What happened
Between onsite examinations, the Federal Deposit Insurance Corporation (FDIC) lists banks rated 1 or 2 that may be weakening on an Offsite Review List. Three models build the list each quarter from banks' financial reports, called Call Reports. SCOR and SCOR-Lag flag banks with a 50 percent or higher probability of a downgrade within a year. GMS flags banks in the 98th percentile or higher of downgrade risk over three years. In an evaluation dated 18 December 2019, the FDIC Office of Inspector General (OIG) examined downgrades from 31 December 2014 to 31 December 2017. In that period the models had flagged 67 of the 244 banks downgraded from a 1 or 2 rating to 3 or 4. With regional additions and a supplemental list, the program identified 94, or 39 percent. The OIG found that many missed downgrades stemmed from weak management or credit administration, which these reports do not capture. Only 47 of 1,665 reviews prompted by the list, or 3 percent, changed the supervisory strategy. The OIG did not compare supervision with and without the models.
Sources
-
Audit report, by authors with no tie to the system.
From December 31, 2014 through December 31, 2017, 244 1-or 2-rated IDIs received composite downgrades to a 3- or 4-rating. Of these, SCOR, SCOR-Lag, and GMS identified 67 of the downgraded institutions at some point during this 3-year period.
Details
AI system
- Workflow
-
SCOR and SCOR-Lag scored quarterly Call Report data and flagged 1- and 2-rated banks with a 50 percent or higher probability of a downgrade within a year. GMS flagged 1- and 2-rated banks in the 98th percentile or higher of downgrade risk over three years, based on rapid growth or noncore funding. Case Managers reviewed each flagged bank within 3½ months after the Call Report date and could recommend a visitation or an earlier examination.
Follow-up
- Conditions for replication
-
The weakness must show in a bank's Call Report data, and the models missed downgrades driven by weak management, credit administration and board oversight. Banks must file accurate Call Reports, and about 25 percent of the missed banks in the OIG's table filed amendments to correct errant data.
Entry details
- Sector
-
Bank supervision.
- Purpose
-
The FDIC built the SCOR, SCOR-Lag and GMS models to flag banks rated 1 or 2 whose quarterly financial data point to a likely downgrade before their next onsite examination.
31 more questions not yet checked
model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). SCOR and two other FDIC models flag 67 of 244 banks later downgraded (AIBD-1037). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1037.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)