A message from the FCA to ISDA on LIBOR fallbacks

In an interesting development, the FCA just made a public plea to ISDA to extend the current definition of the LIBOR discontinuation triggering events.

On Monday, Edwin Schooling Latter (the FCA Director of Markets and Wholesale Policy) delivered a speech at the ISDA Annual Legal Forum about the LIBOR Transition. More specifically, he chose to focus on the thorny issue of the “cessation triggers” in the fallback language that would apply to legacy contracts referencing a LIBOR index.

The main idea pitched here, is to include a supervisor announcement that a benchmark is no longer economically representative of its market as one of the fallback triggers for OTC derivatives.

Here is the link to the full speech.

Rising adoption of alternative reference rates (ARRs): SONIA & SOFR

The message softly starts with a heartfelt acknowledgement of the progress the industry made so far, highlighting encouraging signs of rising liquidity in SONIA and SOFR denominated instruments.

“In fact, on a monthly basis, cleared notional in SONIA swaps is now higher than that for sterling LIBOR.”

Of course, on that specific metric, one might object that a true measure of success would be the adoption of ARRs by loan and derivatives end-users (corporate treasuries especially). Yet, anecdotal evidence points that this still isn’t really happening – as shown by the low volumes of non-cleared SONIA swaps from the 2018 statistics ISDA published yesterday (see below):

As a point of comparison, in 2018 9.5% of GBP LIBOR swaps where non-cleared. This being said, financial institutions really started to embrace the new benchmark since this summer, so - of course - more time is needed and the signs are indeed quite encouraging for SONIA (as for SOFR, the jury is still out).

After also highlighting the progress made with the ISDA fallback rates consultation for SONIA, TONA & SARON, Mr Latter emphasized once again that fallback mechanisms are only a last recourse safety belt and that market participants should work on actively migrating their legacy LIBOR positions ahead of a discontinuation event.

Now, on to the juicy bit !

The Zombie LIBOR hypothesis

Picturing what the end of LIBOR might look like, we can really think of two scenarios: (1) a sudden death, or (2) a slow and painful one :

1. The benchmark administrator (IBA) announces in advance that it will cease the publication of a given LIBOR tenor
2. Contributing banks gradually pull out of the panel and the supervisor (the FCA) declares that a LIBOR index is not representative of the underlying market anymore.

Under that second scenario, we would live for a while with a Zombie LIBOR with dwindling liquidity and increasing difficulties to hedge exposures from legacy LIBOR positions.

An additional challenge with the zombie scenario is that market participants may suffer from a bifurcation in the definition of the fallback cessation triggers for different products (e.g. cash vs. derivatives) which would see legacy LIBOR positions migrate to the ARR at a different pace depending on the asset class! Hello un-hedgeable basis risks and broken hedge-accounting relationships...

How is such a mess even possible, I hear you say?

ISDA fallback triggers only consider a discontinuation of the publication!

For a start, in the new fallback language that ISDA proposes, cessation triggers aim to offer a clear and unambiguous definition of a discontinuation event applicable across relevant markets. As such, they cover the following cases only:

1. A statement by the Regulator that the Benchmark Administrator is insolvent
2. The Administrator announces that it will cease publishing a given LIBOR permanently
3. The Supervisor of the Administrator announces that the publication of a given LIBOR has ceased permanently
4. The Supervisor announces that a given LIBOR "may no longer be used"

This last point is where the FCA is raising an objection, since ISDA made it explicitly clear that "may no longer be used" can only mean an actual public prohibition on the use of this LIBOR.

In other words, statements that the supervisor "no longer supports and index" or "deems it not representative" are excluded and considered too vague (cf. point 8 of the ISDA FAQ - link in the banner below).

Here, the FCA seems to disagree - all the more that, in the US, the Alternative Reference Rates Committee (ARRC) has been running a series of market consultations on the topic of USD LIBOR fallbacks for cash products with a different view on what fallback triggers might be.

Pre-cessation triggers may cause FRNs, Securitizations and Loans to transition earlier than Derivatives!

Without going into the details of the options the ARRC laid down for the various products they are consulting on, they have introduced the concept of "pre-cessation" triggers.

This would let loans or FRNs users describe precisely and unambiguously situations where USD LIBOR-referencing instruments should automatically fallback to their alternative reference rate (SOFR), even though the publication of the index may still go on.

These options address the zombie LIBOR scenario, i.e. cases where low LIBOR liquidity would prevent institutions from efficiently hedging their legacy LIBOR positions exposures, or where the LIBOR quotes would cease to be economically representative of the underlying market.

And here Mr Latter points that - based on the responses Cash market participants submitted to the ARRC - there appears to be strong support for including such a "representativeness trigger" in the fallback mechanisms.

Unfortunately, many participants also clearly expressed their preference for aligning the fallback triggers between Derivatives and Cash products - what a predicament! 

To resolve this dilemma, the FCA pleads that ISDA revisits their triggers definition to include the case where the supervisor deems the relevant rate no longer capable of being representative - typically, following the departure of a panel bank.

How would this work in practice ?

Mechanics and benefits of a "representativeness" trigger

• The date would be announced in advance (though, perhaps with little lead time), as the FCA could anyway force a departing bank to continue publishing for up to 4 weeks.
• In any case, the FCA expects that the timing difference between a trigger based on a regulatory announcement and one based on cessation would be small, as the benchmark administrator would probably cease the publication of the rate shortly after the announcement. This is not guaranteed though, and as we will see later may not always be the most desirable outcome.
• As the IBA reduced submission policy makes it possibility that the last available value of LIBOR is published for some days before the publication ceases - this could be an issue for derivatives relying on a cessation trigger:

1. derivative contracts hedging an exposure to floating rates would be subject to an uncertain period of fixed-rate payments
2. more annoyingly, the historical mean/median spread used to determine the applicable alternative rate (i.e. the method that ISDA is working on) would include a number of observations that would "essentially be non-market fixed rate".

The proposed representativeness trigger would solve this.

One could imagine that CCPs would also favour this approach for cleared-derivatives, as it would be challenging to model Initial Margin requirements adequately in the period following such an announcement and preceding a complete cessation (with heightened risks of liquidity issues and unusual market activity).

And of course, to avoid unmanageable basis risks, it will be key to maintain an alignment of the triggers and fallback mechanisms between cleared and non-cleared derivatives (cf. LCH current supportive position on the ISDA fallback recommendations - circular No 3999 ).

Are there potential issues with the FCA proposal?

This idea of including the "non-representativeness announcement" as a fallback triggers would certainly address a number of issues, but would ISDA make such a change?

A first question is whether lawyers can confidently describe such a case in a way that would unequivocally apply to all the benchmark administrators and their supervisors facing a potential discontinuation of their index - and this in a clear and unambiguous way. The case described by the FCA can be generalized to all those falling in scope of the European Benchmark Regulation (which requires the supervisor of the benchmark administrator to make this assessment of representativeness of the benchmark, every time a contributor leaves the panel) - but how about bodies outside the EU and the UK?

Another possible issue is the risk that a key contributor to the LIBOR panel (or a group of such dealers) decides opportunistically to time its exit (aiming to trigger the "non representativeness" announcement) for self-serving reasons. This seemed far-fetched, perhaps? Yet, we know that the "mean/median historical spread" fallback method currently evaluated by ISDA can generate significant value transfers between counterparties - so, I am not sure we can safely discard this possibility. This would be quite a spectacular way to close the LIBOR manipulation chapter, though!

Hard to transition markets / products : the LIBOR show must go on ?

Towards the end of his speech, Mr latter made a valid yet slightly surprising observation about the EU Benchmark Regulation - if we consider the usual regulators views on the LIBOR discontinuation and the earlier statement that the time lapse between a loss of representativeness of an index and its full cessation should be short.

The Benchmark Regulation allows the administrator to go on publishing an unrepresentative rate for a "reasonable period of time" - what is a reasonable duration is left to interpretation...

Noting that there are markets where changing the contractual references to LIBOR is not practical, Mr Latter acknowledges that there may well still be a material volume of such legacy contracts in cash markets even into the 2030's. To manage such legacy instruments where there is no safe legal mechanism to remove the dependence on LIBOR, a solution could be to allow the continued publication of the non-representative LIBOR, but strictly restrict its usage to specific legacy instruments!

"[this approach] could help to prevent otherwise unavoidable disruption in cash markets. (...) It is important that the way critical benchmarks come to an end is not unnecessarily disruptive to markets."

E.S. Latter

Under that "tamed Zombie LIBOR" scenario:

• it would not be possible to enter new LIBOR contracts,
• fallback mechanisms would have been triggered for all instruments including a pre-cessation trigger considering a non-representativeness announcement (which may or may not include derivatives)
• the continued publication of the rate would offer a welcome relief to the market to work out a transition solution for hard-to-transition LIBOR-referencing cash products.

Indeed, there is still a material risk that a number of cash instruments remain without a proper fallback mechanism for the medium term.

In the US for instance (arguably the most mature market on the LIBOR transition topic), it is currently unmanageable to amend existing FRNs definitions to include suitable fallback conditions as the US Trust Indenture Act requires all bondholders to agree to the change! As expected, there are different legal interpretations (which could lead to disputes), which is why it is understood that the ARRC will work with the New York State legislature to offer a suitable legal framework in the future. This type of issue will certainly be found in other markets and instruments, which is a concern if contributing banks start leaving the LIBOR panels by early 2021.

Considering this situation, Mr Latter states that continuing to publish the rate after the regulator's announcement could even help the portion of derivatives contracts designed specifically to hedge the problematic cash market contracts (it is understood that it would not be practical/reasonable to go on using the deteriorated LIBOR to keep on hedging a large back-book of legacy LIBOR positions though - since it would not be possible to enter and clear new LIBOR contracts).

But how could that work, if ISDA follows the FCA advice and includes a "non-representativeness" trigger in its fallback language for derivatives, then?!

LIBOR Exit: Hard fallback vs. Soft fallback vs. No fallback ?

This would require institutions to split their derivatives books between positions hedging "fallback-problematic" cash products and the rest of the derivatives book. The latter would need to adopt the (proposed) ISDA fallback definition, but the former could not because these contracts still need to fix on LIBOR after the index has been declared unrepresentative!

For such contracts, keeping the original ISDA 2006 definition is not an option either, as we need to cover the case where the benchmark administrator announces it will cease to publish the index. Instead, they would need to use a more restrictive definition of the fallback triggers - one that is in line with the current ISDA proposal (i.e. "cessation-only" triggers)!

Leaving aside the technical and operational complexities of managing differently the derivatives hedging cash products without proper fallback clauses, I doubt this would be a workable option. A derivatives contract could hedge a problematic security for one party, but would have a different purpose for its counterpart. Therefore, it would be difficult to choose between 2 definitions of the trigger, as one of the two parties would necessarily be exposed in the scenario presented by the FCA.

Personally, I would thus favour consistency over bifurcation of the triggers definitions - even at the expense of some broken hedge relationships over the transitional period preceding the full cessation of the index publication.

-----

Looking at the bigger picture, the risks of market bifurcation upon a LIBOR discontinuation are a reality. The best cure would be aligning the fallback mechanisms (i.e. both triggers definitions, and determination protocols for the alternative rates and spreads) between cleared derivatives, non-cleared derivatives, securities, loans and other cash products.

Is such an outcome achievable before the sun sets over LIBOR?

The clock is ticking and it certainly seems a challenge considering the legal and regulatory complexities as well as the conflicting needs from different business lines (such as the call for pre-cessation and optional pre-cessation triggers in some cash markets).

Only time will tell, but I am confident the coming months will see industry and regulatory bodies take some decisive steps which will finally give shape to a LIBOR transition path that has eluded us so far. Do you share my optimism?