The Financial Times ran a piece last week reporting that Senegal — yes, that Senegal — had borrowed €650 million through instruments it had not disclosed to the IMF, to its existing bondholders, or apparently to anyone who might have objected. Which, it turns out, is everyone. I clicked on it because it was about Senegal. That sentence is doing more work than it appears to.
In my experience, the FT does not typically concern itself with Senegal. And, furthermore, when it does, the news is almost never good for Senegal. This piece did not disappoint — unless, I suppose, you are a longstanding creditor of Senegal, in which case it probably disappointed you considerably.
I want to explain why this is interesting beyond the ordinary spectacle of a small country in financial distress, because the interesting part is not the distress. The interesting part is the instrument.
(Ed: upon learning the subject of this post, observed that my idea of “weird” and his idea of “weird” have continued to diverge, and that he has developed a reliable heuristic: if I describe something as “quotidian” or “seemingly mundane,” it means there is a formal impossibility result within three paragraphs. He is not wrong.)
A total return swap, in the form Senegal used, works roughly as follows. Senegal issues domestic bonds — in CFA francs, backed by the full faith and credit of the Senegalese state, such as that currently is — and transfers title to those bonds to a lender: in this case Africa Finance Corporation (AFC) and First Abu Dhabi Bank (FAB). In return, Senegal receives cash. At the end of the agreement, if everyone has behaved, the bonds come back and the cash is repaid with interest. If Senegal defaults in the meantime, the lenders keep the bonds — now worth less than they were — and Senegal owes a cash penalty to compensate for the gap.
The financial press describes total return swaps as “increasingly popular with countries close to default.” This is the most polite possible way of noting that the instrument is essentially a near-default technology: you use it when your access to conventional borrowing has been cut off, which is itself a signal to any observer that you are in serious trouble. AFC and FAB knew this. Senegal knew this. This matters, and I will come back to it.
The swaps gave AFC and FAB something Senegal’s existing bondholders do not have: seniority. In a debt restructuring, not all creditors are treated equally. Senior creditors get paid first; junior creditors wait in line and hope. The TRS structure was explicitly designed to place AFC and FAB ahead of the existing bond market in the event of a default. Whether it succeeded is another question entirely.
Here is the first problem. Seniority is not a property of a contract. It is a property of an enforcement architecture.
In corporate debt, seniority is meaningful because there is a bankruptcy court that will honor it: an automatic stay on creditor actions, a process for ranking claims, a judge with the authority to cram down a restructuring over the objections of minority creditors. The hierarchy written into the contract maps onto a hierarchy in the real world, maintained by an institution with the power to enforce it.
Sovereign debt has no such architecture. There is no international bankruptcy court. There is no automatic stay on sovereign obligations. There is no cram-down mechanism. There is only a negotiation — typically among the IMF, the bondholders, and the debtor government — whose outcome is determined by relative bargaining power, not by the priority rules written into private contracts signed in Abuja and Abu Dhabi. I wrote about a version of this problem a decade ago, when Greece and Varoufakis were playing a similar game with different instruments and similarly unenforceable threats; the post is here if you want the longer version.
This means the TRS’s seniority claim is, in a precise sense, an instrument that presupposes an institutional environment that does not exist. The contract is trying to specify its own priority within a game whose rules are ultimately set by a negotiation the contract cannot control — which is structurally identical to the problem at the center of the recent post on self-enforcing rule systems: a rule system cannot certify its own consistency from within. It requires a meta-level that the system itself cannot supply. The TRS seniority claim is a contract trying to do exactly that. The meta-level, in the form of the IMF, has already signaled it intends to classify the swaps as external debt subject to restructuring. The IMF is the court. The court has ruled that AFC’s and FAB’s seniority claim is not recognized in this jurisdiction.
There is a second problem, logically independent of the first, concerning the credibility of the penalty structure.
The TRS is, among other things, a commitment device. Senegal is promising AFC that it will service this debt preferentially even under distress, because deviation triggers a painful cash penalty: the AFC documentation specifies that in the event of default it “will very likely” mark the bond security to zero, generating a large cash claim against Senegal. The theory is that this threat, combined with Senegal’s formal acknowledgment of AFC as a preferred creditor, makes the commitment credible.
But the penalty is a cash claim against a country in default. The mechanism dissolves precisely at the moment it is supposed to bite: if Senegal had cash, it would not be defaulting. What the commitment device requires is a solvent counterparty against whom to enforce the penalty, and the trigger condition for the penalty is that counterparty’s insolvency. This is not a minor implementation detail. It is the structure of the problem. A commitment whose enforcement mechanism requires a condition that the trigger condition negates is not a credible commitment. It is a very expensive promise.
This is the logic behind the familiar aphorism: if you owe the bank $100 and fall on hard times, you have a problem. If you owe the bank $100 million and fall on hard times, the bank has a problem. AFC and FAB are €650 million creditors against a $40 billion-plus debt stack — not large enough, on their own, to flip the aphorism through scale. The TRS was their attempt to flip it through contractual hierarchy instead: to manufacture, through seniority, the leverage that size would otherwise provide. The problem is that manufactured seniority, unlike scale, depends on an enforcement architecture that sovereign debt lacks. So they have built a paper version of the $100 million dynamic. It looks like leverage. It is denominated in leverage. But when Senegal defaults, AFC’s recourse is to send a strongly worded notice to Dakar.
I want to complicate this framing, though, because AFC and FAB are actually closer to “the bank with the $100 million problem” than I have just implied — not because of their scale, but because of something more interesting.
“Falling on hard times” is not exogenous here. The existence of the TRS arrangements, and specifically the seniority they confer, is itself a partial cause of the crisis they were designed to navigate. If Senegal’s existing bondholders knew that €650 million in claims had been quietly elevated above theirs in the creditor hierarchy, they would not sit still. They would demand accelerated payments, seek legal remedies, reprice Senegalese debt downward, or simply exit. Any of these responses would bring forward the default the TRS was designed to bridge.
AFC and FAB and Senegal all knew this. How do we know they knew it? Because they wrote a confidentiality provision into the contract. Senegal was required to keep the existence of the AFC loan secret — including, remarkably, advising AFC before informing the IMF, if at all possible. The secrecy is not incidental to the instrument. The secrecy is load-bearing. The deal functions only as long as it is not observed by the parties whose behavior it would alter if observed. There is a rich older literature on the strategic value of not knowing things, and on the conditions under which transparency actually makes outcomes worse. A more recent post explored how a principal who publicly announces ignorance enrolls themselves as a named party before a meta-principal who observes outcomes regardless; the Senegal case is the private-sector mirror image, where the instrument’s concealment is not accidental but structurally required. The Senegal case is a fairly pure example: disclosure of the instrument would change the behavior it was designed to manage.
This is performativity in a precise sense: the instrument is not merely a bet on an independently-determined outcome. It is partially constitutive of the outcome. The observation of the instrument changes the behavior of the other players, which changes the probability of the event the instrument is pricing. The map is quietly redrawing the territory, and everyone else is still navigating by the old one.
I want to plant a seed here, to be harvested in a later post: this structure — an instrument that hides itself from the system it is embedded in, whose disclosure would alter the very outcome it is designed to address — is not unique to sovereign debt. It will reappear when we turn to classification algorithms and the systems that run on them. The connection was made explicit in the post on IRS-ICE data sharing and the endogenous base rates problem; the Senegal case is the same structure in a bespoke suit, carrying a term sheet. Senegal’s bondholders are conducting debt sustainability analysis on a portfolio that is missing €650 million in senior claims. That is not only an information problem. It is a classifier that does not know its own training data has been altered.
There is a version of the “$100 million” aphorism in which the bank has also secretly borrowed against the loan it made you, at preferential rates, without telling you or your other creditors. The version without that clause is already bad enough. The version with it is what Senegal’s bondholders woke up to find in the Financial Times. Which is, come to think of it, exactly why I clicked on it.
With that, I leave you with this.