proposal would also bring the MDBs one step closer to operating as an integrated system.