Rethinking the Credit Scoring Model

I recently had the opportunity to meet with Clark Abrahams, chief financial architect at SAS, on a trip he took to New York. He has helped me out with some articles in the past and it was nice to put a face to a name.

Clark, who actually wrote a column for BS&T, is working on creating a new framework for the financial services industry around scoring the risk of customers/potential customers in a manner that's a bit more three-dimensional than the tried and true credit score. His progress in convincing the industry to accept this new concept, called the Comprehensive Credit Assessment Framework (CCAF) was the subject of our conversation.Clark is basically proposing the creation of a new credit score. Kind of a credit score "plus" that takes into account other dimensions of the consumer, such as their capacity to pay bills and their capital.

Clark says the banks need to look at their customers' information in the context of what's going on in the rest of their lives (for example, changing a job for higher pay) and as they relate to consumers in similar situations. In doing so, claims Clark, banks will get a more in-depth view of a consumer's risk than they do today. Furthermore, banks need to adopt models that are flexible and can adapt at the market changes and as the customer's financial condition changes as well. (This is just a basic description of his ideas. For more information, please see his column I referred to above and his own blog.)

This would be ideal. Imagine being able to drill down to such a degree into your customers' risk profile. If the industry had models in place like this, there is a good chance the subprime crises could have been averted or detected much earlier, according to Clark. He is not advocating that banks discontinue using the credit score. On the contrary, he says the credit score would be a part of an overall more broad-based credit rating process.

This all sounds great, but are the banks ready to take this step? Sure, they'll have a great deal more information at their fingertips, but there is such a thing as information overload. The complexity of performing a credit check could increase ten fold. Would banks be willing to bite the bullet and invest in what is needed to accomplish what Clark proposes? He kind of smiled at me when I asked him that. His answer was that it would most likely happen only if the regulators pushed it (and, by the way, he has presented his case to the regulatory bodies and the industry associations). This is, after all, a new way of thinking for banks, right?

Regardless of one's opinions on Clark Abraham's idea, it is clear from the state of the financial services industry that perhaps some new way of thinking is in order. Most of the technology is there today to help banks gather the information they need. Already, some are advocating the use of alternative credit data, such as records of utility bill payments, to score those with thin or non-existent credit files. Could this move into the thick file area? Who knows?

I'd like to know if any of our readers think it's time for the industry to reexamine the way it scores customers-not necessarily in the manner described above, but some new method. Is a new credit risk model necessary as the kinds of financial products offered become more complex and as banks look to different markets for business?