What to Do About the Liquidity Difficulties in Low-Income Countries

The IMF and World Bank have highlighted that many low-income countries (LICs) face “liquidity” difficulties as debts taken on prior to the pandemic (typically from Chinese policy banks) and after the pandemic (often from the multilaterals or via domestic borrowing) come due. Treasury Under Secretary Jay Shambaugh further highlighted the issue in his April speech.

These liquidity concerns are certainly real.

The stock of external debt taken on by LICs, as well as some weaker frontier economies, increased markedly between 2010 and 2020. Now, those countries have to refinance in a world of much higher interest rates.

As the Bank and Fund note, debt amortizations (domestic and external I think, though the IMF and World Bank do not specify) of the world’s LICs increased from $20 billion a year during the past decade to around $60 billion year. The LICs have a combined GDP of only $500 billion, so $60 billion is a significant sum.

The World Bank’s data shows external debt service rising from $6 billion in 2022 to $11 billion in 2024 in no small part due to of rising payments to “non-concessional” bilateral creditors (cough, cough, Chinese policy lenders). That can be assessed relative to $100 billion in exports—and in many cases, a large need to import food and fuel. I at least am not comfortable with the ongoing IMF-World Bank effort to move toward “it’s all fiscal” measures of debt sustainability and prefer to focus on the traditional external debt metrics.

Moreover, the set of countries with immediate liquidity difficulties and potential incipient solvency problems is not just confined to the LICs.

Consider a set of lower middle-income countries that currently have, or have had, IMF programs: Angola, Kenya, Ecuador and Pakistan.

Their stock of external debt is up significantly (as is the debt of all of sub-Saharan Africa) and has now started to amortize rather relentlessly. That means all these countries need to raise a lot of new funds in the market just to refinance their existing debts.

All four of these countries also happen to have (or to have had in Angola’s case) IMF programs that are built around adjustment to restore the confidence of bilateral and market creditors rather than debt rescheduling. These programs are—to use a phrase I like—designed as “bailouts” (of existing lenders), not “bail-ins.”

Kenya was obviously squeezed on the cost of refinancing its $2 billion bond that matured earlier this summer. It pays more in interest on the new $1.5 billion bond it issued (generally to the folks who got paid the $2 billion) than it did on the old $2 billion bond and made up the difference by borrowing $500 million from the multilaterals. Kenya also is paying a pretty penny on its Chinese policy loans, some of which carry a punitive LIBOR + 360 basis point interest rate. Kenya pays significant amounts of interest on its rising domestic debt stock as well.

Even though Kenya managed to avoid a default amid the liquidity pressured associated with the 2024 spike in external payments, it only did so by paying interest rates that put its future solvency at risk.

Through its IMF program, Kenya is required to undertake significant fiscal adjustments to reduce future borrowing needs and cover the rising cost of debt service. The associated tax increase generated a political crisis. Social stress from paring domestic spending and raising broad domestic taxes to cover rising debt servicing costs is thus quite obvious.

Ecuador has paid a decent sum back to China over the past couple of years, running down its reserves in the process. See the line items for “bilaterals” and “banks” in Table 3 of the IMF’s latest staff report.

Ecuador now has a (poorly designed*) IMF program to rebuild those depleted reserves. However, given the structure of its existing bond payments and timing of the disbursements in its new IMF program, Ecuador faces a debt service hump in 2026. It is not at all clear if Ecuador received real assurances from its Chinese lenders that they won’t keep taking funds off the table in the interim.

In the section in Ecuador’s program on financing assurances, the IMF only notes that “The authorities are actively discussing options with the main bilateral creditor to maintain exposure to Ecuador.” Actively discussing options isn’t an agreement in my book.

Angola exited from its IMF program back in 2022 when oil prices soared, but it has (quietly) been making big payments back to Chinese policy banks after a rescheduling in 2020. In 2022, it made net payments of over $1 billion back to bilateral creditors and commercial banks (commercial banks include the CDB and the Chinese state commercial banks), as well as paying $5 billion on “private” debts. Principal due to bilateral creditors and commercial banks is now around $6.5 billion a year for several years—an amortization profile that will quickly erode Angola’s limited buffers. Reserves are around $15 billion relative to $45 billion or so in external debt; it won’t take much to push Angola back into the distress

Pakistan is in a similar position. See Murtaza Syed in the Financial Times, among others.

It really has no reserves; at least net of the funds the State Bank of Pakistan owes directly to the PBOC and the Saudis.

Its debt amortization schedule is now relentless.

Net flows were negative in 2022 thanks largely to large repayments back to (mostly Chinese) commercial banks—and now Pakistan faces close to $7 billion in payments in 2025 on its bilateral loans. Those payments should be judged not relative to Pakistan’s overall economy, but relative to its meager international reserves and tiny export base.

Pakistan also has a new IMF program that is built around full payment of maturing debt with hopes of restoring market confidence via fiscal and currency adjustment.

The external debt profile of these four countries combined is in fact much more dire than the debt profile of the LICs (many of whom are too poor to borrow much).

Indeed, the issue that the IMF and World Bank highlighted for the LICs actually applies to all of sub-Saharan Africa several weaker lower middle-income countries in Latin America and South Asia.

Net flows to Africa fell off a cliff between 2018 and 2022 and are now almost certainly still negative, as more and more of the debt taken on during the peak inflow years is now coming due.

Debt service has spiked across the continent, not just in Angola and Kenya, and will stay high for some time.

The point being that resolving liquidity issues is hard, and it takes a combination of real money and hard-headed analysis to assess whether countries are better off trying to pay or biting the bullet and seeking a restructuring of their external debts.

The IMF and World Bank have thus drawn attention to an important issue, even if they have framed the challenge much too narrowly by focusing entirely on LICs.

But it isn’t clear what the two institutions actually intend to do about the problem. This is what the Fund and the Bank say:

“New solutions are needed to support countries that do not have solvency problems but need to manage the high debt servicing levels. These include mechanisms by multilateral or bilateral partners to mobilize new financing, including from the private sector, at affordable terms using credit enhancements to refinance existing debt. Countries could also pursue liability management operations, including debt-for-development swaps and debt buybacks where appropriate.”

That wording muddles together a set of conceptually different solutions, and at no point are the cost of different options spelled out.

For the sake of clarity, it is useful to focus on a series of “corner” solutions.

One would be to have multilateral and bilateral “partners” solve the problem by lending the countries’ needed sums at a low cost. This isn’t necessarily a bad idea.

One can take the view that a set of countries gambled on the bond market and on expensive loans from Chinese policy banks, it didn’t work out, and they now need to shift back to relying on concessional—or at least reasonably priced—official support to create fiscal and balance of payments space.

For the LICs, covering all external debt coming due is a manageable $16 billion over the next four years; for my set of four troubled middle-income countries, it is more like $60 billion over four year—and for all of sub-Saharan Africa it is around $110 billion over four years.**

Of course, some of the redemptions will be to existing multilateral creditors, so that bit would not require net new financing.

For the LICs, $4 billion of the $10 billion is due to the MDBs, who would normally be expected to provide a comparable amount of new money. Much of the rest is due to existing bilateral creditors (again, Chinese policy banks)—one obvious next step would be for the Bank and the Fund to identify just who is owed what. Chinese policy lenders have been quite willing to take money off the table in Ecuador and Angola—so it isn’t obvious their participation in any initiative can be assured. The true corner solution, though, would use official lending to refinance both expensive Chinese policy lending and maturing private bonds/loans.

Another corner solution is simply to rely on market borrowing. That is, however, extremely expensive for the country—and likely pushes some countries that are on the edge of unsustainability into deep debt distress and even default, as not all countries would be able to raise the needed funds in practice. It clearly isn’t the option that the IMF and World Bank prefer.

The apparent middle ground option is to use official funds to lower the cost of market borrowing (“credit enhancements”). Stressed countries then would not need to rely entirely on the official sector, nor would they have to pay a market interest rate to rollover their debt. Such credit enhancements thus offer hope that the existing stock of debt could be sustained in a way that doesn’t drown out spending on health and the climate.

But credit enhancements have one major problem which is often glossed over by proponents: almost all credit enhancements price as an inefficient mix of official money and private money. In general, the country would be better off, say, covering 50 percent of its borrowing need with official money and 50 percent with private money rather than seeking to raise funds with an instrument that is 50 percent guaranteed. With enhancements, and for that matter collateralized structures, the whole is usually worth less than the sum of its parts.

Bonds with MDB credit enhancements also tend to complicate restructurings (the holders of the bond don’t want to strip it down to a guaranteed component and an unguaranteed component that is part of the broader restructuring).

The IMF and the World Bank did not mention another corner solution: reschedule maturing bonds and maturing Chinese policy loans at a reasonable interest rate (while the MDBs provide new money equal to or bigger than the sums they are due).

Kenya borrowed at around 6 percent 10 years ago, and it could have negotiated to extend the maturity of that bond for another 10 years at 6 percent. The bondholders, of course, would object.

The Export-Import Bank of China (Exim) could have been asked to do something similar, extending its railway loan for 10 years at LIBOR rather than LIBOR plus. China would also object.

But all the creditors live to fight another day and need not give up hope for payment in full over time.

In such a scenario, IMF and World Bank funds also would help cover coupon payments on existing external debts, and World Bank, African Development Bank, and other regional MDB funds could support new projects as well.

But programs based on rescheduling require a willingness to stop payments for a time if creditors aren’t willing to renegotiate legal terms. It thus takes courage from all involved.

Based on the language in the blog, I worry that the IMF/World Bank proposal falls in the mushy middle—a bit of this and a bit of that without any real concrete plan beyond trying to avoid large net outflows on a case-by-case basis.

To make the Fund-Bank proposal real, it would help to have detailed data on the redemptions due over the next few years, as well as a hard-headed assessment of whether covering a large fraction of those redemptions to buy down the cost of market borrowing (either directly or through providing credit enhancements) is the best use of official resources.***

That would provide real insight into how much money is needed to sort out the high cost of already existing debt under different policy options.

I also obviously think the analysis should be extended to a broader set of countries, including a set of countries that have existing IMF programs that have not been constructed around any debt treatment.

*Ecuador’s debt sustainability was assessed under the new “it’s all fiscal” market access debt sustainability framework, which ignored Ecuador’s most obvious vulnerabilities, namely its large stock of external debt relative to its foreign exchange reserves and its large external debt service relative to its reserves.

Focusing exclusively on the overall gross financing need (independent of the currency composition of the debt or whether the debt is owed to residents or non-residents) and the shape and slope of the fan chart isn’t a sufficient basis for solid judgments about debt sustainability.

**Costs of course increase if cheap external debt replaces expensive domestic debt; my analysis focuses on the external debt payments identified in the World Bank’s global debt dataset.

*** It would also help to provide a clear breakdown between domestic and external debt, with a separate category for the intermediate case of non-resident holdings of local currency/local law debt (such debt needs to be broken out for clarity on how exactly it is being counted; it will be small in most cases).