The language of the contract holds mortgage lenders hostage

As reported in the National Mortgage News in May, the Federal Housing Finance Agency announced a comprehensive Libor transition plan for Fannie Mae, Freddie Mac and the Federal Home Loan Banks for residential mortgages. While the Libor deadline ends in December 2021, lenders have been advised to stop pricing loans against the benchmark by September – leaving less than 350 days for the transition.

However, it seems that not everyone is ahead of the game. According to Moody & # 39; s Investors Service, while many financial institutions are prepared, their customers are not – borrowers are "very passive" about the transition.

Libor is deeply ingrained in many mortgage contracts, particularly in the specialty loan market, MBS and storage areas. Replacing one sentence with another can mean changes to the clauses and a whole new variation in terms that must be approved by both the lender and the customer.

According to a recent analysis, Libor serves as the reference for more than 100 million contracts worldwide, including mortgages, representing over $ 400 trillion. Worse still, two out of five of these contracts do not contain any language related to setting the benchmark and therefore need to be corrected before Libor expires. If financial institutions poorly manage the transition or fail to fully address the problem, they risk improper payments and potential fraud fees, as well as business disruption and loss of competitive advantage.

Mastering challenges

You may wonder how a seemingly small change in protocol can turn the industry upside down, panic, and tens of billions of dollars reviewing contracts and finding solutions to remedial action. The answer is easier than you think – it's because the language of the contract, the place where Libor is encoded into the lending business, is firmly anchored in all lending rate setting practices.

Searching this digital mountain of documents to search for a specific contract language requires trained legal specialists, attorneys, and regulators whose time and money are simply beyond fathom, let alone a high risk of human error.

In fact, Momenta Group has now warned that lenders far underestimate the number of experts required to tackle the complicated contract transition – with a shortage of up to 250,000 skilled workers. It is estimated that a simple individual contract will take at least an hour to fix.

By applying artificial intelligence and machine learning to the complex Libor correction process, digital intelligence enables lenders to implement a smooth process of reviewing hundreds or thousands of documents, extracting Libor-related entities, and forwarding identified documents to the legal team. With the tedious, time-consuming, and error-prone task of sifting through mountains of documents being processed for them, legal professionals can focus 100% of their efforts on applying their expertise to the task of cleanup.

If ever there was a use case for AI with ML

Let's talk about the contract language and how it holds lenders hostage. The billion dollar question is, why are consultants needed, and how can artificial intelligence make a difference in machine learning? The answer is simple, even if the solution is sophisticated. As I mentioned earlier, Libor is embedded in the tens of thousands of contracts and agreements through which lenders do business with each other, with client companies, and government regulators.

While AI has been hyped with ML in recent years, AI creates computer programs (also known as computer vision) to help them find, interpret, make decisions, and take action on a corpus of data or documents are embedded as expert or highly skilled users would. AI with ML is not a substitute for high-paying professionals such as lawyers, paralegals, analysts and consultants, but rather helps them adapt to global challenges like the Libor transition.

One particularly useful application of AI for Libor transition is recent advances in named entity extraction, which AI-based programs can use to learn how to recognize legal entities in gigabytes of contracts and related documents. An entity is a specific person, place and / or action that consists of several data fields that can be found anywhere in a contract. Often implied and never in a predictable location, named companies can be especially annoying for legal professionals when entering into contracts.

Libor rates, clauses, and affected parties are all units in contracts that are rarely expressed in the same way in a single contract. However, machine learning AI can learn all permutations of these entities, including any specified and implied references to such entities. You can identify and understand them on a machine in seconds, while a legal professional can take an hour or more to complete the same review. But AI doesn't get tired or distracted the way humans can and can deliver results more consistently across thousands of documents.

With recent advances in named entity detection and extraction, such as the ability of data scientists and legal professionals to train the software with extensive taxonomies and variations (thesaurus, legal codes, aliases), modern AI solutions can be in production in a few days or weeks at a fraction of the cost of hiring an army of advisors and legal teams to get the job done. Libor reorganization experts can now train their "virtual legal team" for AI / ML software and support RPA bots as an army of expert assistants to find, reorganize or flag all affected contracts and legal entities for additional review.

This is where AI can shine with ML as an expert solution for the Libor transition. By adding and learning to work as an army of virtual assistants, a specialized named entity recognition solution supported by modern AL with ML can save the time to find, understand, and correct the sheer volume of Libor-related documents from dozens of people cut years to weeks or months of computing time. The amortization of risk mitigation, time and money can speak for itself.

Related Articles