LIBOR transition: How NLP can help you overcome the challenge

Amid a global pandemic and economic crisis, banks, asset managers, insurers and financial firms are still working towards the LIBOR transition deadline of the end of 2021.

While the Financial Conduct Authority (FCA) in the UK pushed back the cut-off date for the use of LIBOR in new loan agreements until the end of Q1-2021, it is still the expectation that LIBOR will cease from the beginning of 2022. Immediate priorities have changed as COVID-19 plunges businesses, industries and markets into turmoil, but financial firms must continue at pace with their transition plans to ensure compliance.

The phase-out of LIBOR, which underpins around USD 350 trillion in assets globally, would present a huge logistical undertaking at the best of times. Now with additional pressures placed on firms, it’s a truly daunting task that exposes banks to further economic risk if they fail to comply.

Many firms have thousands of documents that reference LIBOR in some way – which all need to be repapered or updated in line with the new regulations. Before the remediation process can even get underway, the right documents need to be identified and then sorted based on the type of remediation required. Remediating contracts for the LIBOR transition is essentially a three-step process:

1. Find and flag all the contracts that need remediation.
2. Identify the type of remediation required and bucket the contracts accordingly. For example: repapering, renegotiation, wind-down etc.
3. Carry out the remediation work.
   

Using natural language processing (NLP), it’s possible to automate and accelerate the first two steps of the process. For the uninitiated, NLP is a branch of AI specifically focused on processing, understanding and interpreting the human language in ways that are valuable to an end-user. NLP has many applications; one of them is analyzing large volumes of unstructured, siloed text, such as contracts, to scale up data extraction and speed up processing times.

Using NLP, salient information is extracted accurately in a fraction of the time it would take to review and process the documents manually. NLP can significantly speed-up compliance with regulatory changes now and in the future. It’s a scalable solution that, when integrated with your existing document management system(s) and downstream systems will reduce the future cost of data reconciliation while also improving access to data.

How does NLP for LIBOR work?

NLP relies on machine learning to teach the system the context, nuances and relevancy of the language contained in different documents. A sample of contracts are uploaded to the system, and a user then labels the relevant data fields or sections within these documents. This information is used to build a machine learning model. More documents and labelling may be required to improve accuracy and attain the desired accuracy rate. The model then analyzes all contracts to retrieve the necessary data points identifying those that contain certain references that will need remediation, as well as picking out the particular clauses that need to be reviewed.

With NLP, a firm can automate the following LIBOR transition processes:

• Review all documents and contracts in its legacy book
• Identify those that reference LIBOR
• Identify those that mature post-2021
• Flag for further action problematic ‘fall back’ language
• Extract fields from the contacts that need further work or attention

Why choose Eigen for LIBOR?

We’ve gained considerable experience supporting clients with their LIBOR transition as well as being embedded in the front, middle, and back office at some of the most respected financial institutions in the world.

Our clients are currently using our NLP solution for analysing documents and contracts

affected by LIBOR, including Syndicated Loans (USD-LIBOR), interest rate swaps, forward rate agreements, interest rate options and cross-currency swaps (USD LIBOR; GBP LIBOR).

And we’ve been awarded the Financial Times Intelligent Business Award for Repapering Technology in recognition of our work and expertise.

For more information about Eigen for LIBOR, you can read this case study.