The switch from Libor to new benchmarks under the EU’s Benchmarks Regulation presents a huge data management challenge, but also an opportunity to digitise contracts and analyse client relationships.
The regulation was passed in December 2015 in the wake of the Libor scandal, and its deadline is December 31, 2021, when Libor will cease to be calculated and all organisations must have adopted new benchmarks such as the UK’s Sterling Overnight Index Average (SONIA).
To comply with the regulation, firms will need to locate all their contracts – these could be kept in a variety of formats, including paper and scanned PDFs, as well as electronically. Contacts in the first two formats will need to be scanned using optical character recognition (OCR) technology. Some firms are then using a simple word or phrase search to identify contracts that need fixing – for example, searching for the word ‘Libor’. This, however, is the hard way to go about things, according to Ned Gannon, co-founder and president of eBrevia, a US-based electronic contract FinTech that was acquired by Donnelley Financial Solutions in December 2018.
Gannon says: “Part of the challenge is that it’s more than just simply doing a keyword search for Libor. Some of the concepts that are important to understand in this part of the analysis can be expressed in a wide variety of different ways, using different types of vocabulary. This is where you really see the benefit of machine learning technology – it is able to identify these concepts no matter where they are buried in a contract or how they are expressed.”
For example, by using machine learning and analytics, firms can identify related concepts that are important to understand when deciding how to treat a contract or a group of contracts, such as the type of financial instrument concerned, the length of the contract, or the presence of specific types of terms and conditions.
Machine learning is also capable of identifying unclear words. Gannon says that about 80% of the documents the company typically processes are scanned images that have been converted into electronic text. If a word is smudged or otherwise unclear in the original, OCR can translate it incorrectly, meaning that a simple word-based search might overlook it. Machine learning can be tuned to pick up this imperfect data and decide if it is relevant, leading to a more complete and reliable document search for contracts that are impacted.
Becoming benchmark compliant
For many contracts where the benchmark will need to be changed, chunks of the contract language will have to be renegotiated – for example, there are important structural differences between Libor and the new benchmarks that could have financial implications, and for some contracts even the legal jurisdiction may need to be altered.
Anastasia Dokuchaeva, head of partnerships at ClauseMatch, which provides a document workflow platform, says: “Financial institutions come to us after they OCR documents and need to put them into a structured database format so that they can separate a paragraph and then start collaborating with other firms and lawyers around wording of a specific term or condition.” She adds that firms need to have this real-time editing capability today, as well as an audit trail that records the changes that were made, why they were made, who signed off the changes, and who approved the final contract. An audit trail is becoming increasingly important to financial firms that are being asked for this kind of evidence by regulators more and more as part of their focus on conduct risk and ethical client outcomes. For Libor repapering, firms need to be able to demonstrate to regulators that they handled the process fairly.
Looking to the future
Firms can reap benefits from digitising contracts and applying a data governance framework to them, according to Gannon. For example, they will be able to adapt to regulatory change with more agility in future as, in minutes, the business or compliance team will be able to use contract analytics to see all of the documents that are impacted by a rule change. This will dramatically increase efficiency, reduce compliance risk and lead to better client outcomes.
As well, firms can drive reporting analysis out of the digitised documents for management purposes. Says Gannon: “We find, particularly in corporate legal departments, that people oftentimes don’t know what’s in their contracts. Sometimes they are paying for things they no longer use, or they are unaware of risks. Maybe they are missing out on revenue opportunities. Just having the contracts digitised can give people better insights into their relationships with customers, vendors, and strategic partners.”
Firms with digitised contracts will be able to analyse contracts in bulk and export the information into Word, Excel, a contract management database, or another type of database. This allows teams, by way of example, to use information about existing derivative contracts or loan contracts in detailed credit risk modelling, or for new product development.
Standardising terms and conditions
Firms may also want to consider using a database of digitised contracts to standardise some of their terms and conditions as they are finding they have ‘70 versions of the same clause in 70 different ways’, says Dokuchaeva. “Why not have one master version of that clause and, yes, maybe a variation of it in certain special cases,” she adds. The clause can become a unique data object in itself, and contracts could be built out of clauses as individual data objects. Such contracts could also be approved clause by clause, making contract negotiation simpler and helping firms to better manage the risks associated with contracts.
So, although benchmark reform poses some significant compliance challenges, the digitisation of contracts also presents interesting opportunities to rethink an important aspect of client engagement, and to gain new insights about client relationships from contract analytics.