U.S. banks are looking to automation to aid in the yearend phaseout of the London Interbank Offer Rate (Libor) and to comply with a not-so-friendly push from the Federal Reserve to do so by Dec. 31.

Libor is the most widely used benchmark by which financial institutions determine their currency swaps, loans, mortgages and credit card rates. Although regulators issued a final date of June 30, 2023, for the phaseout on existing transactions, the Fed is telling banks to refrain from using that date on new transactions.

Randal K. Quarles, vice chair for supervision at the Federal Reserve, noted that, “Continued use of Libor in new contracts after 2021 would create safety and soundness risks, and we will examine bank practices accordingly,” when speaking before the U.S. House of Representatives’ Committee on Financial Services on May 19.

The proverbial cart before the horse

The automation challenges for banks with regard to the Libor phaseout are many, Ronak Doshi, vice president at IT research firm Everest Group, told Bank Automation News. These include changing how documents are processed and risk-managed — even on new transactions — and the added complication of banks possibly opting for competing reference interest rates like the Secured Overnight Financing Rate (SOFR), American Interbank Offered Rate (Ameribor) and Bloomberg Short-term Bank Yield Index (BSBY).

“There are several numbers on the market right now — $300 trillion to $350 trillion worth of contracts, which are still outstanding in Libor,” Doshi told BAN. Before the tech can be implemented, banks are “identifying where they are using Libor contracts and going through all of the documents,” he added. Renegotiation of interbank contracts, mostly based on Libor and extending beyond 2021, require the human touch, making it a resource-intensive process.

Bots are being employed to update calculations, but they do not provide a perfect solution. “As these rates are changing, or the sources of where you collect this data is changing, bots need to be updated, which is costly,” Doshi said, adding that there is also the chance bots could fail in the production environment.

The automation edge

However, cognitive document processing, such as digital signature document verification tools, as well as chatbots and virtual agents to process frequently asked questions, all have a place in the phaseout, Doshi told BAN, but added he thinks “none of the automation vendors are offering this as a complete solution.”

No matter the solution, the difficulty for banks and, consequently, vendors remains with banks adopting differing rates, David Musto, professor of finance and director of the Stevens Center for Innovation in Finance at the Wharton School of the University of Pennsylvania, told BAN. “The anointed one, SOFR, doesn’t track credit risk, and so people are looking for a benchmark that does track credit risk.”

That bifurcation of the market is already apparent. According to a May 3 Bloomberg News report, Bank of America and JPMorgan Chase “struck the first swaps trade,” a $250 million one-year basis swap with one side tied to the BSBY index. The $82 billion Zions Bank also announced it intends to use Ameribor in most of its commercial loan contracts beginning this summer.

Covering the bases

Many fintech vendors are gearing up for various reference rates. Lending and risk management solutions provider Automated Financial Systems (AFS) has already announced its software products, AFS Level III and AFSVision, are ready for all four SOFR rate methods, including credit-sensitive spreads and compounding rate and balance, as well as processing multiple alternative rates, including Ameribor, BSBY, Fed Funds and Prime Rate, said Dean Snyder, executive vice president of business solutions at the company.

“We have been upgrading our commercial loan servicing software so that it can support the alternative rates,” Synder told BAN. “The predominant one, of course, is the SOFR rate, which required some new ways to calculate interest on commercial loans.” He noted that AFS had to develop new capabilities within their APIs to support the new interest rates and calculations. AFS counts the $518 billion Truist Bank, $126 billion Huntington Bank and $90 billion Santander Bank as some of its clients.

Strangely enough, the phaseout of Libor has brought competing stakeholders to the same table, noted Snyder. AFS is a part of the Alternative Reference Rates Committee (ARRC), a group of public and private market participants, including major fintech vendors, banks, asset managers, insurers and industry trade organizations, brought together by the Federal Reserve Board and the New York Fed to ease the transition to a new reference rate. ARRC is recommending that the financial services industry adopt SOFR.

“The challenge was getting the industry to get together and agree on standards for how these calculations should work,” Synder told BAN. “We compete with the other major vendors but, at the same time, our systems have to communicate to each other, so we need to make sure we’re all on the same page.”

The next steps

AFS’ development is complete, and Snyder believes most of the other major vendors have developed new code capabilities. Now, it’s a matter of the banks doing the implementation. “SOFR is not necessarily going to be the predominant replacement,” he said. “So that’s kind of thrown a wrench into it.”

Credit spread adjustments are needed to change contracts from Libor to the new reference rates, which could be particularly complicated if banks are using more than one rate. Even though one of the biggest challenges does come from the fact that “global lending rates could be calculated more dynamically, banks can use automation to calculate the rates on a real-time basis,” Rajesh Agarwal, senior vice president at IT software and consulting firm Datamatics, told BAN. Datamatics currently deploys robotic process automation, intelligent document processing and artificial intelligence (AI) for its banking clients, with an “intelligent solution suite” that can identify at-risk documents, then automate their selection and classification without the need to employ large teams of people.

Agarwal noted that, even though banks still use paper-based processes, automation “technologies like intelligent [ocular character recognition], AI and [natural language processing] can help banks to identify and classify the documents where the Libor rate has been applied.” Of course, documents will still need rewriting or editing to comply with the new terminology and rates, he said.

The final hurdle

At the end of the day, the financial institutions themselves remain the most complicating factor for the phaseout. Everest Group’s Doshi pointed to the expense and employee hours required for banks to comply. A report from EY indicates that the budgets of most global banks’ to replace Libor exceed $100 million, with the bulk of that going to technology costs.

Despite the significant outlay expected, banks may still not be well positioned for the phaseout. An April report from financial consultant Duff & Phelp indicated that more than 50% of respondents to its survey from financial services firms “had not set a date to cease new LIBOR linked issuance. In addition, 5% expected to be unable to meet the December 31 deadline.”