Insurance scoring 'Color Blind,' Says Arkansas Insurance Commissioner

OLDWICK, N.J. 04/14/2003 (BestWire)-Credit-based insurance scoring is a fair predictor of risk, said the president of the National Association of Insurance Commissioners, adding he doesn't expect the association to conduct a study on whether it has disparate effects on certain classes of people.

Mike Pickens, Arkansas insurance commissioner and NAIC president, said insurance scoring is valid and credible, pointing to a recently released University of Texas study showing a high correlation between credit scores and frequency, probability and degree of loss.

The study "was the first one not bought and paid for by an insurance company," Pickens said. "It basically legitimizes everything we heard. Why it works, I don't know, but it does work."

If people take care of their most important asset --their finances--they are likely to exercise the same amount of responsibility in other areas of their lives, Pickens said. They're also more likely, if they have a minimal loss, to pay for the loss themselves "because they have the financial wherewithal, rather than file a claim."

The NAIC has a task force, led by the Washington state and Oregon insurance departments, looking at insurance scoring. They've issued an educational brochure for consumers, and the efforts at the NAIC have been aimed at disclosure, Pickens said.

But funded consumer representatives at the NAIC want state commissioners to go further and determine whether the use of insurance scoring adversely impacts "minorities and other protected classes," he said. "It's a tough issue to get your arms or mind around. It could open up a lot of other legitimate underwriting criteria to scrutiny," he said.

Pickens said he was speaking as the Arkansas insurance commissioner. "Credit scoring seems to be a fairly color blind way of making a determination of risk," he noted. " When you look at credit you don't know what their ethnicity or age is."

A valid and credible study on whether the use of credit has a disparate affect on certain classes of people would involve polling consumers, he said. "I don't know if you'll see the NAIC pursue a study in this area because it's time consuming and costly and probably wouldn't be very constructive at the end of the day."

Insurance commissioners have taken a balanced and thoughtful approach on this issue, which is what's taken place legislatively in a majority of states, Pickens said.

The Texas University study helps demonstrate causality, but more needs to be done to explain why there is a correlation, he said. "Why is the primary question for regulators and legislators. They want more information on why. The Texas study didn't try to answer that question."

Speaking during an April 11 Deloitte & Touche presentation entitled "Credit Scoring: The Regulations, the Models and the Alternatives," Pickens said a "great deal" of legislative activity is going on in the states this year concerning the issue of insurers' use of credit information in underwriting.

Some 41 bills on the subject have been introduced in legislatures around the United States, he added.

In his home state for example, two bills were introduced in this legislative session. One was based on a model bill approved by the National Conference of Insurance Legislators. Pickens said his department worked with the Independent Insurance Agents & Brokers, and the American Insurance Association.

But another legislator wanted to place a total ban on the use of credit information for personal passenger automobile insurance rates, along with other credit information-related restrictions, Pickens said.

What ended up being passed by the Arkansas legislature was a bill that sets "pretty reasonable standards on the way credit scoring is used," and one with which independent agents and insurance companies are comfortable, Pickens said.

As in Arkansas, the scope of the bills introduced around the country varied. Pickens said many states started out with legislation that was pretty restrictive, "but most states have done something reasonable and come up with something close to the NCOIL model," he said.

Part of the reason for the consumer outcry and flurry of legislative activity, not only this year, but for a few years, is that insurance companies didn't do a good job of educating consumers or agents prior to using credit information in underwriting, Pickens said.

"The industry didn't educate consumers and agents as they should have," Pickens said.

Regulators would be open to other predictive variables besides insurance scoring if insurers explain what it is and why it works, he said. But if the industry is not open to more regulation and scrutiny.

(By Dennis Kelly, Washington bureau manager, BestWeek: Dennis.Kelly@ambest.com)