Can a bond that no longer exists trigger a default?
This is the quasi-metaphysical question facing the panel of experts who have to determine whether Credit Suisse’s credit-default swap contracts will pay out. Call it Schrödinger’s swap.
While most in the market had assumed that Switzerland’s decision to vapourise $17bn of Credit Suisse’s Additional Tier 1 bonds would not trigger the bank’s CDS (for reasons we’ll get into below), a couple of US hedge funds have wagered that nearly everyone has overlooked crucial quirks in the derivative contracts that could lead to a payout.
Bloomberg revealed last week that FourSixThree Capital and Diameter Capital Partners have loaded up on Credit Suisse CDS aiming to profit from a ruling that the swaps have in fact been triggered. On Thursday, a question was lodged with the Determinations Committee — the all-powerful panel of market representatives who make often-controversial rulings on whether credit-default swaps pay out — asking whether a so-called credit event has occurred at Credit Suisse (CDS aficionados will have noted that the question went in just a few days shy of the 60-day deadline after the AT1 bonds were written down).
The rationale behind why these funds are betting on a payout is not yet public, however. Bar some clues in the phrasing of the question submitted to the DC and the supporting documents attached, the wider world is none the wiser as to why two sophisticated hedge funds believe most of the market has missed a seismic event in the world of CDS.
In a bid to fill this void, FT Alphaville has spoken to several players with knowledge of the trade in order to set out as clearly as possible the thorny argument in favour of a CS CDS payout.
We can’t promise that the explanation will be straightforward. But you know you’re in for a wild ride when the supporting materials include documents as esoteric as a bond prospectus from the year 2000 and a more than a century old New York court judgment.
Tiers for fears
Before we get into the nitty gritty of the argument in favour of a trigger, it is first worth explaining why the majority of CDS traders had assumed there would be no payout on the Credit Suisse swaps.
CDS contracts written against bank bonds have two main classes: senior and subordinated, referencing the debt at the top and bottom of the capital structure, respectively.
You might naturally assume that when Swiss authorities wiped out $17bn of Credit Suisse’s deeply subordinated AT1 bonds, as a condition of the bank’s shotgun wedding to UBS, it would trigger payouts on the latter set of contracts.
Indeed, the International Swap and Derivatives Association overhauled the standard CDS contract definitions in 2014 to include a new “governmental intervention credit event” to cover exactly this sort of eventuality (one of the reasons for the overhaul was a fiasco surrounding the CDS of Dutch bank SNS Reaal the previous year, where swap holders received no payout even after the government zeroed its subordinated bonds).
The problem is that bank capital structures are rather more complicated than just being split into “senior” and “subordinated” bonds. There are multiple tiers of debt. And for the purposes of subordinated CDS, it is the Tier 2 bonds that matter.
This is because AT1 bonds, which banks only first started issuing around a decade ago, have a number of structural features that make them incompatible with CDS. For example, their perpetual maturities (which mean they never have to be repaid) clash with the requirement in standard CDS contracts that the relevant bonds have a maximum maturity of 30 years.
This distinction did not matter the first time that AT1 bonds were written down in Europe — in the case of Banco Popular in 2017 — because regulators zeroed the Spanish bank’s Tier 2 bonds in tandem. This led to a full payout for those who had bought CDS protection on Banco Popular’s subordinated bonds.
In the case of Credit Suisse, however, Swiss authorities left its Tier 2 bonds completely unscathed (rather miraculously, given that the Tier 2 debt had an incredibly similar writedown clause to the Swiss bank’s AT1 bonds).
With the Tier 2 bonds unimpaired, it seemed logical that there would be no payout on Credit Suisse’s subordinated CDS contracts.
Or so you might have thought.
Y2K bug
The first hint that all might not be as it seems comes from the fact that among the supporting documents in this week’s question to the DC is the offering circular for a £250mn subordinated bond that Credit Suisse’s US subsidiary issued in the year 2000.
Students of banking M&A history may be interested to discover that these bonds were issued to fund the acquisition of US investment bank Donaldson, Lufkin & Jenrette, back when an expansionary Credit Suisse First Boston was trying to make deeper inroads into the US junk bond market:
Crucially, however, this bond matured back in 2020. How can debt that no longer exists have any bearing on Credit Suisse’s CDS contracts in the here and now?
The answer lies in one of the many ways that bank capital structures have become even more complex as the result of the global regulatory framework introduced following the 2008 financial crisis.
On top of the standard capital requirements, globally significant banks also need to have a certain amount of bonds that can be easily written down in a crisis, not-so-catchily known as “total loss-absorbing capacity”. To fill up these buckets, banks began issuing a new tier of bonds that ranked between senior and Tier 2.
In the EU, banks mostly opted for so-called contractual subordination, where bonds are issued from the same entity and their contracts specify different orders of priority. Rather than giving them the more logical name Tier 3 bonds, bankers christened this new class of debt with the rather oxymoronic moniker “senior non-preferred”.
In the UK and Switzerland, however, banks instead adopted structural subordination, where bonds are issued out of different entities to determine their order of priority. While traditional senior bonds were issued from the main banking entity, the new bonds were issued out of a holding company. The two classes of debt became colloquially known as “OpCo senior” and “HoldCo senior” respectively.
The shift created even more complexity for European bank CDS, with different contracts emerging to reference the differing ways the two tiers of senior debt were issued across the continent.
It means that for Swiss CDS, the distinction between bonds at the HoldCo and the OpCo matters. And Credit Suisse’s subordinated CDS references Credit Suisse Group AG (the HoldCo) not Credit Suisse AG (the OpCo).
You would probably expect that the relevant bond for this contract (the “reference obligation” in CDS parlance) to be the Credit Suisse Tier 2 bond that escaped a writedown in March. And yet that bond’s offering memorandum makes clear that this bond is an obligation of Credit Suisse AG, not Credit Suisse Group AG:
At this stage it makes sense to turn to the “Standard Reference Obligation list” maintained by IHS Markit to get a definitive answer on the contract’s relevant bond. And lo and behold, it lists the £250mn bond issued in 2000 that matured in 2020:
A quick check of the bond’s prospectus shows that in contrast to the Tier 2 bond, it was guaranteed by the same entity that the subordinated CDS references, Credit Suisse Group:
While it might seem deeply weird for CDS to reference a bond that no longer exists, ISDA’s standard CDS definitions do allow for this possibility, with the concept of a “prior reference obligation” for situations where no reference obligation is outstanding.
So, incredibly, it could well be the bond we need to look most closely at in determining whether there has been a CDS default.
Unwritten ranking?
So far, so strange.
But you could be forgiven for still wondering what relevance this legacy bond has to Credit Suisse’s AT1 wipeout. After all, it goes without saying that a bond that was repaid nearly three years ago was not written down as part of the UBS merger.
The next clue lies in the fact that the DC has not only been asked to rule on whether a governmental intervention credit event has occurred at Credit Suisse Group. The question submitted to the panel has a supplementary query:
“The Eligible Market Participant who submitted the question noted that in order to determine the DC Question the EMEA DC would need to determine whether the AT1 bonds were Not Subordinated to the Reference Obligation or the Prior Reference Obligation and asks the EMEA DC to consider this point in connection with the DC Question.”
This is relevant because it is not only a default on the specific “reference obligation” that leads to a CDS payout. The swaps should also be triggered if the borrower defaults on debt that ranks at the same level or higher (or, to use CDS parlance, debt that is “not subordinated to the reference obligation”).
If the AT1 bonds that were zeroed were deemed to rank pari passu with the £250mn subordinated bond that matured in 2020, that may be enough to trigger the swaps.
On the surface, this does not appear to be the case. The form of legacy bank capital that Credit Suisse issued in the year 2000 was colloquially known as “lower tier 2”. In the order of priority that makes up a bank’s capital structure, it follows that “tier 1” ranks behind “tier 2”.
And yet, in the 2020 maturity bond’s terms and conditions there is no reference to it being “tier 2” debt. The bond’s terms only set out that they are generic “subordinated obligations”. Unsurprisingly, there is no reference to the bond’s ranking in relation to AT1 debt, given that these securities did not yet exist when the bonds were issued in 2000.
The hedge funds that are wagering on a payout are betting on the fact that, while market participants generally understand that the subordinated bond issued in 2000 ranked ahead of the AT1s, there is not anything making that explicit in the documents relevant to a CDS decision.
In this logic, there is an amusing echo of the debate between Credit Suisse’s AT1 holders arguing that their debt was commonly understood to rank ahead of the bank’s common equity (which paradoxically was not zeroed in the merger and received $3bn) and sceptical observers telling them that this was never actually enshrined in the documents.
The final piece of the puzzle is a New York court decision from 1918, which FT Alphaville understands has been submitted to the DC as part of a belt and braces approach to proving that perpetual bonds are still legally considered debt (the ruling deals with a “perpetual debenture” issued in 1901, a sort of early 20th century version of an AT1 bond).
Senior moment
Let’s imagine for the sake of argument that this gambit works. The DC agrees that there has indeed been a governmental intervention credit event in respect to Credit Suisse Group’s subordinated CDS. What then?
CDS contracts do not all pay out the same amount when they are triggered. In most instances, an auction of the relevant bonds is held to determine what the loss to bondholders has been and to compensate them accordingly (for example, an auction that valued the bonds at 25 cents on the dollar would lead to a 75 cent payout on the CDS).
Given that the subordinated bond Credit Suisse issued in 2000 has already been repaid and its other bonds that are eligible for a CDS auction were not impaired in the UBS merger, you might assume there is little reward in store for those betting on a surprise CDS default.
This is where the hedge funds betting on a payout are banking on the sharp moves higher in interest rates over the past year coming to their aid, however.
Credit Suisse Group has a bond outstanding that is trading at around 65 cents on the dollar, purely as a function of the dramatic change in underlying rates since it was issued at the end of 2021 (the bond in question has a skimpy 0.625 per cent coupon and a decade-long 2033 maturity, making it particularly sensitive to interest rate changes).
While this is one of the so-called “HoldCo senior” bonds we discussed earlier, CDS contracts explicitly allow for bonds that are more senior than the reference obligation to be submitted into an auction. In the event that Credit Suisse’s subordinated CDS is deemed to have triggered, the hedge funds stand to bag a tidy 35 cent pay out.
That’s nice in theory but…
That’s the trade as FT Alphaville understands it. Essentially, these two funds think the ambiguity over the ranking of a dead bond is enough to turn the AT1 wipeout into a CDS payout.
The big question remains: will it work?
¯_(ツ)_/¯
The logic behind it all may at times seem counterintuitive and faintly ridiculous, but then CDS decisions are frequently counterintuitive and faintly ridiculous. The members of the DC are not supposed to rule on how market participants intended the contracts to behave, but on their strict legal definitions, unintended quirks and all.
The DC accepted the question late on Friday and will hold their first discussion on the matter later today, so the panel certainly thinks it is at the very least worthy of debate.
It is also worth noting that there are serious players involved. FourSixThree and Diameter are both distressed debt specialists whose principals have extensive experience trading CDS. They have also enlisted a heavyweight law firm in Kramer Levin to support of their cause.
One thing that does strike us, however, is that there are several steps that the DC needs to go along with for a payout to occur. This means the trade has multiple points of failure.
For example, while the four corners of the 2000 vintage bond’s offering memorandum do not contain an indication that it is a Tier 2 bond that ranks ahead of the AT1, the order of hierarchy can be found in other publicly available documents. This document Credit Suisse issued in 2019 explicitly states that the 2000 bond has a “Tier 2” ranking, while also stating the AT1 bonds’ “position in subordination hierarchy in liquidation” is behind the Tier 2.
If the DC does not agree with the notion that the AT1 bonds are “not subordinate” to the “prior reference obligation”, then the whole trade falls apart. Athanassios Diplas, one of the architects of the modern CDS auction, has tweeted that this could be an “uphill battle”:
A question has been submitted to the DC asking to confirm that the CS AT1 bonds were not subordinated to the RefOb of the CS CDS contracts.
Don’t see how that would be possible given the nature of the AT1s.
Feels like uphill battle.
PS I cannot access some of their attached docs https://t.co/JbPW18HXrt— Athanassios Diplas (@AthanaDiplas) May 11, 2023
Whatever the outcome though, one thing is for certain: the trade is delightfully weird and fun.
And FT Alphaville is strongly in favour of weird and fun trades, whether or not they are ultimately successful.
Further reading:
— Who killed Credit Suisse? (FTAV)
— Beware of Greek lessons for AT1 bondholders (FTAV)
— Why bank capital has a problem (FTAV, 2017)