Statement of ALLSTATE INSURANCE COMPANY by wuxiangyu


									                 Statement of ALLSTATE INSURANCE COMPANY

                                    Steven R. Sheffey

                   National Conference of Insurance Legislators
           Property-Casualty Insurance Committee Hearing on Proposed
                            Credit Scoring Model Act

                                San Francisco, California

                                    November 21, 2002

Thank you for the opportunity to present our views. We believe that NCOIL has a
unique opportunity to address concerns that have been raised regarding the use of credit
history by insurance companies by adopting a model bill that will provide meaningful
regulation while still allowing consumers to benefit from insurance scoring based on
credit history. I’d like to take this time to explain how Allstate uses credit history and
why consumers benefit, and to also touch on some correlation issues.

Allstate derives insurance scores based on information contained in credit reports, a
practice that we believe benefits the insurance-buying public because it allows us to write
more insurance, because it increases the accuracy of our risk evaluation, and because it
allows us to reward people less likely to incur losses with lower premiums. The use of
insurance scoring is a significant advance in cost-based pricing because it allows us to
reward customers in already-existing risk pools with lower premiums. This means that
we can give lower premiums to many customers who otherwise would pay more for
insurance. As an insurance company, we need to collect a certain amount of premium to
cover expected losses. Our use of credit history allows us to more accurately identify
customers who are less likely to incur those losses, and reward them with lower

Our use of insurance scoring is premised on the same principle all of our underwriting is
based on: those less likely to incur losses should pay less for insurance. We have found
that insurance scoring is an extremely powerful predictor of loss.

Our customers--your constituents--want insurance at the lowest possible price. Our
customers—you constituents—want maximum insurance availability. What they don’t
want is higher taxes. We need to be careful that any model act does not prevent
insurance companies from continuing to reward customers with lower premiums. The
last thing we want is to inhibit the use of credit history to the extent that customers who
are paying less now because of credit history are forced to pay more. That would amount
to a hidden tax on people less likely to incur losses and a windfall for people more likely
to incur losses. Hidden taxes and undeserved windfalls are not good public policy.

Let me now explain our practices.
Although the information we use happens to come from credit reports, we are not looking
at "credit" or "credit worthiness." Unlike many users of credit reports, we do not look at
or consider income level. Someone turned down for a credit card or a bank loan could be
eligible for coverage in our preferred company based on insurance scoring, and could
qualify for a lower priced tier within either our preferred or nonstandard company.

Our scoring model does not consider race, ethnicity, income, wealth, geography, age,
gender, health status, marital status, nationality, net worth, occupation, religion, absolute
size of balances and limits, medically related items identified as such, insurance inquiries,
or items shown as being under dispute.

Our scoring models allow positive credit characteristics to offset negative credit
characteristics. This means that no one negative item will necessarily preclude anyone
from receiving our lowest rates.

We use credit history in addition to, not instead of, other risk assessment tools, to
increase the accuracy of our risk evaluation. Where Allstate Insurance Company has
begun to underwrite customers into tiers based on insurance score, we have been able to
broaden our underwriting guidelines, and offer insurance in our preferred company to
higher risk segments that we would have previously been offered coverage in our
nonstandard company. In some places we expect to soon be able to eliminate
underwriting based on credit history. In Nevada, for example, because we use insurance
scoring for tier placement, no one is rejected for auto or homeowners insurance in whole
or even in part based on credit history. About 20% of the new preferred auto business
we’ve written in Nevada since implementation of our scoring model in July 2001 would
not have met our previous underwriting guidelines, and we’ve seen this result replicated
in other states. If we had to stop using credit history in these states, the result would be
that some policies currently accepted into our preferred company would be rejected or
offered coverage in our higher priced non-standard company. That would not be good
for our customers or your constituents.

Allstate wants to write as much profitable business as it can. We are a publicly traded
company that is committed to growth. Through our use of insurance scoring, we can
write more business than we could before, and reward customers less likely to incur
losses with lower premiums. The net effect of our use of insurance scoring is that more
people are written and get the benefit of the rates in our preferred company because our
ability to consider credit history allows us to relax other underwriting standards. In
addition, within our preferred company, customers less likely to incur loss can qualify for
lower-priced tiers.

Allstate has found that auto customers in the worst 10% of scores will have over 60%
more losses than auto customers with the best 10% of scores. Homeowner customers in
the worst 10% group of scores will have well over twice as many losses as those
customers with the best 10% of scores.

I have attached an internal Allstate document that provides detailed evidence for the use
of credit history information in predicting insurance loss potential.

For Allstate, what is important is the fact that insurance scores predict loss. Allstate
decides whether an applicant qualifies for insurance based on a variety of factors.
Experience shows that these factors help predict the likelihood of loss. For example, we
underwrite against traffic violations only because our data shows that the group of people
with speeding tickets in the past tend to have more losses in the future, not because a one
traffic violation causes another traffic violation. The same type of analysis shows that
auto customers with the worst 10% of scores are about 60% more likely to incur losses
than those in the best 10% of scores. That analysis has been validated by the actual loss
experience of millions auto exposures. We would be irresponsible to ignore it.

Yet some people have asked why insurance scores predict losses. The link between
credit history and loss potential has been studied extensively by many scholars
independent of the insurance industry, in fields such as psychology, safety engineering,
occupational medicine, consumer research, and risk perception. Over 30 articles and
studies that we analyzed point to two basic explanations. These explanations are
persuasive, but the fact of insurance scoring’s predictive power—-not the reason for its
predictive power--is the basis for Allstate’s use of this important tool.

The first explanation relates to stress. People under stress are more likely to have auto
accidents. They may be more easily distracted or not react as well to certain situations
(the difference between an accident and a near-miss is often just a few seconds).
Financial problems are a known cause of stress. Therefore, some people with poor scores
are more likely to experience stress and thus more likely to incur losses.

The second explanation relates to risk-taking behavior. Different people have different
aversions to risk. Some people like to sky-dive. Some people are afraid of the
amusement park roller-coaster. Some people will run a yellow light if it was yellow
when they first saw it. Some people will stay under 55 on the highway. People who are
more likely to take risks are more likely to get into serious financial difficulties
(bankruptcies, liens, foreclosures, etc.) than those who are more risk averse. People who
are more likely to take risks are also more likely to get into auto accidents. Therefore,
some people with poor scores are more likely to engage in risky behavior and thus more
likely to incur losses.

Either or both of these theories may be true for a particular individual. In some instances,
financial difficulties might not be caused by risk-taking behavior but will still produce
stress. In other instances, however, it is the risk-taking behavior rather than stress that
leads to a greater likelihood of loss.

Allstate’s use of insurance scoring as risk evaluation factor is based on the fact of its
predictive power, not the explanation for its predictive power. Just as it would be foolish
to ignore the fact that apples fall down from trees because scientists have yet to fully
explain how gravity works, we would be foolish to ignore (and it would be unfair to our

customers to ignore) the fact of the predictive power of insurance scoring . Nevertheless,
we take great comfort that support for the link between credit history and loss potential
exists in the academic literature and is intuitively satisfying.

Finally, let’s not forget that the use of credit reports is already the subject of Federal law.
The Fair Credit Reporting Act, as amended in 1996, states that the use of information
from credit reports in connection with the underwriting of insurance is a permissible
purpose. This bill, passed by a Republican Congress and signed by a Democratic
President, enacted many important consumer protections. We recognize, though, that
some people think even more regulation is necessary, and we respect their views. We are
here to urge NCOIL to lead the way by adopting a model act that will lead to responsible
regulation, one that will provide meaningful consumer protection without forcing
customers to pay more or forcing insurance companies to reduce the availability of

I’d be happy to answer any questions you may have.


To top