A new start: regulators want Libor phased out by December 31 2021 © Leon Neal/Getty Images

Frequently described as the world’s most important number because it underpins trillions of dollars of transactions, the London interbank offered rate (Libor) has persisted until now despite a scandal that caused lasting reputational damage to the entire financial system.

Libor is the key interest rate benchmark for mortgages, loans and contracts but it has been tainted since 2012 when it emerged that banks had misstated their Libor rate submissions, often in collusion, to make better returns. 

The controversy led to at least five traders going to jail in the UK, and US and UK regulators extracting penalties totalling about $10bn. Regulators want Libor phased out by December 31 2021, and banks are pivoting to alternative risk-free rates such as Sonia (sterling overnight interbank average rate).

Its demise is already a headache for law firms and banking clients, which must examine hundreds of thousands of legal contracts containing references to the Libor rate and then rewrite and “repaper” them to ensure they include the new reference rates. Contracts must also be analysed for “fallbacks”, the legal rules that spell out what would happen if Libor ceased to exist. 

‘Defcon 1 litigation event’

The scale of the repapering exercise is dubbed “immense” by those involved and last year was publicly described as a potential “Defcon 1 litigation event” by Michael Held, general counsel at the New York Federal Reserve.

This has led many law firms and their bank clients to turn to artificial intelligence (AI) technology, often provided by start-ups, which can review large numbers of documents using natural-language processing to identify legal clauses and obligations. Harnessing technology to do the grunt work means banks can avoid paying for armies of sleep-deprived junior lawyers and paralegals to sift through contracts, which in some cases may still be paper documents.

“It is possible to train AI to look for Libor or how Libor is referenced in various guises and to do that in various languages,” says Deepak Sitlani, head of the derivatives and structured products group at law firm Linklaters in London, who says some clients have developed their own technology tools.

“This is taking up a lot of resources and if you are a bank it’s a massive project,” says Adam Ryan, chief legal innovation officer at law firm Freshfields.

Lewis Liu, chief executive of Eigen Technologies, a natural-language processing company that works with banks such as ING and Goldman Sachs, estimates that about half its work this year is from clients using technology to help ease the switch from Libor.

Eigen’s technology helps automate the process of finding and flagging the contracts and then identifying the types of remediation needed. However, the technology cannot write a new contract or carry out repapering work.

No more grunt work: Eigen Technologies uses artificial intelligence to search contracts for references to Libor

Mr Liu says banks are at different stages of readiness, despite the regulators urging them to make progress on leaving Libor. “I know of one big US investment bank [that] has completed its programme and another bank [that] has not yet started the work,” he says. One bank has even opted to review contracts manually with the help of lawyers rather than use technology.

For a large wholesale bank, he estimates, it could take 1,000 lawyers more than two years to identify and switch over all its contracts manually, whereas technology can do this in four months with 20 paralegals and lawyers. The attraction is not just speed, he adds: “Banks are using it as an opportunity to do large-scale digitalisation of their documents,” he says.

In the past few years AI has been deployed in litigation cases to help large companies scan databases or email archives to find a particular word or search term, and then apply relevant changes. 

Charlie Connor, chief executive and co-founder of US-based Heretik, says machine learning is being used in a sophisticated way for the Libor transition.

It enables a 700-page Libor contract to be searched and even to pick up punctuation and the part of the sentence that would determine the remediation strategy.  Contract disputes in the past have turned on punctuation such as a misplaced  Oxford comma  — a mark that comes before the “and” or “or” at the end of a list but which can change the meaning of a sentence.   

Mr Connor says the Libor transition is making banks digitise paper contracts that might be scattered in various offices. “Banks and financial institutions are being forced into digitisation through Libor and can see its benefits,” says Mr Connor. “The software is very effective and it means law firms will be doing less mundane tasks and more higher-quality work, which is what counsel is paid for.”

He says doing this work now means that the same AI technology can be used to examine contracts en masse for other financial risk factors, such as negative interest rates. “You can run it to look at the effect of negative interest rates which could be impacted by, for example, Brexit,” he says.

Some lawyers suggest that while the biggest effect of the Libor transition will be on leading banks, other non-financial services companies could also be involved if they use Libor-related business contracts to buy and sell goods, for instance, for late-payment clauses or cost increases in long-dated contracts.

If no replacement for the Libor rate has been specified in a contract, the two parties would then have to decide what other benchmark to use. “If there is money involved there might be a fight about it,” says one lawyer, who declines to be named. He also suggests that banks could start to use the replacement of Libor as an opportunity to reopen older contracts with customers and renegotiate other terms.

For now, it is clear that the task of extracting the industry from using Libor cannot be left by banks and corporates till next year, even if they are already grappling with more immediate priorities such as the coronavirus pandemic and the departure of the UK from the EU.

The case studies below are a shortlist of entries to the FT Intelligent Business awards event held online on November 19, where the winner of the Financial Services award was announced.

All the entries showcase the combined use of data and tech in business operations. Source: RSG Consulting

Financial services

© Alamy Stock Photo

WINNER:
ING Bank and Eigen Technologies

In 2020, Dutch bank ING launched its Saber Data Extraction Platform, which uses natural-language processing technology from artificial intelligence company Eigen Technologies to extract information from documents. The platform accounts for the nuances and inconsistencies in human language and was first used in ING’s interbank offered rate (Ibor) benchmark transition.  

Of thousands of documents reviewed, 80 per cent required no further review and there was a 75 per cent reduction in overall review time. ING Bank and Eigen Technologies estimate their collaboration cut the cost of this review process by 60 per cent.

D2 Legal Technology

A number of investment banks have implemented the legal data consultancy’s new tech-enabled due diligence service to identify its counterparty in regulatory capital calculations. Previously, experts in investment banks would conduct this research independently, even though many were searching for the same information.

By offering this service to banks, D2 Legal Technology has built a global database with the information, completing anonymous searches. D2 says the solution reduces the cost of the process by half. 

Dürr and Targens

Dürr, a mechanical and plant engineering company, last year raised a €750m syndicated loan with the support of 13 international banks. The loan was raised through a blockchain platform developed in-house, using a service from Targens, a German consultancy, that creates digital identities for business-to-business transactions. Raising a loan digitally reduces the time required to establish legally binding contracts with multiple parties.

NatWest and Nuance Communications

NatWest bank fields 17m calls from customers every year. For security checks on these calls, the bank previously relied on “static” data for identification (such as addresses or mother’s maiden name), but such details are easily stolen online. The bank’s fraud-prevention team is using voice recognition technology created by AI company Nuance Communications to catch known fraudsters impersonating customers.

NatWest is starting to use biometric data to identify customers, removing the need for static data. Through its use of a variety of technology, the fraud-prevention team spots fraud before customers do in 80 per cent of cases.


Financial Services category research and award supported by Ashurst

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments