The London Interbank Offer Rate (LIBOR) is a series of benchmarks that underlies $240 trillion of financial products globally. LIBOR was established as the average interest rate at which institutions lend to one another and has been used as a benchmark for everything from interest rate swaps to consumer mortgages, commercial lending, and business loans.
Following the 2008 financial crisis, the frequency and scale at which institutions lent to each other diminished significantly, and in 2017, the UK Financial Conduct Authority announced by 2021 it would no longer require banks to submit the information required to calculate LIBOR. As a result, LIBOR will no longer exist and trillions of dollars in financial contracts relying on LIBOR or its overnight equivalents must be revalued using alternative benchmark rates.
In April 2018, the Federal Reserve began publishing SOFR as an alternative to USD-LIBOR. One month later, the Bank of England began publishing reformed SONIA as an alternative to GBP-LIBOR. Since these alternative rates published, a concentrated number of the world’s largest banks have traded futures based on these new benchmarks, supported by institutions such as the Intercontinental Exchange (ICE), London Clearinghouse (LCH), and the CME Group.
The transition from LIBOR to alternative rates will generate significant risks for financial institutions. Fall-back clauses in existing contracts were often designed for situations where the benchmark is not available for a single day, not for a complete benchmark cessation. Switching to an alternative benchmark can generate significant gains or losses to a financial institution depending on the alternative rate used and the duration of the fall-back clause period. Financial institutions worldwide must revalue entire books of business using alternative rates, a massive undertaking that may dramatically change balance sheets for thousands of financial institutions. Artificial intelligence provides a powerful approach to solving the LIBOR transition crisis and allowing the global financial system to meet the 2021 LIBOR deadline.
Synechron’s Data Science Accelerator for LIBOR Impact Analysis enables financial institutions to identify and quantify their LIBOR exposure at either a contract level or across all contracts within an institution. It does so by aggregating information that already has been captured in bank systems with explicit fields for underlying benchmarks, but also adding information generated from contracts that only exist in less structured form such as confirmations, master agreements, and more.
The Accelerator then allows a financial institution to revalue a contract or set of contracts using alternative rates and valuation models, rapidly allowing an institution to quantify its exposure and potential gains and losses for a contract or book of business under various interest rate assumptions. Alternative rates already agreed upon in the underlying contracts, discovered by the Accelerator, could be presented as options to choose.
This powerful solution automates the time-consuming and potentially error-prone process of identifying LIBOR terms and provides significant flexibility in applying different valuation techniques and interest rate curves to accurately calculating the value of a contract. When a choice has been made (and agreed with the client) about an alternative benchmark, both the information captured in bank systems and the underlying legal documentation must be amended. This Accelerator will help keep the two connected.
The Accelerator combines a powerful combination of Optical Character Recognition and Natural Language Processing to identify any reference to specific benchmarks (e.g. LIBOR) and data attributes contained in a contract and to return those results in a structured form. The NLP model analysis goes beyond simple keyword recognition to understand the intent, content, and materiality of that reference gaining meaning from a complex, proprietary financial services lexicon developed by Synechron.