
COVID-19: A Positive Disruptor to Expected Credit Loss Models
The current murky picture delivered by traditional data has forced an increased focus on new methods that better capture credit risk.
The current murky picture delivered by traditional data has forced an increased focus on new methods that better capture credit risk.
When conducting recent examinations for Bank Secrecy Act/anti-money laundering compliance, regulatory officials flagged deficiencies in risk assessments, a need for more maturity in compliance systems and processes and data integrity issues, particularly as a result of mergers, as areas of concern during a panel discussion at the American Bankers Association/American Bar Association Financial Crimes Enforcement Conference in Washington, D.C., today.
The Government Accountability Office this week found that the two of the Federal Reserve’s guidance documents on large bank supervision—which were issued in 2012 and 2014 by the Fed’s Large Institution Supervision Coordinating Committee—are considered rules for the purposes of the Congressional Review Act.
Regulatory guidance wants models that can be challenged. But when artificial intelligence turns a model into a black box, how can bankers manage model risk?