Register today for our Generative AI Foundations course. Use code GenAI99 for a discount price of $99!
Skip to content

Industry Spotlight: Credit Scoring

In the U.S., credit scoring is dominated by three companies – Experian, TransUnion and Equifax, employing roughly 30,000 people. An important player in the scoring methodology is FICO, previously Fair Isaac Corporation, and the scores are typically called “FICO scores.” Credit scoring is the oldest application of predictive modeling, fulfilling a need that has been around for millenia. Ever since the development of money and, hence, money-lending, lenders have needed to assess the creditworthiness of borrowers.

Fair Isaac first started selling credit scoring services in the late 1950’s, but it was not until the 1990’s that credit scoring agencies began leveraging the wider availability of data and applying statistical models. Failure to pay, credit risk, is not the only risk facing the lending business. There is also interest rate risk – the risk that interest rates will move up or down, changing the value of loans with fixed interest rates. Interest rate futures markets are effective mechanisms for managing this risk. Prepayment risk is the risk that a loan will be prepaid. This deprives the lender of a specified stream of income, but at least the principal is returned.

Credit risk, on the other hand, entails the prospect of losing all or a portion of the principal. Predictive models that can predict credit risk have proved transformative for the financial industry. Accurate predictive models turned credit risk from something to be minimized, managed or avoided into something that could be priced. Loans can be subdivided into tranches by their riskiness, and that risk can be accurately predicted and priced for each tranche. Now, instead of hazard and uncertainty, you have new products to sell!

Unfortunately, the financial industry took this process too far and began playing elaborate, obscure and sometimes fraudulent statistical tricks with their products. Actual credit risk, the probability that a given borrower would repay a mortgage, and the real world predictors that can be used to estimate it, got lost in the haze. The result was the financial collapse of 2008.