With $25 billion in fraud reported by the U.S. property and casualty insurance industry last year, it cant be said the industry is winning the battle against fraud. The claims department and the special investigation unit are on the frontlines in this battle, but some insurers believe it is time to bring in the heavy artillerythe IT department.
CIOs are starting to get more involved, says Ralph Carlile, one CIO moving in that direction for Grange Insurance Group in Seattle. There are systems implications when you start bringing in a third-party system that will help detect fraudulent areas. Those systems have to be purchased and integrated into everything else.
Bill Kenney, CIO at Colony Insurance Group in Virginia, says he is not overtly involved in fighting fraud for the E&S insurer, but his IT department works closely with other departments on various projects.
Its not just fraud-fighting software that is drawing the IT department into this fracas. CIOs can contribute to the effort by nurturing what TowerGroup consultant Jamie Bisker calls an insurers greatest natural resourceits data. The basic weapon is having clean data so you can do data mining, cluster analysis, pattern matching, rule induction, and just basic analysis of your data, says Bisker, director of research for insurance practices for the Needham, Mass.-based research group.
We do a lot of due diligence on our data, Kenney says. We cross check it and balance it. Colony has a balancing team, made up of IT and accounting staffers. They make sure the transaction system and the data warehouse system balance. The nice thing about the client/server environment is we have edits on the front end that basically narrow down what [agents] are allowed to enter. Once we set that criteria, a data-entry person could sit there and beat the system up all day long, and the data just wont go in.
Carlile indicates his role in fighting fraud for Grange will increase substantially over the next year as the company moves data from its mainframe to a relational database. He says this would allow Grange to implement data marts. In order to move data into the new system, we have to convert it, he notes. While were moving one line, there are other parts of that line being transacted off the mainframe. If we dont move the transaction and the data together, we end up with two different systems that cant communicate with each other.
Carlile believes it will be worth the effort. The SQL Server 2000 comes with a variety of analytical programs, he says. Once thats in place, we can look at how to slice and dice the datadoes a claim fall into a template so the system will raise a flag to our adjusters or the SIU?
The Need for Speed
Having quick access to the data is important. Our data warehouse is invaluable because we can get at the data quickly, Kenney says. The storage devices are so cheap now. When we were buying mainframe data drives, they were $50,000 to $60,000. Now you have data disc drives for $5,000 to $10,000. And with the Internet, you have data available everywhere.
Bisker believes insurers need to be more proactive in fighting fraud. Traditionally, its been a retroactive situationthey had to wait until a claim was closed to analyze it. Then you had to find a claim you knew was fraudulent, look it over, understand it represents a pattern, and look for a pattern in the data you have, he says. There is nothing wrong with a reactive system. Im just advising the reaction time be shortened. There is technology to do that.
He asserts one of the biggest steps an insurer can take is to cleanse the data it has in its own database. An insurance company is an information processing company, he says. They get information about people, products, property, cars. They need to treat the data they have just like they would treat natural resources. They need to put the same type of energy and effort into data management, data quality, and data engineering.
This can be done by periodic edits of duplicate records, common storage schemas, cleaning up the data by having a schema, and not having customer data spread across 14 different files, Bisker says. Just like software engineering cleaned up the way code was written, we need the same type of structure and routine for data.
You wont get any argument from Carlile. Its a huge issue, he says, adding that throughout the industry, 20 percent of vehicle identification numbers (VINs) collected by insurers are incorrect. Grange is no exception, he admits. That takes a lot of clean-up effort, he says. Name and address consistency is another data problem, often going back to the agents filling out a policy application. Every insurer has horror stories about bad data, and Carlile relates one about a business having a dozen different addresses for the same building in its files. With most applications being done electronically today, he says, insurers can force an application to have a particular format for the data.
Grange is using a tool from ChoicePoint that matches the VIN to the most likely make of the car. The insurer can then match that information against the data on the policy. You make a presumption thats the correct number, he says. Validating Granges data is part of the overall project of moving data off the mainframe.
Kenney also supports the effort to achieve clean data. It is important that the data going in is correct, he says. If we catch [mistakes] at the point of sale, it makes it a lot easier because cleaning it up is just cumbersome.
Carlile does believe if an insurance company establishes a reputation as being tough on fraud, it will scare off some potential fraudulent claimants. There are a lot of benefits to being considered not friendly, he says.
Bisker also thinks insurers need to pressure the states to recognize insurance fraud as a serious crime. The industry has been successful in a lot of states in raising the stature of such activity to a major crime, but in some states it is still considered a misdemeanor.
He believes CIOs will be working closely with claims people in the future. The CIO has to be aware the technology can help prevent fraud, and the technology necessary for data cleansing and data engineering are a part of a good IT department anyway, he says. Its really just closing some loopholes, tightening some controls, and doing a better job of preventing fraud or at least identifying it.
With the figure at $25 billion (and that is just what is reported), it would seem to be worth the effort.
Insurers Join War on Terrorism
Its hard not to be supportive of the federal governments war against terrorism, but the insurance industry has to be a little nervous about what implications the Patriot Act will have. Sometime this fall, guidelines for the industry will be handed down for such areas as claims and underwriting, and until then, there is speculation about its impact on the industry.
Jamie Bisker, director of research for insurance practices with the research-consultant firm TowerGroup, believes life insurers in particular have more to be concerned about with the law, passed in November 2001 in the wake of terrorist attacks against the United States. Of particular concern is the anti-money laundering aspect of the law.
MIB Group (www.mib.com) currently provides life insurers with data, such as the governments list of Special Designated Nationals and Blocked Persons developed by the Office of Foreign Assets Control (OFAC). The Patriot Act prohibits U.S. businesses from engaging in transactions with people on the OFAC list. Violators face fines, prison sentences, and civil penalties.
The OFAC list is also available on the Treasury Departments Web site (www.treas.gov), and that is where Colony Insurance Group, a Virginia-based E&S insurer, goes to compare names on its policies with the OFAC list. Bill Kenney, Colonys CIO, says the company has added the list to its database so agents can check the list first, but he is worried the list is getting unwieldy. They started off with about 75 names, and now its up to over 50,000, he says.
Further complications arise when the people you are looking for are constantly changing names or appearances. If you are making a concerted effort to check this list, and we are, I think thats what they [the government] wants to see, Kenney says.
Bisker believes life insurers have to be careful with whom they are dealing when certain types of policies are purchased. He points to three areas of concernsingle premium life insurance policies, loans out of life insurance policies, and viatical loans. Those mechanisms will be given closer attention, he says, adding all insurers will have to preserve more data on their customers.
But Bisker notes there are other things insurers need to check as well. There will be different kinds of scenarios being looked at, such as whether a policy makes sense for an insurable interest and where the money is going, he says. If its going outside the U.S., thats certainly going to raise a flag.
For the P&C industry, the Insurance Services Office has the go-to database for claimsClaimSearchand Vincent Cialdella, vice president of ISO-ClaimSearch, says once the government hands out its regulations, insurers will be able to adjust quickly. Weve been approached by the industry to help it out with the Patriot Act, he says. Companies send claims to ClaimSearch for initial screening, so it stands to reason they would want to incorporate the process of checking those records against a list provided by the government as part of the ClaimSearch flow.
© Touchpoint Markets, 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.