Following is PART III of the testimony presented by Lawrence S. Powell, PhD before the United States House of Representatives Financial Services Committee Oversight & Investigations Subcommittee on May 21, 2008, by:
Lawrence S. Powell, Ph.D.
Research Fellow - The Independent Institute (http://www.independent.org/), and
Whitbeck-Beyer Chair of Insurance & Financial Services
University of Arkansas-Little Rock
*********************************************
Section 3: Appropriateness of Insurance Scores
Regulators require insurance rates to meet three criteria. They must not be inadequate, excessive, or unfairly discriminatory*. A rating criteria is unfairly discriminatory if it is does not bear a reasonable relationship to the expected loss and expense experience among insured exposures. Given the evidence presented in Section 2, insurance scores clearly meet the third criterion. However, some people remain uncomfortable with application of credit information in insurance rating. In this section, I describe the individual and societal benefits of insurance scoring. Finally, I present evidence that competition in insurance markets prevents discrimination based on any factor other than expected losses.
(* Almost every state also imposes additional restrictions on the use of insurance scores in the ratemaking process.)
Insurance scoring benefits society in several ways. All of the benefits accrue from improved efficiency and accuracy of risk estimates. The first benefit is that insurance scores provide a very high level of accuracy for a relatively small cost. Using insurance scores reduces cost for insurance companies. Because the market for insurance is competitive, this savings is passed through to consumers as lower premiums. Data from a recent report by the Arkansas Insurance Department indicates that if insurance scoring were eliminated as a rating factor, nearly 91 percent of automobile and homeowners' insurance consumers would incur a rate increase. Using a slightly different method, the FTC (2007) study estimates that insurance scoring results in a decrease in insurance premiums for 59% of drivers.
The next benefit of insurance scoring is that improved accuracy may make insurers more willing to offer insurance to high-risk consumers for whom they would otherwise not be able to determine an appropriate premium (FTC, 2007). For example, insurance scoring information can allow an insurer to offer coverage to drivers living in a geographic area with high traffic density at a price the driver can afford. Without information from insurance scores, insurers would not be able to differentiate sufficiently among these drivers. Therefore, they would not be able to offer the coverage at a lower price for the lower-risk drivers living in the area. Consistent with this assertion, FTC (2007) finds limited evidence that the advent of credit scoring in automobile insurance coincided with substantial decreases in residual market mechanisms. This suggests insurers, with the benefit of credit information, are more willing to offer coverage to high-risk drivers (at a risk-based price) than they were before the introduction of insurance scores.
Another advantage of using insurance scores is it improves accuracy of information used to classify drivers. In addition to calculating more accurate loss predictions, the scores, themselves, are less likely to contain material factual errors than are several of the driving history variables used to underwrite insurance. Studies by Associated Credit Bureaus (ACB, 1992) and Trans Union report material errors in credit information in only 0.2% of credit records. In striking contrast, a study by the Insurance Research Council (IRC, 1991) found public information available on only 40% of a sample of known automobile losses. Underreporting of traffic citations also appears problematic. IRC (1991) indicates less than a third of all traffic citations are accurately reported in state driving records. Furthermore, consumers have a strong incentive to correct inaccurate credit information; whereas the opposite incentive exists for driving records. This is true because recorded driving events can only be adverse events. Data describing instances in which drivers avoid collision by defensive driving and alertness are not collected.
The final benefit of insurance scoring I would like to address is that, because scoring produces more accurate loss estimates, it results in outcomes that are more equitable for individuals and society as a whole. As noted in Section 2, insurance scoring is likely to make the price of insurance better match the risk of loss posed by the consumer. Thus, on average, higher-risk consumers will pay higher premiums and lower risk consumers will pay lower premiums (FTC, 2007). This addresses a very common problem in the insurance mechanism called cross-subsidization.
When insurers cannot accurately classify applicants for insurance, they must either decline applications, or charge the same premium to high-risk and low-risk drivers. The latter case obviously leads to cross-subsidization - when low-risk drivers must overpay to make up for underpaying high-risk drivers. However, the former case, declining applications for insurance, ultimately leads to the same outcome. This is type of cross-subsidization is facilitated by residual markets for insurance.
Each state has a residual market mechanism to make insurance available to drivers whom the voluntary market will not cover. Residual market mechanisms effectively set a maximum price that insurers may charge for insurance. If insurers are not willing to offer coverage at this price, consumers may purchase coverage at this price from the residual market. However, if the premium is not enough to cover losses and expenses, insurers in the voluntary market must make up the deficit in proportion to their market shares.
FTC (2007) shows that as insurance scoring has become more common in ratemaking models, the populations of states' residual markets have decreased. This suggests insurance scoring results in more equitable or fair outcomes compared to less accurate rating models that do not use insurance scores.
Perhaps the most controversial result appearing in FTC (2007) is the study's assertion that insurance scores exhibit a proxy effect for race. Objective consideration of this result leads me to doubt its validity. The econometric test used to support the existence of a proxy effect is flawed such that it would not withstand the scrutiny of a legitimate academic peer-review process. Clearly, the lack of objective confidence in the result suggests that public policy should not be altered to address this weak finding.
Another way to address the appropriateness of insurance scoring is to consider the level of competition occurring in insurance markets. If insurance markets are competitive, insurers will not be able to charge excessive or unfair prices. If an insurer tries to set prices based on anything other than expected losses and costs, it will either, suffer substantial losses if the price is too low, or, if the price is too high, it will lose market share as its competitors offer a lower price to the same consumers.
Effective competition is a fundamental characteristic observed in U.S. insurance markets. Competition prevents insurers from charging excessive or unfair prices. In 2005, NAIC data show an average of 157 insurance companies underwriting the private passenger automobile cover in each state. It is, therefore, reasonable to believe that an insurer cannot systematically over-charge a group of drivers because one of the other 156 existing companies, or perhaps a new company, has an opportunity to cover that group of drivers at an equilibrium price. Compare such competition to other "required" services such as phone, gas, electric, etc. where consumers have at best the choice between two companies.
We are not in this hearing because everyone likes insurance scoring. I have heard critics of insurance scoring describe potential or anecdotal unfair outcomes associated with its use. I do not dispute the fact that some consumers have encountered individual rating scenarios that seem to lack intuition. For example, I know of a consumer in Arkansas who received an increase in his premium because his wife cancelled a credit card they were not using. However, he called a few competing insurance companies and found one that offered him the same coverage at a significant discount from what he was paying before the change in his credit. This is an example of competitive markets reaching an optimal outcome.
While competitive markets are very effective at making the goods and services consumers want available to them, critics have voiced concerns that when a drop in credit is unrelated to insurance risk some individuals could be mistreated by insurance scoring.
In response to such concerns, almost every state has regulations in place to recognize the benefits of insurance scoring, while limiting its use in certain scenarios. I think it is worth noting that many insurers offered the same protections as these regulations require before the laws were enacted. This is another example of competitive markets creating an optimal outcome.
Conclusion
Setting reasonably accurate prices for insurance is a difficult task because insurers must establish prices without the benefit of knowing all of the costs involved. To offset this hardship, actuaries have developed complex pricing models using applied economic and statistical tools. While this complexity is necessary, it unfortunately leads to a lack of understanding among people who have not developed such specific expertise.
Insurance scoring is an example of a beneficial tool used in ratemaking that is often misunderstood. Insurance scores are relatively powerful and accurate predictors of losses, even when controlling for other factors known to be correlated with losses. When insurers use insurance scores to improve the accuracy of predicted losses, it benefits individuals and society. It increases the equity or fairness in insurance pricing outcomes because, on average, premiums are closely related to consumers' risk of loss. Insurance scoring also adds value to insurance transactions. It reduces the overall cost of providing insurance because insurance scores are accurate and inexpensive rating variables.
Finally, the vigorous competition exhibited by the property and casualty insurance industry suggests that pricing of insurance based on anything other than expected losses is nearly impossible. Insurance markets show strong signs of effective competition including a large number of suppliers and low barriers to entry.
(The views expressed in this article/commentary are solely those of the author and do not necessarily represent the views of MyNewMarkets.com, the Insurance Journal or Wells Publishing.)
© 2010 by Wells Publishing, Inc. All Rights Reserved. Privacy Policy | Terms and Conditions | Advertise with Us | Contact
Credit scoring is not necessary and it does not prevent the high score folks from being dishonest or negligent. What you have are honest folks with low credit scores, no claims paying more for folks with high scores and having more claims.
I suspect the statistics that support the credit rating theory are flawed. The stats should include every policyholder, with low scores, with every company in every state and comparing each company to the other, types of losses and dollar amount, and location of risk. The stats should also include a comparisons of the high credit scores folks with the same criteria. It is not accurate to survey a segment of policy holders in a certain geographical area.
What should be done is spread the risk, which should be the function of insurance. In fact, regulation should require that if a P & C company wants to sell insurance they must sell it in every town, city and state and to offer all risk coverage without exclusions. This is called spreading the risk. This would lower rates for all nationwide and elimiate hardships for folks who presently can't afford or find to purchase the insurance they need.
There are too many Insurance Companies and most of them aren't big enough to handle the catostrophic losses they insure (i.e. Century 21 in Louisiana). They don't plan too insure them, they will declare bankruptcy. Why? Because they can. These insurance companies hurt the big companies like, State Farm, Allstate and others. How? They're taking away their premium dollars by competing with unfair pricing and offering coverage they don't plan on paying for.
Credit scoring should not be a factor in trying to asses future fradulent or negligent claims. I'm not saying that fraud doesn't occur amoung the poor credit scores. I'm saying that fraud and negligence occurs with the high credit scores too and this makes it an unfair and unjust system. Whether it is more or less for one or the other is irrelevant.
The Industry needs to find another way to fight fraud and negligence. Insurance Credit rating is a cheap out. They don't need it. Not to mention that it in some ways is red lining.
Dr. Phillips suggests that credit scoring is more accurate than driving records, in one, because drivers have no incentive to correct their driving record to put on an adverse event. Does he actually believe credit users have any incentive to add an adverse credit event to their credit reports either?
Dr. Phillip suggest that it works because it causes high risk customers to pay more. Someone who drives safely but doesn't earn much isn't necessarily a high risk customer, nor is someone who earns a lot and pays bills on time necessarily a low risk customer.
He states that residual markets decreased as credit scoring use increased. This is another correlation without causation. The entire industry has been in a soft market cycle for quite some time, the same time that credit scoring increased in use. In soft market cycles insurers, in an effort to write more business, loosen their underwriting standards and thus business previously written in the residual market is written in the standard market. There is nothing shown in his material that the decrease in residual markets wasn't due to the soft market cycle rather than insurance scoring.
The example he cites of the person whose insurance was cancelled because they cancelled an unused credit card is a perfect example of why credit scoring is worthless and harmful. The fact that the person was able through shopping their insurance to find a better deal doesn't make the use of credit scoring by the company that cancelled make sense or not harmful to the public.
Credit scoring is used because it is a cheap way insurers think they can redline that most people won't be able to figure out. In years past insurers sent inspectors out to view the properties they were asked to insure. By comparison insurance scoring is much cheaper and it allows them not to insure many exposures in geographic areas where they'd rather not offer insurance and to people they couldn't refuse to insure for other reasons.
re: the gathering of ssn for uses other that
social security purposes is ILLEGAL
and the agents have been forced to possible
future prosecution by the carriers
Is it wrong to enjoy some fruit of my labors? - I don't think so.
There are a few cases where one doesn't get a preferred score because he / she never borrowed. This needs to be addressed and most [ regional carriers ] are prone to do so. Other than that, don't mess up a good thing.
Say a person who has worked hard all of his or her life through no fault of their own loses a good job and experiences financial difficulties. Their credit score drops and suddenly several factors fall into place that perpetuate their financial downfall. Insurance rates increase, rates for future loans increase, rates for existing credit card debt increase, fees and penalties compound and most damaging of all; future employment is jeopardized due to use of credit scores by potential employers. Just because a parallel can be found actuarally does not deny the right of an individual to protection from discrimination by association. The implication that financial difficulty equals dishonesty is flat out unfair and needs to be legislated against.