IMF Conducts Capacity Building Work on Total Return Swaps as Sovereign Debtors Increasingly Turn to Controversial Debt Mechanism

IMF representatives have held meetings with financial institutions and member countries in recent months to conduct capacity building work aimed at deepening the institution’s understanding of self-collateralized sovereign borrowing transactions structured as repos or total return swaps, or TRS, sources told Octus.

TRS deals, repurchase agreements and other similarly structured sovereign debt instruments have increasingly become a favored tool for countries who are locked out, or priced out, of the traditional eurobond market. As a result, the sovereign credit community, including the IMF, is seeking to understand the significance of these transactions and the potential risks attached.

Sovereign TRS and repos are often bespoke to each country and so vary in their terms, and the details of the deal mechanics are rarely made publicly available. However, the overarching principle is that a country offers its own debt instruments – in many cases, local bonds – to a lender as collateral for a synthetic loan. One key detail that makes these transactions intriguing, and dangerous in the eyes of some, is that the value of the underlying collateral instruments offered is significantly greater than the value of the principal loan made out to the country.

Senegal, Angola, Nigeria, Argentina, Ecuador, and Colombia have borrowed using self-collateralized TRS or repo structures, typically with international commercial lenders. A handful of other countries have shown interest in the deals, according to sources.

These transactions have generated controversy and increased media scrutiny in recent months. Some experts in the field, including a number of prominent lawyers and academics, have raised the alarm over sovereign TRS and repos, which they say do not meet the same transparency and disclosure standards as other types of debt, complicate the capital structures of low-income countries that are vulnerable to default and would create novel dilemmas in the event of a restructuring.

“We’ve known about the possibility of doing these transactions for at least a decade, but they seem to have gotten much more popular recently,” said Mitu Gulati, professor at the University of Virginia Law School. “There are a thousand ways to see why these aren’t good, but it’s also easy to see how they can be attractive to finance ministers.”

The market is watching closely to see what stance the IMF will take toward the growing use of TRS. The Fund often plays a pivotal double role in sovereign restructuring scenarios, as both the lender of last resort and the creator of the debtor’s debt sustainability analysis, or DSA, which acts as the anchor for restructuring negotiations. IMF staff often emphasize that the Fund does not dictate whether or how to carry out a debt restructuring.

There has also been speculation around how TRS debt is currently, and should be, accounted for on sovereign balance sheets, and whether the IMF would treat the collateral instruments attached to TRS transactions as part of a country’s debt stack. In the case of Nigeria’s $5 billion in TRS transactions with First Abu Dhabi Bank, the IMF stated in its most recent Article IV report that staff views the collateral bonds as part of the country’s total debt stock for the purposes of the DSA.

Asked by Octus to comment on the IMF’s ongoing fact-finding work to better understand TRS, a spokesperson for the Fund said:

“Some emerging and frontier markets have increasingly used total return swaps (TRS) to access external FX financing, often when market access is constrained, benefiting from speed, size, and flexibility. As a general principle, such instruments would be treated as debt for the purposes of our debt sustainability analysis.

“While TRS may appear cheaper and allow access to financing when it is constrained, their true cost is opaque [and likely higher] due to undisclosed fees, opportunity costs of collateral, shorter effective duration, and margining features that can significantly raise ex post borrowing costs.

“As with other collateralized transactions, non-conventional borrowing instruments like TRS warrant caution, especially for developing economies with existing high debt vulnerabilities and insufficient institutional capacity to calculate and fully assess the risks involved.”

Below, Octus delves into the mechanics of sovereign TRS and repos, how they have been deployed by sovereign issuers, and some of the key concerns and points of debate.

What are Sovereign TRS and Repo Transactions? How Do They Work?

Although several notable self-collateralized sovereign borrowing transactions in recent years have been structured as TRS using ISDA documentation, the transactions are closer to secured loans in purpose and repurchase agreements in form.

The first confirmed instance of a sovereign using the TRS structure for a synthetic loan was a transaction between Angola and JPMorgan in December 2024, the first of a series of similar transactions by the Angolan government. Because Angola’s transactions used eurobonds as collateral, many details of the transaction were disclosed in bond offering documents.

In the first transaction, Angola issued $1.2 billion of eurobonds without collecting proceeds, and transferred them on a full title transfer basis to JPMorgan in exchange for a financing amount of $600 million. During the life of the swap, JPMorgan has “the right to sell or otherwise dispose of its ownership interest” in the collateral bonds, according to the base offering circular for the bonds. Angola classified the $600 million financing amount as debt, and the $1.2 billion of collateral bonds as contingent liabilities. But the offering documents note that if Angola does not receive all of the collateral notes back and cancel them (whether due to an event of default or other circumstances), any outstanding notes would be reclassified as debt, and that if both the financing amount and the collateral notes are classified as debt, the TRS transaction would contribute $1.8 billion to Angola’s debt stock for a financing of $600 million.

One source familiar with TRS transactions disputed the idea that the debt could balloon to this extent. They argued that in a scenario where a commercial bank were to sell underlying collateral notes in a bid to recover the value of a TRS loan, the bank could never recoup more than its initial loan claim, which would then be extinguished through the sale of the collateral. For example, a bank that has lent $600 million to a sovereign, backed by $1.2 billion of collateral bonds, can only recoup a maximum of its initial $600 million through the sale of those collateral bonds, while any additional proceeds must be returned to the issuer. The $600 million claim is thereby extinguished.

However, the source noted that in a stressed or restructuring scenario, the market price of the bonds would likely be much lower than par value. Therefore, the source argued, a $600 million TRS loan can ultimately result in the appearance of $1.2 billion of face-value paper in the secondary market, with the attendant interest payments, but the bonds are likely to have a lower sale value and/or recovery value in practice.

In January 2025, Angola upsized the initial TRS transaction at a lower collateralization ratio, transferring $728 million of its eurobonds to JPMorgan in exchange for a $400 million financing amount. Shortly before the two initial TRS were due to terminate in December 2025, Angola and JPMorgan entered another TRS with a three-year term, under which Angola transferred to JPMorgan the same $1.928 billion of eurobonds, plus $300 million in U.S. Treasury bills, in exchange for a financing amount of $1.5 billion.

Also in December 2025, Angola entered another three-year TRS agreement with Africa Finance Corp., or AFC, under which the government received $1 billion in financing in exchange for transferring debt securities with a face amount of $1.4 billion, including U.S. Treasury bills and Kwanza-denominated Angola treasury bonds.

The Angola offering documents also note that Angola may be required to pay down the TRS financing amounts or to deliver additional collateral “in certain circumstances.” In April 2025, Angola’s Finance Ministry told Bloomberg that a decline in the price of Angola eurobonds had triggered a $200 million margin call on its TRS with JPMorgan, which the country paid in cash. Quotes for the 2030 collateral bonds dropped to a low of about 86 on April 9, 2025, according to Solve. In May 2025, after the price of the bonds rebounded to the mid-90s, JPMorgan returned the cash to Angola, the Finance Ministry later told Reuters.

Fewer public details are available about TRS transactions entered into by other countries.

Earlier this year, Senegal’s Finance Ministry confirmed that the country had entered into seven TRS transactions between April and November of 2025, for a total of 721 billion West African CFA Francs ($1.26 billion), collateralized by around XOF 1 trillion of its own XOF-denominated debt securities. In its year-end 2025 budget report, Senegal’s Finance Ministry noted that it does not view the collateral securities as an obligation of the state. Senegal said that the net interest rate on the transactions was about 7.1%, lower than the average 11% to 12% yields on Senegal’s eurobonds at the time.

Senegal’s Finance Minister Cheikh Diba acknowledged in a press conference in March that the country’s TRS include an “independent amount” mechanism that can be the object of a call if interest rates rise and the value of the collateral package falls, although he said that the risk of such a call taking place was “almost nonexistent.” Senegal’s sovereign eurobonds were mostly trading in the 50s at the time of Diba’s comments and have broadly remained at those lows despite some fluctuation in the intervening months, excepting the nearest-term 2028 bonds, which have slipped to the mid-50s today from the mid-60s in March.

There has been heavy market speculation that the margin call mechanism may be tied to yields on new issuances of domestic bonds, and that the government’s rejection of a significant portion of bids at certain maturities in recent months is motivated by an attempt to avoid triggering a margin call.

An IMF spokesperson told Octus the Fund is seeking further details of Senegal’s TRS transactions.

“In the case of Senegal, the authorities informed IMF staff about a number of TRS transactions,” they said. “Some of the specific terms of these transactions have now been shared with the Fund, but further details are needed to understand the impact on the structure of public debt.”

“We remain engaged with the Senegalese authorities and continue to emphasize the importance of transparency, sound debt management practices, and a comprehensive assessment of all liabilities to support fiscal sustainability and accountability, and investor confidence,” they added.

The Nigerian legislature earlier this year approved $5 billion in TRS transactions with First Abu Dhabi Bank for the 2026 budget, collateralized by up to 133% of that amount in Naira-denominated government bonds. As of late June, Nigeria had drawn about $1.5 billion of that amount, Bloomberg reported.

The sovereign can be required to make dollar margining payments to First Abu Dhabi Bank in cash at any time if the value of the collateral securities declines below its initial dollar value, whether through declines in the market price of the bonds or depreciation of the naira against the dollar, according to a report by a legislative committee.

The IMF’s last Article IV report for Nigeria expressed concern that the possibility of margin calls tied to domestic bond prices and exchange rates could “give rise to political constraints on monetary or exchange rate policy.” The report also noted that IMF staff included the “full value of the collateral” in Nigeria’s external debt stock for their projections for the DSA.

Most recently, Argentina’s Central Bank, or BCRA, announced earlier this month that it closed $6 billion in repo transactions with 10 international banks, collateralized by bonds issued by the Republic of Argentina. According to Argentina’s legal advisor Cleary Gottlieb, each bank was given the option to structure its participation in the deal either as a repo or “a substantively analogous TRS,” and one of the banks chose a TRS structure. This most recent transaction replaced a series of repos between BCRA and international banks, some of which were collateralized by Republic bonds, and some by bonds issued by the BCRA itself. The BCRA said that it will pay an interest rate of SOFR + 4% on the newest round of repos and TRS, which are scheduled to expire in September 2028.

TRS: The Benefits and Risks

Many of the features that most concern the IMF and outside experts about the effect of sovereign TRS and repo transactions on the broader sovereign debt system are the same features that make them attractive to lenders.

Some of those concerns relate to the pro-cyclical nature of TRS. The fact that a deterioration in a borrower country’s creditworthiness can trigger certain features in TRS that can deepen and accelerate the deterioration. That includes the fact that a default on a TRS will cause the collateral bonds to become live indebtedness, leading to a sudden jump in a country’s debt stock as well as the potential for margin calls.

The threat of disastrous compounding consequences for a default on a TRS may reassure lenders, by motivating debt managers to prioritize repaying these instruments ahead of a restructuring, according to Lee Buchheit, honorary professor at the University of Edinburgh Law School.

“If the swap is not paid off at maturity, the bureaucrat who entered into the transaction is going to have to explain why he or she sold the full faith and credit of the country at an enormous original issue discount a few months before without ever telling parliament, without ever telling the IMF, without ever telling the rating agencies,” he said. “And any right thinking bureaucrat is going to sell the family silver to try to pay off this swap right before a general debt restructuring begins.”

In cases where the collateral includes domestic debt securities, a default on a TRS could also threaten a country’s financial and banking systems, creating an even more effective “ransom arrangement,” according to Anna Gelpern, professor of law at Georgetown University.

Gelpern noted that if the collateral is an asset that is important to a country’s banking system, “you’re buying yourself not just a sovereign debt crisis, but a banking crisis. And that’s part of the value of the collateral, that the borrower won’t mess with the debt that would bring down the banking system of their country.”

The incentives for sovereigns to pay a TRS in full before defaulting on other obligations effectively gives TRS lenders a status akin to superseniority, even if they do not benefit from official priority over other creditors.

Ecuador’s treatment of its self-collateralized repos in 2020 shows how that dynamic plays out in practice.

In 2018, Ecuador entered into two four-year repo transactions, one with Goldman Sachs and the other with Credit Suisse. In each transaction, Ecuador transferred about $1.2 billion of its own bonds to the lender in exchange for financing of about $500 million. Ecuador’s bonds began to trade down significantly in late 2019 and early 2020, leading to hundreds of millions of dollars in margin calls that were cited by the IMF as one factor that contributed to a massive drop in Ecuador’s foreign exchange reserves. In April 2020, Ecuador voluntarily prepaid the repos, sparing the repo lenders from Ecuador’s general debt restructuring later that year.

If a TRS or repo is not paid off before a general default, it’s not clear how the financing amount or the collateral bonds would be treated in a restructuring.

According to Buchheit, the collateral securities have an uncertain status even during the life of the swap, because the issuer (at least in the case of Angola) has made it clear that they “do not regard those securities as live indebtedness,” making them “undead” until they are “vivified” by the swap being terminated without the securities being returned to the issuer.

Even though theoretically, the lender has full title over the securities and the right to sell them at any point, this feature makes it practically complex for the securities to be traded while the swap is still outstanding.

“If the new bonds are vivified only upon termination of the swap and not returned, what would the lender sell if it disposed of its interests before that day? It would be selling zombie securities,” he said. “I would think the lender would be in an awkward position to dispose of those securities before the swap terminates, and indeed, I’m sure that was the wink and nod understanding.”

And even if a sovereign defaults on a TRS, indisputably “vivifying” the bonds, it’s not clear that those bonds will receive the same treatment in a restructuring as other bonds issued through standard channels, Buchheit noted, adding that treating all bonds the same in this situation could arguably amount to preferential treatment for the formerly-pledged bonds.

In that case, “the debate in a sovereign restructuring will be what value should be ascribed to those securities given they were effectively issued at an original issue discount,” Buchheit said. “The other people who bought the country’s debt securities for 100 cents are going to be given some kind of debt treatment, and if you gave that same debt treatment to a guy who paid significantly less than that, he is obviously being given preferential treatment.”

A full assessment of TRS remains difficult, since so much about the actual terms of the contracts is unknown. While some market participants argue that the risks are overstated, others read the opacity itself as evidence that there’s something worth hiding.

“Looking back, how bad does it have to be for the docs to still be hiding despite all the stench that the transactions have produced?” Gelpern said.

The deeper concern, for Gelpern and others, is one of fit: the countries most likely to resort to “hail-mary transactions” like TRS tend to be the ones least equipped to absorb and manage the volatility the structures can introduce.

“These transactions are now being used to gamble for resurrection in 11th-hour, don’t default in the next three weeks sort of situations,” Gelpern said. “And that’s a terrible use of these structures, because it just makes the eventual default that much more painful, and it tends to shift risk from those who are better able to handle it to those who are less able to handle it.”