Canadian and U.S. insurance companies' fixation on Y2K have left them wide open to high-tech scams
Canadian automotive insurers are taking a major hit from organized and computer literate criminals who are staging crashes and taking advantage of dirty data in corporate databases, according to an investigation launched by an American software company.
As part of a U.S.-based insurance fraud investigation, Jay Valentine, CEO of Austin, Tex.-based InfoGlide Corp., and his team of database investigators have been reviewing the databases of three Canadian insurance companies. Due to confidentiality agreements, Valentine can't identify the companies he has been investigating, but says they are among the "top 10" and the findings are not good.
To detect fraud, InfoGlide takes an insurance company's data and -- although the data is in multiple formats -- searches across databases using its similarity search engines. After about three weeks "they're at the point where we can show them a sophisticated fraud," Valantine says.
Valentine says typical perpetrators of this type of crime are computer literate professionals who recruit programmers to use the "technology paralysis" created by the Year 2000 problem to hide them. The programmers take advantage of built-in limitations of neural net technology and relational databases used by most insurance companies.
The key to the scam is the fact that databases use exact match searching to detect multiple billings and other fraudulent practices. The fraud artists, however, change various identifiers -- from single letters in policy holder names to single digits in social security numbers -- to take out multiple policies on a single individual.
Valentine said Canada was part of a study the company did on the Y2K issue, which found insurance companies have so many resources dedicated to getting their billing systems ready for Y2K problems that there's no money available for sophisticated fraud detection.
"We actually interviewed some people who were under indictment that we had discovered as part of our fraud work. Although they were not necessarily in Canada, they were victimizing U.S. insurance companies and their comment was that insurance companies were really tied up on Y2K issues and they were perpetrating fraud for that reason," he said.
The situation in Canada, Valentine says, is similar, but the frauds being perpetrated are more like "family endeavors" than organized criminal activity.
Valentine describes the similarity search engine as a new type of database management system (DBMS). While relational databases and neural nets require the data to be cleansed, a similarity DBMS takes data that may be incorrect and unscrubbed -- including different spellings for a person's name -- and finds correlations. It can take misspellings and derivations and track the data to one person.
"What we've done is created very sophisticated profiles from lots of data," he said. "That data may have hundreds of fields -- many of which are incorrect -- but because they are similar to, in some degree, profiles from other people, there's some correlation." Fraud artists may be able to change various identifiers, "but they can't change everything, so we find them."
Once the data search is complete, the insurance companies are given the results. The affected companies then usually attempt to file a lawsuit against the fraud artists in the hopes of reclaiming their losses, rather than calling police. Valentine said the companies don't file criminal charges because such charges are very difficult to prove and result in negative attention.
"They don't want a lot of publicity over being victimized, they just want it to stop happening to them and let someone else be victimized."
Valentine says there are about four Canadian law firms recruiting people to cause car crashes and then filing multiple claims. The perpetrators are likely one law firm operating with different names or several lawyers doing business under several names, he said.
The biggest fraud in Canada found during the review was perpetrated through several insurance firms and generated $56 million (U.S.) for one fraud ring, Valentine said.
The Canadian Coalition Against Insurance Fraud helps insurers become aware of insurance fraud and works to prevent it. Coordinator Ann Walker said staging auto accidents is a "significant insurance claims fraud in Canada" and the U.S. She added there are a number of "red flags" companies look for to determine the legitimacy of filed claims.
In fake auto accidents, a red flag could arise when a company compares the damage to a car and the lack of injuries of the people involved. "We tell companies in every claim that's been denied, there was always a red flag in the claim," she said. "Insurance companies are aware of them and know what to look out for."
All private insurance companies in Canada contribute to industry databases that keep track of claims and claim costs. Policies on fraud differ from company to company and the coalition doesn't have regulatory power over the companies.
"There are a number of industry tracking systems in place that these companies all report to and they can check these databases and make sure there's no claims history or no denied claims," she said. The databases are run by various industry associations including the Insurance Information Centre of Canada and the Insurance Crime Prevention bureau. They report different kinds of information on the databases and they consult databases according to the kind of policies they are writing.
Walker said insurance fraud costs policy holders $1.3 billion a year in North America -- second only in criminal profits to illegal drug sales.
COPYRIGHT 1998 Plesman Publications
COPYRIGHT 2000 Gale Group