The Complex Regulation That Can Revolutionize Risk

As part of its implementation of the Dodd-Frank law, the Federal Reserve is regularly collecting a large amount of granular information on corporate loan portfolios, a practice that’s causing short-term headaches for data managers, but in the long term should provide benefits for risk managers.

The Fed's new form, FR Y-14Q, requires banks to disclose transaction data for commercial and industrial loans (C&I), small to medium size business loans (SME) and commercial real estate loans (CRE). Any bank holding company with more than $50 billion in consolidated assets must provide quarterly reports that include detailed information on loan portfolios. That information is used as part of supervisory stress test models and to measure actual vs. forecast information.

The problem is that for most banks, the required data either resides in different data sources, or is populated differently in separate data sources. These different data sources can include financial statement spreads — programs that analyze financial statements to track performance over time — and the institution’s loan accounting system, which collects data on payments and transactions.

“Institutions now have to, on a quarterly basis, submit loan-level information to the Fed that includes not only the name of the borrower and loan exposure, terms and conditions, but also [financial data] associated with those borrowers. And sadly, at most institutions the loan accounting system and financial spread analysis has been siloed and isn’t integrated together,” says Brad Saegesser, a director at Moody’s Analytics. Moody’s KMV RiskAnalyst provides financial statement spreading, along with competitors such as Tyler Analytics and Peldec Decision Systems.

The Fed is collecting data on bank holding companies’ asset classes (types of businesses or loans), net revenue before loss provisions, securities risk, retail risk, wholesale risk, private equity, regulatory capital instruments, along with other Basel III and Dodd-Frank disclosures.

Saegesser says that in cases where this information is located in different systems, or is populated differently, there’s not an easy technology solution to integrate the data quickly or to ensure accuracy without manual inspection. “There is no magic silver bullet to do that…you’re talking about two different systems that have different identifiers for the same company when that company appears in each system,” he says.

While there are consultants working to help banks automate the matching of company names or industry categories in the two systems, Saegesser says these are not perfect solutions.

“How many different ‘First Baptist’ churches are there in a bank’s portfolio? The matching exercise today is one of the biggest challenges. It’s a back fill type of exercise,” Saegesser says.

Saegesser says one suggestion is to implement standards for identifiers or other ways of locating companies or information on loans — i.e. company “a” would be identified in a specific way across an enterprise, which would make it easier to verify the financial data tied to that company, regardless of where that company’s data was located.

“Banks are using some name-match rules where they look at fields within systems that may match up, and are using some data management technology to [find the same] customer in two different systems…but even once the system [finds a match] you still have to manually verify it,” he says.