In Cathy O’Neil’s chapter 8 of Weapons of Math Destruction, the discussion turns to credit scores. In many ways I feel the same way about this chapter as I did about chapter 7: O’Neil correctly identifies the large, looming moral problems and implies a reasonable solution (regression, essentially (this time to FICO, a “good” WMD rather than merely human judgement)), but fails to expand on points that should not be glossed over and merit their own discussion. One example of this is the question of why these e-scores are being used when FICO is apparently so good. O’Neil just states that FICO is a neutral feedback loop and implicitly implies that this is what should be the goal. I think the occurrence of a feedback loop bears no relation to what a company will choose as it’s credit rating model. Rather, their concern with risk and thereby, in the loan industry, making money, is instead going to cause them to choose whatever score model makes them the most. While the FICO was a good improvement over the issues with human selection biases that have no relation to making payments, unfortunately, this was probably not the issue that merited it’s creation. I really don’t know what would make things more fair. Fairness is a very tricky question specifically in relation to credit and loans. Each party likes to think they are getting the upper hand, and the decision to even give credit/loans and how much is an even more gray area. I myself have just begun to learn about this stuff recently as I will soon have a 9–5 job and will need loans at some point in the future. The whole system seems complicated and dainty to an outside viewer, something that maybe should not be wisely disrupted as it could send shockwaves should the America/world economy. Maybe FICO is the last, proven bastion of credit rating.
