OFR and Partners Hold Contest to Link Identifiers to Legal Entity Identifier

A messy but important problem for OFR researchers and data experts is linking information about a financial services company from diverse sources. Regulators, investors, and the public face the same issue.

For example, is the “J.P. Morgan” in a financial contract the same entity as “JPMorgan Chase & Co.” cited in an academic journal? Does either use the “JPM” stock ticker symbol? Is there a connection to “JPMorgan Chase Bank” and its identity code of 852218 in the Federal Financial Institutions Examination Council’s (FFIEC) database?

The OFR helps develop techniques to make these types of connections faster and more efficient. In particular, we want to make it easier to link the many existing identifiers to the legal entity identifier (LEI). An LEI is assigned to each legal entity in a financial transaction. The OFR played a leadership role in developing the LEI system after international regulators endorsed the concept in response to the 2007-09 financial crisis. The Office also helped establish the nonprofit Global LEI Foundation, which oversees the LEI system.

This year, the OFR and the National Institute of Standards and Technology (NIST) took a first step toward solving the problem of linking company data across datasets. They launched a series of challenges to attract solutions from experts in academia and industry. NIST awarded the University of Maryland a grant to administer the challenge. Eight teams participated in the first Financial Entity Identification and Information Integration Challenge.

Each participating team began the challenge with the same financial entity identifiers from three datasets:

1. Research, Statistics, Supervision, and Discount (RSSD) codes that banks use in filing FFIEC regulatory reports;
2. Central Index Key (CIK) codes that reporting firms use when filing corporate financial statements for the Securities and Exchange Commission; and
3. LEI codes required by many domestic regulators to buy or sell financial swaps.

Subject-matter experts at the OFR manually found a representative sample of possible matches of RSSD-LEI codes and RSSD-CIK codes. The analysis revealed, for example, that the vast majority of RSSD-identified banks have not yet registered for LEI codes. Most were small banks.

Each team in the challenge was given three assignments. First, match the RSSD identifiers to the LEI codes. Second, match the FFIEC entities to the CIK codes. Third, align entities across the three datasets. Teams were free to use other databases and information sources if they chose.

Overall, the results are quite promising. Contestants on average correctly aligned the majority of RSSD-CIK and RSSD-LEI matches.

On July 1, the eight teams will share their results at an industry conference for database developers and users in San Francisco. The teams also will learn who best matched identifiers.

“We will have an open discussion about what worked and what didn’t work,” said Mark Flood, an OFR Research Principal and one of the challenge’s organizers. “The point is not to have a ribbon-pinning for a winner. It’s to have a conversation. We’re trying to foster a research community around this issue.”

Stacey Schreft is Deputy Director for Research and Analysis at the Office of Financial Research