Simon Squibb oversimplifies and calls them a scam
I paid off my house and my credit score dropped
In 2015 I had no student loan debit, no car loans and no revolving balance on my credit cards. I had never paid a bill late and I was earning in the top 20% of Americans and yet I did not have a perfect score.
When I paid off my house my credit score dropped 15-20 points depending on the company to around 780/850.
Credit scores are composite measures that are fed by imperfect data and are designed around determining how much companies can profit from you.
In my case, my score dropped (and this was clearly displayed on my credit reports) because I had “too few credit sources” and impacted both my credit utilization rate and the mix of credit I had available to me.
I didn’t plan to buy anything on credit at the time so I didn’t care that my score dropped.
It could have increased the financing cost of purchasing a new home or car OR have impacted my job prospects since employers very often check credit reports in hiring processes.
I disagree with the video commentary that credit scores are a scam.
It’s not that the scores are a scam, it’s that they are a proxy for a much more complex picture and they aren’t designed for the benefit of consumers (you’re an input to the product not the purchaser of these scores).
There are any number of services that aim to help you quickly improve your score by understanding how they are calculated and pointing out the controllable aspects of the score
– opening or closing credit cards
– shifting debit types
– timely payments
Credit scores are a great example of why feature engineering in modeling is so important. The summary metric score is likely less predictive or instructive than understanding the component depending on what you’re using it to gauge.
If I want to know if you’ll stay in you’re job maybe knowing that you own your house free and clear WOULD indicate that you have an easier option to walk that someone in a 30 year debt cycle.
I’m curious to know how many companies are using the disaggregated components of credit scores in their financial and risk assessment models?

Leave a comment