Project

General

Profile

Initial Client Questions

Notes and Implications of Questions

From David Lynn

Okay, some thoughts on questions. All of this would pertain to SD County, unless you wanted to look otherwise. These are in no particular order:

1) Avg credit score by zip (or even better if by some smaller designation that is still mappable) - like this US version, do different regions tend to have different credit scores (which we would expect, so just a visual way to highlight, really)
2) Avg credit score by demographic - visual map of credit scores by different demographics, like #1 (could also have a geographic overlay if valuable)
3) Avg credit score by demographic relative to household income -* is a white family with a $60k household income in Carmel Valley likely to have a higher credit score than a minority family with $60k household in City Heights - meaning there is more at play than income level*
4) Business lending amounts by zip (or other geographic coding) - do certain areas get more business capital than others
5) Avg business lending rates and amounts by zip (or other geographic coding) - do certain areas get better rates than others for similar loans
6) Avg business lending rates and amounts by owner demographics -* do certain race/ethnicity/gender/etc get better rates than others for similar loans*
7) Business lending approval rates and amounts by zip (or other geographic coding) - are similar loans in different areas more or less likely to be approved
8) Business lending approval rates and amounts by demographic - *are similar loans from different demographic borrowers more or less likely to be approved
*
And extending on #8 to me is where it really gets interesting, to see if data can be isolated that shows systematic barriers:

9) Is somebody with the same credit score, education, or other variables than can be isolated but a different race/ethnicity/gender/geography less likely to get a loan approved, or likely to get a smaller loan, than somebody nearly identical but with the demographic element varying?

Not sure what story it will tell. We can assume, obviously, but data and presentation has value. Everybody seems to focus on credit score, so the nut of the questions to me comes down to:

10) Is the reliance on the credit score actually what happens? I.e., with same credit scores but different demographics does that change approval rates or loan amounts?
11) Is there another demographic factor that is more likely to correlate with loan approval rates and amounts than credit score?

If possible to get loan repayment info, then:

12)* Do the factors used (credit score, etc) to approve loans actually correlate to successful repayment?* I.e., if black business owners are less likely to get a loan, are they actually less likely to repay, or are their repayment rates just as good as the white business owners (or better)?

Could do the same for consumer or auto lending, presumably.

Overall, this is just a way to highlight inequities. From the Mission Driven Finance and impact investing perspective, the argument is that we can close the gap that traditional lenders won't. And possibly that our method of assessment will allow a black borrower to get a loan that the data shows is actually more likely than a white borrower to repay, and thus is not actually a risky loan though from traditional standards it would look like it.