Artificial intelligence suits up as LIBOR retires
AI can help financial institutions manage the legal risk associated with the LIBOR transition, says Anu Sachdeva, global service line leader for commercial banking at Genpact.
AI can help financial institutions manage the legal risk associated with the LIBOR transition, says Anu Sachdeva, global service line leader for commercial banking at Genpact.
The clock is ticking. In 2021, London Interbank Offering Rate (LIBOR), the benchmark interest rate that underpins trillions of dollars of loans globally, is set to expire. But LIBOR is hardwired into almost all financial contracts that have a variable interest rate component.
As a result, banks face extensive and costly work to review and revise a breath-taking volume of contracts. The time to prepare is now. Saying goodbye to LIBOR won’t be easy, but artificial intelligence (AI) can help banks automate the transition and mitigate the financial and legal risk associated with contracts maturing beyond 2021.
The transition away from LIBOR consists of three phases:
The first step is for banks to evaluate all contracts and assess which of them are linked to LIBOR. This is not a simple ‘yes/no’ question. Instead, this analysis is complex and requires deep domain expertise and an understanding of asset classes. For example, in commercial loans, a credit facility with purely US dollar loans has different implications than a credit facility with LIBOR-linked loans in other currencies.
And a fixed-rate loan contract that is ostensibly not linked to LIBOR may contain an interest rate derivative linked to it. Banks must develop a comprehensive set of contract review questions that addresses these complexities.
Although a replacement rate has not yet been identified, banks must start now to incorporate interim amendments into their contracts. For example, if loans exist that do not have any fallback language, contract terms in case LIBOR is unavailable, banks can incorporate a soft fallback option.
If fallback provisions do exist, 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 flagged and revised.
Banks can also start to develop contingencies. For example, although LIBOR serves seven different maturities ranging from overnight to one year, the replacement rate might not mimic the same tenor structure. So banks can plan for contract changes that should be implemented in this case.
Soon, there will be clarity on the replacement rate, its characteristics, and how it will be operationalised for each asset class. In this phase, the bank must amend contracts, update systems and processes to procure and test data feeds for the new rate, and train staff to address the needs of the new rate.
The effort and cost involved in manually reviewing financial contracts is huge. However, an AI-assisted review process is accurate and efficient, augmenting what humans can do and freeing them up to take on higher-level tasks such as communicating with customers and negotiating deals. AI is well equipped to perform contract reviews associated with the LIBOR transition.
In fact, AI solutions already exist to minimise legal risk, reduce costs, and boost governance with respect to contracts. Such solutions have been applied to interpret tens of thousands of credit agreements and amendments across industries and demonstrated benefits such as 80% improvement in process efficiency and 70% less time spent on processing documentation. These same solutions may now be applied to the LIBOR transition.
Using applied computational linguistics, pattern recognition, and machine learning, AI can extract contract terms and validate them against the contract review questions and other data. For example, AI, specifically natural language understanding, can identify whether LIBOR is being referenced in the loan and direct humans to the exact locations in the loan documentation that cites LIBOR.
Using machine learning, AI can learn to identify the numerous permutations of phrases that cover fallback procedures and whether they are sufficient for the permanent discontinuation of LIBOR. And AI can validate that the amendments processed are reconciling with the guidelines given by the front-office.
Achieving positive outcomes from AI investments
According to the second edition of Genpact’s research in AI, the banking sector stands out as the top spender in AI technology, but is only average in achieving very positive outcomes from AI. This may be because banks are not leveraging or using it in the right way. Applying AI to the LIBOR transition is an ideal way for banks to harness the technology’s power to drive positive outcomes. An AI solution to the LIBOR transition affords:
Getting started
Closing the door on LIBOR won’t be easy. And the transition from LIBOR to an alternative rate has the potential to cause havoc if not handled properly. But AI can help financial institutions get moving to manage the legal risk associated with the LIBOR transition. With AI and the right partner, banks can turn struggle into strength, challenge into opportunity.