From my pivot tables, it seems we’re going in the wrong direction. The real people being discriminated against here are people who don’t provide their race on their applications. What a startling revelation!
But, really, the discrepancies in the rejection rates seem to come from the different loan purposes that’re included in this huge data set. When I looked at only Home Improvement loans, rejection rates jumped up, but in differing amounts between races. White, Hawaiian/P.I., and Asian denial rates are much higher for home improvement loans, while American Indian rates only go up a few points, and rates for Black/African American applicants actually lowered.
For refinancing loans only, all races’ denial rates went up, excluding Native Hawaiian/P.I., which went down 5%. American Indian – Alaska Native applicants went up less, though, from 31% to 35%, as opposed to Whites’ 18 to 26. These differing rates seem to be a possible culprit for the differences we’re seeing in our denial rates between sources.
This is great work. You really dug out all of the trends I was hoping for.