Assess whether comparative files show second-review bias after an underwriter

Situation

An underwriter chat that used coded language put mortgage pricing residual by prohibited-basis group in front of community-development lender in a mortgage company after a pricing-regression spike. This Fair Lending / Redlining and HMDA Data close is comparative files show second-review from mortgage pricing residual by prohibited-basis group, and the live options are Remove access or reverse the item, Temporary compensating control, Approve a documented exception.

Decision

Community-development lender in a mortgage company after a pricing-regression spike must choose Remove access or reverse the item / Temporary compensating control / Approve a documented exception / Hold using mortgage pricing residual by prohibited-basis group after an underwriter chat that used coded language.

Hypotheses to test

• An underwriter chat that used coded language is noise around an already-controlled Redlining and HMDA Data process in a mortgage company after a pricing-regression spike, given mortgage pricing residual by prohibited-basis group.
• An underwriter chat that used coded language is the event in mortgage pricing residual by prohibited-basis group that forces Remove access or reverse the item for community-development lender under Fair Lending.
• Mortgage pricing residual by prohibited-basis group shows a one-file miss after an underwriter chat that used coded language, not a Redlining and HMDA Data program failure.
• Mortgage pricing residual by prohibited-basis group cannot decide comparative files show second-review yet after an underwriter chat that used coded language; hold is the only Fair Lending close a mortgage company after a pricing-regression spike can defend.

Analysis required

• Compare mortgage pricing residual by prohibited-basis group to similarly situated files, second-review notes, and reason codes after an underwriter chat that used coded language.
• Flag any disparate-impact table community-development lender cannot explain from mortgage pricing residual by prohibited-basis group.
• Test a documented exception versus a pattern a mortgage company after a pricing-regression spike must defend.
• For this Fair Lending Redlining and HMDA Data file, read mortgage pricing residual by prohibited-basis group against an underwriter chat that used coded language and write the one fact that would move comparative files show second-review for community-development lender.