Predicting loss in personal lines insurance is a risky business, but carriers have found a tool to offer lower rates to customers who rarely file claims while making those who do file claims pay their fair share. Insurance scoringbased on credit recordsmay confound some consumers (and departments of insurance), but it makes perfect sense to carriers.
By Robert Regis Hyle
Personal lines insurance isnt that difficult to understand. Were in the business of assuming risk, and weve promised our customers we are going to attach the premium to the risk, says Win Logan, director of corporate communication for State Auto Insurance. It isnt always that easy, though. As Logan points out, The relatively few pay for the claims of the many. Rewarding the few, in this case, is becoming easier for insurance carriers in recent years as underwriting tools and rating engines have added another data element to their arsenal: a policyholders or applicants credit history.
You cant predict individuals, but if you have 1,000 people with the same credit report, you might be able to predict losses, says Charles Neeson, senior executive, personal lines products, the Westfield Group. Out of the several hundred items on a credit report, insurance carriers discovered there are 10 or 15 itemsnot necessarily the same ones that were used to predict loan defaultthat were telling in their prediction of insurance loss. [The data] may not be as amazing as noticing a youthful driver has four times the amount of loss that an adult does, but the information probably is better than rating territory, says Neeson. In other words, it is more important to know your credit score than to know you live in Cleveland vs. a rural area.
Its not rocket science to prove the correlation between scoring tiers and loss experience, asserts Michael Puerner, vice president and general counsel for Hastings Mutual Insurance. Its a fairly well-developed trend within the industry, he says. It may be difficult for people on an individual basis to understand how specific credit behavior reflects an increased likelihood of loss, but thats true for a lot of actuarial-based rating tools. They measure large numbers and correlate large data.
Whats in a Name?
Insurance scoring is what insurance companies use to predict future insurance loss, Neeson notes. Credit scoring is what banks and other lending institutions use to predict loan default.
Puerner cautions Hastings doesnt use scoring to underwrite risks. We use it only to determine the rate, he says. [The insurance score] goes through our underwriting system, which has a rating function, as well.
Insurance carriers always are looking for the best way to assign that risk, Logan believes. If it works, if its accurate, and if the modeling says this is the best way to decide who is most likely to have an accident and who is least likely to have one, then we go that route, and thats what we do with insurance scoring, he explains. As long as other companies are using credit information, we really dont have much choice because it works. If we dont use it and our competitors do, we end up with adverse selection.
Robert Hartwig, chief economist for the Insurance Information Institute, maintains the use of credit information is one of the most powerful predictors of loss the insurance industry has come across in decades. The statistical correlation is there, and thats what matters, he says. Statistically what it means is people who are very cautious financially are unlikely also to be people who get in their car and drive 90 miles per hour down a residential street, he adds.
Hold On There
The opposition to insurance scoring among consumers relates to privacy issues, Matt Josefowicz, group manager of the insurance practice for Celent, contends. If you look at it from a consumer standpoint, it is very easy to understand if you are going to borrow money from others, they have the right to find out what your history is in terms of paying back money you borrowed, he says. With an insurance policy, a) youre giving the carrier money, its not giving you money until theres a claim; and b) your credit history is not directly linked to your risk profile the way your driving record is, although there is statistical evidence related to the quality of the risk. Consumers are concerned about the issue, which makes state legislators concerned and eventually state insurance commissioners. As Josefowicz sees it, scoring is going to create a more difficult insurance environment for people with bad credit, which is a very large portion of the population these days. I think thats part of the fear, he says. It will become very difficult or expensive for people with bad credit to get coverage.
Some in the consumer public have a mistrust insurance companies only are attempting to increase everybodys rates, Neeson believes. Actually, companies just are trying as accurately as possible to assign a premium to people related to their expected future loss, he says.
Logan understands the public relations challenge that comes with insurance scoring. For years, all insurance companies talked about was driving safety and driving record, he says. That still plays an important role in assigning the risk, but then theres something such as insurance scoring, and theres a disconnect with the consumer. Consumers wonder what their credit record has to do with what kind of driver they are, says Logan. In fact, what it has to do with is what kind of risk we think you are.
Things are turning in the right direction for insurers with this issue, Hartwig indicates. Many people were surprised when they heard insurers were beginning to use [scoring], he says. They didnt understand the relationship. [Insurers] are not using all the elements of a credit score. Theyre using only [the data] that correlates with future loss. Thats why we call it an insurance score.
The use of credit-based information to judge performance under a variety of different contracts or responsibilities is becoming more commonplace, Hartwig continues, adding a credit check is commonly done as part of an employment check these days. Your credit record is part of your financial identity today, and that spills over into the area of insurance, too, he says.
I certainly would tell my children, If you are going to get a good insurance premium when you are on your own, youve got to be a good driver and you have to behave responsibly, says Logan. That doesnt go away. But [insurance] scoring also is a part of how the risk is evaluated. Thats where our industry is challenged. Weve got to do a better job of explaining to consumers why this is important and why we use it.
Who Goes There?
Westfield started researching the use of credit information in 1999 and looked at the technology and the loss-predicting value of insurance scores before deciding it was something the carrier wanted to do. After a review of different models, Westfield decided the model from Fair Isaac best fit its needs.
There are three different groups of carriers with regard to insurance scores, Neeson suggests. Some companies arent using credit at all, he says. Others are companies that have created their own customized insurance scoring model. Generally, smaller companies wouldnt do that, says Neeson. Those with their own solutions purchase credit reports and run the data through their particular algorithm, he adds. The third level of carriers consists of those purchasing industry-standard models, such as the one Westfield purchased from Fair Isaac. ChoicePoint and iiX are two other popular insurance solutions, and Neeson points out credit bureaus have developed their own insurance program, as well.
Carriers are taking the credit information they receive, putting it into their actuarial models, and building it into their rate engines just to give them one more data input into underwriting decisions, says Josefowicz. Underwriting profitability is all about effective rating, so, if there is any piece of data that is going to make their underwriting models tighter, they are going to seek to use it.
Profit or Loss
Another point of confusion Logan would like to correct is the view insurance scoring has increased premium for carriers. The idea of scoring was not to make more money, he says. [State Auto] always has been a very strong underwriting company. Prior to [insurance] scoring, we were one of those rare companies that made an underwriting profit.
He states late payments are not an automatic credit-score killer. Overall good credit can outweigh one or two instances of late credit-card payments, he says. By the same token, having no late payments on your credit does not mean a perfect score will be generated. Payment history is just one piece of information as we calculate a score.
The insurance scores work within the Westfield system similarly to motor vehicle reports. Our computers here at Westfield pull MVRs, CLUE reports, and credit reports; assemble them all together; run a scoring model that calculates an insurance score; read the motor vehicle report; and come out with a premium, says Neeson.
Westfield closely monitors the insurance scores, he adds. We know that people with better insurance scores perform better than people with bad insurance scores, he says. For the industry, Neeson estimates between two-thirds and three-quarters of the population are seeing improved prices because of insurance scoring. It should be pretty much that way because most people dont have claims, so its not surprising to see if you have a good pricing mechanismwhether it is age of driver, territory, insurance score, or prior lossesyou should see an efficiency occur where people with less likelihood of claims pay less, he says.
Banks have been using credit information for decades, Hartwig comments. Oftentimes, the use of credit information was an arduous, labor-intensive process that in the end proved to be fairly subjective. But today, because of advancements in computer storage and processing technology, more data elements can be stored cheaply and be processed and analyzed quickly.
The fact insurers are able to use credit-based information in their underwriting process is the result of two things, Hartwig remarks. One is technology, he says. Credit-based information in the database is sufficiently rich and can be attained relatively inexpensively and quickly. Also, and this is obviously the most important fact, this information actually has predictive value when it comes to estimating future insurance losses for individuals who are applying for insurance.
From a technology standpoint, Josefo- wicz believes the credit bureaus are good at sharing their information electronically in real time by using Web services. Most of the carriers are building into their underwriting systems the ability to accept the information that way, he reports. Its very automated because the credit bureaus are used to providing this information in real time for the mortgage industry and the lending industry.
Regulators have been telling insurers to use less information related to where someone lives, and insurance scoring achieves exactly that, points out Hartwig. [Scoring] has nothing to do with where you live, your income, your race, or your gender, he says. Credit scores are absolutely blind to all of that.
Making Insurance Law
While several states have debated the issue of insurance scoring, the practice actually ended up in a courtroom in Michigan earlier this year as the Insurance Institute of Michigan and several insurers successfully challenged an administrative rule issued by Linda Watters, the Michigan Office of Financial and Insurance Services (OFIS) commissioner, to ban insurance credit discounts.
Two earlier bulletins issued in 2003 by Watters and her predecessor seeking to put restrictions on the use of credit information led to an ill-fated attempt by insurers to have legislators codify the use of scoring for rating personal lines insurance. When Watters followed with a third bulletin banning the use of credit information, Michigans insurers went to court.
We opposed the rule during the rule promulgation process, and the rule became final, says Michael Puerner, vice president and general counsel for Hastings Mutual Insurance. It was shortly thereafter the litigation was commenced in Barry County, Mich., seeking to have that rule declared invalid.
On April 25, Judge James Fisher of the Barry County Circuit Court ruled in favor of the insurers, finding the prohibition on the use of insurance scores was contrary to Michigans insurance code and stating the administrative rule was an attempt to rewrite the insurance code through administrative rule-making. The state has appealed the decision.
[The state] apparently feels scoring is not authorized as a discounting tool under our insurance code framework for personal lines insurance, and we disagree on that point, says Puerner. Apparently, [that state] also believes insurance scoring is not an accurate predictor of loss, which again we disagree with.
When the litigation was first filed, insurers won an injunction that delayed inception of the rule until the issue was resolved in the court. Insurers can continue using scoring while the state court of appeals considers an appeal.
Risk Management, Modeling, and Actuarial Tools
3i Infotech, Inc.
Edison, N.J.
732-225-4242
www.3i-infotech.com
AIR Worldwide
Boston, Mass.
617-267-6645
www.air-worldwide.com
Algorithmics, Inc.
Toronto, Ont.
416-217-1500
www.algorithmics.com
Aon RiskConsole
Chicago, Ill.
704-343-8661
www.aonriskconsole.com
@Global, Inc.
Golden, Colo.
800-419-4449
www.atglobal.com
Brightwork, Inc.
Chicago, Ill.
773-529-0465
www.brightworkinc.com
Camilion Solutions, Inc.
Thornhill, Ont.
416-346-3713
www.camilion.com
CheckFree Corp.
Norcross, Ga.
678-375-3000
www.checkfree.com
CMS Bondedge
Santa Monica, Calif.
310-479-9715
www.cmsbondedge.com
COSS Development Corp.
Mequon, Wis.
262-241-8989
www.cossdev.com
CSC
Austin, Tex.
800-345-7672
www.csc-fs.com
CS Stars
Amarillo, Tex.
800-858-4351
www.csedge.com
Delphi Technology
Boston, Mass.
617-259-1200
www.delphi-tech.com
Digital Sandbox, Inc.
Herndon, Va.
703-871-3980
www.dsbox.com
Duck Creek Technologies
Bolivar, Mo.
866-382-5832
www.duckcreektech.com
e2Value, Inc.
Stamford, Conn.
203-975-7990
www.e2value.com
Effisoft
Boston, Mass.
617-437-9600
www.effisoft.com
Envision Technology Solutions
Salt Lake City, Utah
801-568-1818
www.envision-ts.com
EQECAT
Oakland, Calif.
510-817-3100
www.eqecat.com
Explore Information Services
Eagan, Minn.
800-531-9125
www.exploredata.com
Fair Isaac
San Rafael, Calif.
415-472-2211
www.fairisaac.com
Fiserv Insurance Solutions
Cedar Rapids, Iowa
800-943-2851
www.fiservinsurance.com
FJA-US
New York, N.Y.
212-840-2618
www.fja-us.com
Forensic Analysis &
Engineering Corp.
Raleigh, N.C.
919-872-8788
www.forensic-analysis.com
GenSource Corp.
Valencia, Calif.
800-949-9192
www.gensourcecorp.com
Infoglide Software
Austin, Tex.
512-532-3500
www.infoglidesoftware.com
INSTEC
Naperville, Ill.
630-955-9200
www.instec-corp.com
Insurance Information Exchange
College Station, Tex.
800-683-8553
www.iix.com
International Security
Technology, Inc.
New York, N.Y.
212-557-0900
www.ist-usa.com
Motion Computing
Austin, Tex.
512-637-1100
www.motioncomputing.com
MountainView Software
Kaysville, Utah
888-533-1122
www.mvsc.com
Portellus
Irvine, Calif.
949-250-9600
www.portellus.com
Risk Management Solutions
Newark, Calif.
510-505-2500
www.rms.com
Risk Sciences Group, Inc.
Schaumburg, Ill.
800-619-0224
www.risksciencesgroup.com
RISKTRAC
Portsmouth, N.H.
800-294-3316
www.libertymutual.com/risktrac
RMIS/Lab
Hartford, Conn.
860-543-7341
www.rmislab.com
Safe Risk Management Systems
Baltimore, Md.
410-230-0022
www.saferms.com
SAS Institute, Inc.
Cary, N.C.
919-677-8000
www.sas.com
Searchspace
New York, N.Y.
212-422-5100
www.searchspace.com
SilverPlume
Boulder, Colo.
800-677-4442
www.silverplume.com
SOLCORP
Toronto, Ont.
416-673-9900
www.solcorp.com
SRC Software
Portland, Ore.
503-608-3300
www.srcsoftware.com
SunGard
Miami, Fla.
305-858-8200
www.sungard.com
TDM-Ltd.
Aberdeen, U.K.
44 (0)1224 209 234
www.tdm-ltd.com
Thazar
Frisco, Tex.
972-377-1110
www.thazar.com
Valen Technologies
Denver, Colo.
720-570-3333
www.valentech.com
Valley Oak Systems, Inc.
San Ramon, Calif.
925-242-4600
www.valleyoak.com
Vanguard Integrity Professionals
Las Vegas, Nev.
702-794-0014
www.go2vanguard.com
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.