Fighting Fraud in a Faltering Economy

Insurers Use Predictive Modeling, Search and Rules-Based Technology to Identify Fraud

It wasn't a big insurance fraud case by any means. Last year, a Pennsylvania homeowner reported a break-in and the associated property loss totaling $13,500. During an investigation, MetLife Auto & Home discovered alterations on the receipts that were submitted by the woman for the stolen electronics, and the claim was denied. Even worse, a neighbor reported that the homeowner had moved away weeks before the alleged break-in.

The homeowner was arrested, sentenced to two years of probation, and fined $500. This type of opportunistic fraud is not uncommon, but special investigative unit (SIU) investigators fear it's the tip of the iceberg in America's faltering economy. SIUs in both personal and commercial lines are bracing for a new wave of fraud. Those who were around during the last major recession in the early 1990s remember the spike in fraudulent claims and the emergence of highly organized fraud rings. Today, soaring gasoline prices, falling real estate values, and unprecedented financial crises have formed a perfect storm of economic circumstances that tempt ordinary people to commit fraud.

"Desperate people do things they normally wouldn't do," noted Michelle King, manager of Claim Auto Property Damage & Subrogation at Unitrin Business Insurance (UBI), a division of AmTrust. North America. "Fraud has always had the problem of appearing victimless to the average person. Many people just don't see shaving a deductible or inflating a claim as the same kind of wrong as stealing a neighbor's lawnmower. Obviously, burning your own car up or faking a theft is completely different. But at the point when someone has already submitted a few inflated claims here and there - when they're really desperate - committing insurance fraud is just the next step."

Gas Guzzler Claims

The overarching impact of the U.S. economic crisis means insurance companies need to be vigilant on all fronts - auto, homeowners, and business. Insurance companies are exploring ways technology can help consistently identify suspicious claims as early in the process as possible. Carriers such as Unitrin and MetLife Auto & Home are using software, predictive modeling techniques, and external database searches to continually improve their detection capabilities and stay ahead of the latest trends in fraud.

According to the Coalition Against Insurance Fraud, evidence that auto fraud is on the rise already exists. An October report by the coalition noted a 14 percent increase nationally in vehicles recovered burned from 2006 to 2007. Throughout the summer of 2008, as gas prices crossed the $4 mark in many parts of the country, arson investigators saw spikes in cases by as much as 25 percent in states such as New York, New Jersey, and California. The coalition even noted alarming increases in smaller states such as South Carolina, Mississippi, and Utah.

MetLife Auto & Home SIU Director John Sargent says his team has noticed "incremental increases" in such claims. In fact, earlier this year the company fine-tuned its internal software for scoring suspicious claims with something informally called the "gas guzzler" rule.

MetLife's fraud evaluation system scores all claims at first notice of loss and re-scores them as new data elements are captured. When the aggregate score reaches a threshold defined by the company, the claim is electronically referred to the SIU. The new rule automatically scores claims involving high-end, low-mileage vehicles with eight cylinders or more with indicators such as arson, theft or a single-vehicle accident without injuries.

Even though gas prices fell late in 2008, economists report that consumers are still more likely to shop for cars with high fuel economy, which means the resale values of SUVs, Hummers and high-end trucks are plummeting. Gasoline prices aren't the only factor. Many Americans awakened to the fact that they are saddled with debt because of the depreciation of the values of their homes and automobiles.

King, who is based in UBI's office in Dallas, points out some anecdotal cases of expensive commercial vehicles reported stolen.

"When somebody parks a car on the border and leaves it, you've got to wonder if there's something suspicious going on, but proving it is another challenge," she said. "The value of some of these vehicles has dropped from $75,000 to less than $45,000, but with gap insurance we're basically paying off their notes. That certainly is fortuitous for somebody who is very deep in debt."

In addition to internal systems, companies are also using external databases to focus on auto-related fraud. MetLife Auto & Home, for example, is working on a project to interface its system with an external national claims database. With rules in place to search the database for related claims, salvage history, claimant history, and other factors, the system will score suspicious claims at an early point in the process.

"We have been doing this on a manual basis, but it will become part of the fraud-scoring process," said Jim Murphy, SIU business consultant and manager of the SIU system at MetLife Auto & Home. "So if we have a VIN number for a vehicle that the database shows was previously salvaged, it will be scored and possibly sent to an investigator. At the very least, we'll know the vehicle was a prior salvage. Overall, it's going to help our analysts in the triage process."

Another critical area involves homeowners' claims. Banks throughout major regions of the country report a record number of foreclosures and falling real estate values, leaving a growing number of homeowners with high monthly costs and an unwieldy amount of debt. In 2007, MetLife added rules for homeowners' claims involving fire, theft or mysterious disappearance to its system.

"It's been very successful for us," Murphy said. "Another thing we've been doing is taking that data in a batch and working with a public records service to search for pre-foreclosure or foreclosure information. So far, we've gotten some very good information on our homeowners' claims. You can definitely see a correlation in the data."

The Unlucky Worker

Arson and theft tend to plague commercial lines companies as well during economic downturns. However, with the continued climb of medical costs, workers' comp policies are ripe for fraudulent activity. To identify suspicious claims faster, UBI included workers' comp rules in its fraud evaluation system to analyze factors such as severity, injury type and claim frequency.

"When you suddenly see that someone has filed a fourth workers' comp claim, you start wondering how unlucky one person can be," King said.

The capability to perform similarity searches on addresses and other data has helped to uncover claims by travelers, professional fraudsters who typically move from place to place and one scam to the next. Identity theft is increasingly a factor in fraudulent claims. Last year, for example, a Philadelphia man was convicted of 62 counts of insurance fraud and 22 counts of identity theft. Police say he had filed 70 claims on multiple policies with seven different insurance companies. He had identified himself as a police officer, an accountant, a bounty hunter, a soldier, and also cited various other occupations when talking to numerous adjusters.

"We've seen lots of claims with the same address even though the name is different," King said. "These travelers are usually very good about using fake names and social security numbers, but sometimes they get sloppy. Eventually they're going to want the check to come to them in the mail. And if the address they gave has some commonality with other addresses involving suspicious claims, then the system will flag it."

The company adds in other factors, such as publicly sanctioned doctors and lawyers to develop a unique score for each claim. While a claim might appear to be run-of-the-mill at the first notice of loss, additional information gathered during the claim-handling process may flag it as suspicious.

"All claims are continuously rescored as we get more information," King said. "It may be rescored a half a dozen times, and then something like a social security number, a diagnosis code or a doctor's address is added that pushes the claim over the threshold. Our adjusters still flag suspicious claims but your ability to detect fraud is only as good as the person who flags the file. On any given day, a human being can look at something and just miss it. The system looks at every claim objectively, and if certain criteria are met, it's going to alert the SIU."

Outwitting Organized Fraud

While opportunistic fraud is on the rise, organized fraud rings show no signs of slowing down. If anything, they're better organized and attracting even more members. In October, for example, New York police rounded up 61 people including the owners of a Manhattan medical clinic in an alleged fraud ring that bilked insurers for an estimated $1.6 million in claims for staged accidents.

"I'm not sure if it's related to the economy or just pure competition, but we have seen a big increase in the amount of money a typical runner gets for referring a patient," said MetLife Director John Sargent. "In places like New York, it's up to $3,000 to $4,000 per patient, compared to the $2,000 they used to get. And in other places around the country, they're getting $1,000 instead of $500. So, the incentive is definitely out there."

Using external sources such as the National Insurance Crime Bureau (NICB) database, MetLife Auto & Home has used predictive modeling to create fraud detection rules for staged accidents typical of organized fraud rings. In addition to looking at the claim type, the company looks at the relationships between the individuals involved in the accident as well as their medical providers and attorneys.

"Fraud rings are getting more sophisticated at committing these types of crimes, so it's really crucial that we use all of the tools at our disposal to stay ahead of them," Sargent said.

A number of years ago, Unitrin investigated a claim by a policyholder whose car caught on fire in a vacant parking lot. "He had done a really good job except for the fact that the camera at the convenience store nearby had a wide angle lens and caught him burning up his own car," King recalled. "When you have video tape, it really changes the dynamics of the case. Otherwise, isolated incidents are hard to prove. That's why it helps when you can pick up a pattern."

There's no substitute for the basic investigative work needed to properly investigate claims but a growing number of companies are using technology to identify fraudulent patterns and relationships. Not only does it get higher-quality referrals to the SIU faster, but it also helps investigators work smarter and make the best use of their resources.

This article, excerpted from Claims magazine, was written by Ed Charlton, vice president of Claims and Legal Solutions for the Property and Casualty Insurance Division of CSC, which supports more than 750 insurers and self-insured organizations. Charlton has spent more than 25 years focusing on claims and litigation management.