What happened
The Federal Housing Administration (FHA) insures home loans made mostly to low-income, minority and first-time buyers. From 1998 to 2004 it worked with the contractor Unicon Research Corporation to build the TOTAL Scorecard. Unicon fitted a logit model to 1992 data on 9,867 FHA loans that ended without an insurance claim and 4,818 that ended with one. From 2004, lenders ran FHA applications through TOTAL, which accepted 65 to 70 percent automatically and referred the rest for manual underwriting. In a report dated 13 April 2006, the Government Accountability Office (GAO) found that TOTAL had produced more consistent underwriting results. The two scorecards that it replaced had sometimes disagreed, with one accepting a loan that the other referred. GAO also found that lenders using TOTAL needed less documentation for accepted loans and that some lenders had a faster approval process. GAO drew on agency documents and interviews and measured no time or cost saved. It also found that the development data were 12 years old by 2004 and that FHA had no formal plan to update the scorecard.
Sources
-
Audit report, by authors with no tie to the system.
U.S. Government Accountability Office. Mortgage Financing: HUD Could Realize Additional Benefits from Its Mortgage Scorecard. GAO-06-435, report to the Chairman, Subcommittee on Housing and Community Opportunity, House Committee on Financial Services, 13 April 2006. Section 'HUD Could Benefit Significantly More from TOTAL', pp. 15-16, and Table 1.
TOTAL has produced more consistent underwriting results and, for some lenders, has streamlined the approval process and reduced paperwork.
Details
AI system
- Workflow
-
Unicon fitted a logit model to a 1992 sample of FHA-endorsed loans, with credit score, loan-to-value ratio and cash reserves among its variables. Lenders' automated underwriting systems pulled credit data and sent each application to TOTAL, which returned accept or refer. FHA set a cut point that accepts 65 to 70 percent of applications automatically.
Follow-up
- Conditions for replication
-
Lenders must manually underwrite every loan that TOTAL refers. Accepted loans must still pass FHA's additional credit checks, which can downgrade them to refer. The scorecard must be updated as borrowers and products change, and GAO found that FHA had no formal plan to do so.
Entry details
- Sector
-
Housing finance: mortgage insurance underwriting.
- Purpose
-
FHA and its contractor Unicon Research Corporation built the TOTAL Scorecard to rank applications for FHA-insured loans by the likelihood of default and to accept most of them automatically.
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). TOTAL Scorecard cuts paperwork and evens out FHA underwriting (AIBD-1036). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1036.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)