Assess whether comparative files show second-review bias (c4bf31)

Situation

Model-risk partner for credit scoring in a credit-card issuer changing line-assignment logic has one working extract — appraisal-gap outcomes in majority-minority tracts — after a vendor score change with no disparate-impact test. Model-risk partner for credit scoring in a credit-card issuer changing line-assignment logic has appraisal-gap outcomes in majority-minority tracts after a vendor score change with no disparate-impact test. If that extract cannot support comparative files show second-review, the honest Fair Lending CRA and Special-Purpose Programs output is hold.

Decision

Model-risk partner for credit scoring in a credit-card issuer changing line-assignment logic must choose Remove access or reverse the item / Temporary compensating control / Approve a documented exception / Hold using appraisal-gap outcomes in majority-minority tracts after a vendor score change with no disparate-impact test.

Hypotheses to test

• The population in appraisal-gap outcomes in majority-minority tracts is the one a vendor score change with no disparate-impact test named, so Remove access or reverse the item follows for this CRA and Special-Purpose Programs file.
• The population in appraisal-gap outcomes in majority-minority tracts is adjacent only to a vendor score change with no disparate-impact test; Temporary compensating control is the honest Fair Lending call.
• A credit-card issuer changing line-assignment logic already contained a vendor score change with no disparate-impact test before appraisal-gap outcomes in majority-minority tracts arrived; no new CRA and Special-Purpose Programs path.
• Provenance on appraisal-gap outcomes in majority-minority tracts after a vendor score change with no disparate-impact test is broken; do not pick Remove access or reverse the item or Temporary compensating control yet.

Analysis required

• Test a documented exception versus a pattern a credit-card issuer changing line-assignment logic must defend.
• Match the adverse-action language to the facts in appraisal-gap outcomes in majority-minority tracts.
• Check HMDA coding and underwriting policy against comparative files show second-review.
• For this Fair Lending CRA and Special-Purpose Programs file, read appraisal-gap outcomes in majority-minority tracts against a vendor score change with no disparate-impact test and write the one fact that would move comparative files show second-review for model-risk partner for credit scoring.