A Nose for Fraud

The game of cat and mouse between auto insurance fraudsters and the carriers who pursue them may well be an eternal one, if you ask Ben Kosic.

The CEO of Canatics, an anti-fraud non-profit, concedes that the nature of the business and the amount of money at stake may make it impossible for insurers to completely eliminate cheats from the system.

“But that doesn’t mean they always have to be four steps behind,” Kosic says.

For the past two years, his organization has helped insurers “close the gap” on their perpetual foes, increasing the chance that carriers will identify fraudulent claims before paying out on them.

“Fraudsters are ultimately in it to make money, so if you can catch up before the payment, they will move on,” Kosic says.

Canatics works by scouring claims information from multiple large insurers for suspicious patterns and connections gleaned from millions of pieces of data surrounding the drivers, vehicles and locations involved in collisions. Red flag alerts are then passed on to individual members on a weekly basis for further investigation.

The formation of the organization, whose eight members account for about 75 per cent of Ontario’s auto insurance market, is just one example of how insurers across the country are harnessing the power of data analytics to fight fraud in the system.

The power of analytics

Further west, the Insurance Corporation of British Columbia (ICBC) enlisted the services of security firm BAE Systems to create a similar tool to boost its fraud fighting efforts at the end of 2016. The insurer estimates that fraudulent or exaggerated claims together account for between 10 and 20 percent of its annual claims costs for a maximum of $600 million per year, or the equivalent of $100 to $200 for every ICBC policyholder.

Chris Fairbridge, the manager of ICBC’s special investigation unit, says the insurer has been on the lookout for new tactics as part of its effort to keep premiums down for the province’s drivers. Spiralling claims costs have contributed to an insurance rate rise of around 30 per cent compared to 2011 levels, with ICBC recently warning they could go up by as much as a further 40 per cent by the end of the decade.

And the analytics tool has already had a transformational effect on Fairbridge’s department. After running data from the last three years of claims through it, the insurer generated more than 3,000 potential fraud alerts on open cases, prompting around 90 new investigations into suspected fraud, with plenty more to come.

“It picks up links and connections our adjusters wouldn’t otherwise see when they go through a file,” Fairbridge says. “By getting flags as early as possible, we’re trying to become more proactive, rather than reactive.”

Even if cases slip through the net and are only later uncovered as fraudulent, Fairbridge says his team is not afraid to come after perpetrators. Civil courts have been used to claw back payouts, while provincial law also allows SIU officers to recommend criminal charges directly to Crown prosecutors. Since 2010, they have recorded around 100 convictions per year that resulted in penalties ranging from fines to short jail sentences. “This isn’t a victimless crime, and courts do take it seriously,” Fairbridge says.

To cope with the extra work, the SIU will boost its staffing levels by 50 percent, from 100 to 150 by the end of 2017, according to Fairbridge. But claims costs are projected to fall $21 million for policies written over the same period thanks to the new system and, by 2019, ICBC expects to be saving $44 million per year as a result of the tool. “The good news is that most people are making honest claims, and all these activities are intended to protect them. It’s only a small percentage that are costing everybody,” Faribridge says.

Gordon Rasbach, a former police detective who is now vice-president of legal and fraud management at Aviva Canada, says he’s encouraged by the industry’s general embrace of data analytics in its efforts to fight fraud.

“There has been a lot of investment in technology to help,” he says.

Twenty years ago, when he first entered the business, “insurance companies didn’t have fraud units,” Rasbach says. “Now they all have a fairly substantial one.”

He identifies 2012 as a big turning point. That’s when an Ontario anti-fraud task force reported back to the provincial government, urging the industry to take a more aggressive approach to combating fraud. A KPMG study commissioned by the task force pegged the cost of fraud at between $800 million and $1.6 billion per year.

A year later, Canatics was formed around the same time the provincial government pledged to cut average auto insurance premiums by 15 percent. However, getting the non-profit off the ground was a feat of governmental and industry relations. Canatics had to wait two years before issuing its first alert in 2015, while it dealt with the myriad of competition and privacy concerns that come with an attempt to combine private and personal information from customers of direct competitors.

The software gap

Still, Kosic says it was worth the effort, since the pooling of data from multiple insurers is key to the program’s success.

“Collaboration is the way to make a difference with this type of fraud. In the past, it was easy for fraudsters to hide in the data by moving around from insurer to insurer, or by subtly changing their names or other information about themselves,” he says.

The system only identifies suspicious patterns, so it’s up to individual members to decide how they act on the notification from Canatics.

“We do the triaging by giving them the best possible lead,” Kosic says. “They are finding a lot of stuff, ranging from theft rings to vandalism to the more traditional accident benefit abuse.“

Kosic says Canatics hopes to expand its membership beyond Ontario in future, noting that fraudsters don’t respect geographical boundaries.

“When you put the clamps down in Ontario, these people don’t go and get jobs. If it’s harder to detect fraud in Alberta, then they’ll move. There’s no reason to wait for auto insurance fraud to become a crisis before you do something about it,” Kosic says.

According to Rasbach, individual insurers are also making their own investments in software, designed to cut off fraudsters even further upstream by denying them access to policies in the first place.

“We can’t go on forever chasing people who commit fraud after the fact. We’ve got to find a way to stop them ever getting an opportunity to do it in the first place,” he says.

However, he says early advances in the area are based on similar tools used in banks that don’t translate as well into the insurance field. As a result, Rasbach says predictive technologies have so far delivered relatively high false positive rates, meaning that many innocent customers could end up facing extra, unnecessary scrutiny.

“If you’re not careful, at some point that will start to impact on retention levels of your own customers,” he says. “We need to send software companies back to the drawing board to come up with products that are more tailored to our business.”

A ‘firewall’ of brokers

Rasbach says brokers are a big asset to insurers in stemming fraud before policies are issued. He says they act as a kind of firewall that will become even more important in future, as technological advances allow more consumers to purchase insurance online without any human contact.

“Fraud at the policy end is much harder to do with brokers involved than it is when you’re filling out a form on the Internet. That’s a huge issue for companies that rely on direct business in any way,” he says.

David Elliott, president at Elliott Insurance Services in Port Hope, Ont., says he calls on his background in financial investments when meeting clients for the first time. “There, everything operates on the basis of knowing your client. As brokers, we try to do that too, instead of simply securing them a policy. You go a long way to avoiding fraud when you take a little more time to talk to your client.”

When a potential customer hesitates over answers or raises Elliott’s suspicions, he takes the opportunity to educate them on the impact of insurance fraud.

“It means higher premium costs for everyone but, on an individual level, it could also mean denial of claims, fines and even imprisonment,” he says.

That personal touch also makes brokers useful in instances of opportunistic fraud, when genuinely injured individuals pad out their claims to include unnecessary and unwarranted benefits or fictitious losses. Insurance Brokers Association of Ontario spokesperson Brett Boadway says most brokers have a sixth sense about claims that “don’t smell right.” Brian Purcell, the organization’s president-elect, remembers receiving one claim from a customer following a car fire who probably wasn’t aware of the broker’s side job.

“I’m a volunteer firefighter, so I was the one on the end of the hose putting the fire out, and there seemed to be a lot more in the car according to the claim than what I saw,” Purcell says.

“Sometimes it’s people trying to get a little bit more than they are entitled to, rather than some big money laundering scheme,” he adds.

Elliott says a tip-off to the insurer’s claims adjuster is often a good option for brokers who suspect claims exaggeration.

“You take it case-by-case, because you don’t want to accuse someone of fraud without any evidence. If there’s something glaring, then you can always call in the police,” he says.

__________________________________________________________________________
Copyright © 2017 Transcontinental Media G.P. This article first appeared in the April 2017 edition of Canadian Insurance Top Broker magazine

This story was originally published by Canadian Insurance Top Broker.