Institutions with preapproval programs should pay close attention to how they incorporate preapproval applications in fair lending reviews. Fortunately, LendingPatterns™ and Fair Lending Magic™ provide robust tools that allow you to produce a thorough analysis.
Importantly, LendingPatterns™ allows users to run the popular Introductory Fair Lending Analysis and Lending Disparity Profile reports with a filter on the HMDA Preapproval Status field. This allows users to run reports focusing separately on 1) apps that came in through the preapproval program, and 2) apps that did not come in through the preapproval program. When analyzing apps that came through the preapproval program, the Denial Disparity Index (or DDI™) will capture records with action taken code 7 (preapproval request denied), and the Fallout Disparity Index (or FDI™) will capture records with action taken code 8 (preapproval request approved not accepted).
I would note a couple points about the peer analysis:
- Within your preapproval program, comparing your FDI™ from 2017 and prior years to peer institutions may be tricky, if not impossible, because reporting of preapproval requests that were approved but not accepted was optional.
- Identifying peers within a region to compare your preapproval program to is inherently problematic because geography is not reported on apps with action taken codes 7 and 8. You may be limited to selecting all lenders who have any apps (i.e., including apps that did not come through a preapproval program) within a particular state or set of states, even though many of these lenders will not have preapproval programs.
In Fair Lending Magic™, you can determine whether your denial and fallout rate disparities are statistically significant. Then you can run your approve/deny analyses by loan product three ways:
- First do it the traditional way which is action taken code 1 and 2 (approve) versus action taken code 3 (denied), irrespective of preapproval status.
- Exclude apps that came through the preapproval program.
- Conduct an approve/deny analysis on apps that came in through the preapproval program, including statistics and match pairs. This analysis would include action taken code 8 in the approved group and action taken code 7 in the denied group.