Skip to content

Search the site

LIBOR laggards win a "synthetic" reprieve from the FCA. Wait -- Artificial Intelligence didn't fix it?

Nobody wants "widespread market disruption and a disorderly cessation"

The UK's Financial Conduct Authority (FCA) says it will compel continued publication of six sterling and yen Libor versions after the end of 2021 using a "synthetic" methodology to "ensure an orderly wind-down", amid concerns that there are over 235,900 tough legacy contracts worth £470 billion that reference the sterling Libor versions alone still sitting untouched on banks' books -- despite years' of notice from regulators.

Libor, one of the main interest rate benchmarks used in global financial markets, is a measure of the average rate at which banks are willing to borrow wholesale, unsecured funds. It underpins financial contracts worth trillions of pounds, including derivatives, bonds and loans. It is being killed off at the end of 2021 in a long-flagged move by the FCA and BOE. (The latter had suggested in March it would show some flexibility.)

Follow The Stack on LinkedIn

The Libor scandal triggered robust action from regulators. Yet despite the the phase-out beginning in 2017 and Bank of England Governor Andrew Bailey warning in March 2021 that "with limited time remaining, my message to firms is clear – act now and complete your transition by the end of 2021" the FCA said September 29 it had identified billions-worth of contracts still referencing short-term sterling Libor versions, including:

  • 35,000 outstanding syndicated and bilateral loans worth "at least" £350 billion
  • 659 outstanding securitisations valued at over £57 billion
  • 278 outstanding bonds with a total value of around £28 billion.

"The Authority expects that while it is possible to amend these contracts, some may be difficult to amend in the time available before end 2021, particularly the more complex multi-currency, large syndicated loans. In addition, given the large numbers of loans, there could be market disruption caused by all counterparties undertaking the consent and / or renegotiation process at the same time," the FCA said sympathetically.

The FCA added that it "considers that the existence and the scale of these contracts referencing the 6 LIBOR Versions means that widespread market disruption and a disorderly cessation would likely be caused, if the Authority did not exercise its Article 21(3) power to compel IBA to continue publishing the 6 LIBOR Versions."

Synthetic Libor: "Not indefinite..."

In the wake of the widely reported Libor rigging scandal, the FCA stripped responsibility for its supervision away from the British Bankers Association  and turned it over to the Intercontinental Exchange's Benchmark Administration (IBA) -- a subsidiary of exchange operator, Intercontinental Exchange (ICE).

ICE said today that it had been told to publish the 1-, 3- and 6-Month sterling and Japanese yen Libor settings under a synthetic methodology* while those with legacy contracts continue to get their houses in order. (The FCA toyed with the idea of having Libor panel banks continue to contribute data to sustain the six Libor versions in question; few were keen and rather than force them, the new synthetic Libor route was settled on, it suggested.)

Users of Libor "should continue to focus on active transition rather than relying on the synthetic replacement, which will "not be published indefinitely" ICE noted of the FCA's move.

As Genpact's Anu Sachdeva earlier told The Stack's Ed Targett: "the transition from LIBOR to alternative rates poses an enormous task. Banks need to revise contracts maturing beyond 2021 to incorporate fallback provisions, terms in case LIBOR is unavailable, or to transition to an alternative reference rate.

"While fallback provisions may exist in some contracts, it’s likely that they were designed to address the temporary unavailability of LIBOR, such as a computer systems glitch or a temporary market disruption, rather than permanent discontinuation of LIBOR and would therefore need to be revised. Banks face extensive and costly administrative work to change contracts, update computer systems..."

"When they tell me “speed!” I need to think speed, and scale, and security, and stability" -- JPMorgan Global CIO Lori Beer

KPMG is among the organisations that have been urging banks to see the colossal contractual revisions needed as a golden opportunity to update their systems, add AI-powered document discovery, etc.

"Ultimately the LIBOR transition opens the door to a fundamental rethink on the use of data and machine-learning techniques to better provide products and services to customers" KPMG's Traci Gusher and James Lewis noted back in 2019.

"We are advising clients to view the current Libor challenge as a 'once-in-a-lifetime' opportunity — a catalyst to making major advances toward innovative digital solutions that will transform their front-to-back infrastructure and processes."

Many have indeed seen it as a chance to overhaul governance, contract identification, inventory of systems, infrastructure and functions and test the ability of AI and Matural language processing (NLP)-based tools to help automate back-end processes. But at the coalface there's been no silver bullet, to mix metaphors. Even with automated document discovery, complex contracts with multiple counterparties can not be "AI'd" away and as the FCA's shown today, with £470 billion's worth of contracts still exposed to Libor, there's a lot of work still to do.

*For 1-, 3- and 6-Month sterling LIBOR, the relevant ICE Term SONIA Reference Rate provided by IBA, and for 1-, 3- and 6-Month Japanese yen LIBOR, the Tokyo Term Risk Free Rates (TORF) provided by QUICK Benchmarks Inc. (adjusted to be on a 360 day count basis); plus the respective ISDA fixed spread adjustment (that is published for the purpose of ISDA’s IBOR Fallbacks Supplement and Protocol) for each of these six LIBOR settings

Has your bank made use of innovative tech to help tackle the Libor transition? We'd like to hear from you. Get in touch with editor Ed Targett here.