The report here is a Pricing Disparity report. This is one of the ComplianceQC Fair Lending module's reports, and is designed to explain whether a protected class (such as a racial minority or a particular gender) receives inexplicably different treatment from other borrowers. To see the entire report in PDF form, click here.
This page will break the report into three segments. The first segment is the header, which looks like this:
The header identifies the range of loans that have been selected, in this case loans originated between 10/01/01 and 12/31/01. On the left hand side, it then goes on to explain that it will compare loan officers at a particular branch with each other to determine whether a particular loan officer is discriminating in determining the rate of interest for a loan. On the right hand side of the header it indicates that the comparison involved deals with race, and that all minorities are being checked against white borrowers. The Fair Lending Module will allow you to tailor these comparisons to your needs so that you can assess disparities by gender or age, and compare branches, loan products, or any other component of the origination process.
The second part of the report appears directly below the header and on the left hand side of the page. The names in the example below have been scrambled for privacy reasons, as the report was generated with actual data:
 |
The branch code number is in the far left column, and the loan officer's name is immediately to the right of it, followed by the branch name. The words "Traditional Disparities" near the top of this segment indicate that the disparities discovered are ones where the protected class does not fare as well as the unprotected class. The system will also check for reverse disparities, which are situations where the protected class seems to receive disproportionately better treatment than the unprotected one does, and also mixed disparities, which occur when one indicator points to a traditional disparity and another indicator points to a reverse disparity. Those types of disparities are sought out because they typically only occur in samples involving bad data; they thus serve as a self-correcting mechanism on the report's accuracy. |
The third segment of the report appears directly to the right of the second. It looks like this:
| The "Loan Counts" column explains how many loan records were examined for purposes of this analysis. The "Magnitude (Average)" column describes the magnitude of the independent variable for the report, which, as is indicated in the header of this report, is the Annual Percentage Rate. The columns below illustrate that the APR on loans for the protected class is no less than a percentage point higher than for the nonprotected class for the loan officers reviewed in this table. The final column, labeled, "T-Test Probability," represents the probability that the disparity we observed is the product of chance. Notice that for the first loan officer, the probability that the disparity is due to chance is .5%, and for the others, it is statistically impossible. A report like this is a clear warrant for corrective action. |  |
|
|