At this time of year, insurance claims rise dramatically – but many of them are fraudulent, writes William Lawrence, regional practice lead: fraud and financial crimes at SAS Institute.
Mike overindulged at the year-end office function this weekend. While driving home after a few too many beers, he veers off the road and crashes his BMW into a tree. His car is in bad shape but Mike knows that if he reports the accident to his insurance company now, his claim will be rejected as he has been drinking. He also knows that he can’t afford to pay for the damage himself, especially since he spent his bonus on a holiday for this family.
So Mike waits a few days before submitting his claim, telling his insurance provider that he drove into the tree after swerving to miss a dog while driving to the shops at 1pm on Monday afternoon – not that he was drunk at 3am on Saturday morning. He convinces himself that, since he’s paid his premiums regularly and hasn’t claimed in years, he doesn’t need to feel guilty – especially as there were no witnesses to verify his story.
The volume of insurance claims increases dramatically during December and January for a number of reasons. Increased holiday traffic and more social events – and therefore more people driving under the influence – could potentially mean more accidents. Also, as families head to the coast, homes sit empty for weeks and are easy targets for burglars. And all this happens at a time of year when we experience extreme weather patterns, resulting in more claims for hail and flood damage.
By the time the new year rolls around, people are generally under financial pressure from having spent too much over the festive season and worry about how they’re going to pay January’s bills. While they might have a legitimate loss event, these unique circumstances could motivate policyholders to submit false claims or to inflate their claims – for example, while claiming for a pair of sunglasses and a cellphone that was stolen out a car, the policyholder throws in an iPad and camera, even those these were never in the car.
Too many claims; too little resources
Insurance companies have a limited number of claims investigators, making it impossible to scrutinise every claim for fraud. As a solution, they may apply a claim threshold, say of R20 000. Any claim under this amount will not be reviewed and will be immediately processed for payment. This means thousands of fraudulent claims could be slipping through the cracks, amounting to huge losses.
Separating the wheat from the chaff
With advanced analytics, insurers have the ability to scrutinise every single claim. By applying robust fraud rules and models, claims with a high fraud propensity are automatically flagged. Claims considered low risk will be immediately processed for payment while high-risk claims will be sent to the fraud department for further investigation, allowing insurers to better allocate their investigative resources. To achieve this, analytics systems need to be fast and flexible and provide the ability to create new fraud rules in response to trends and unique circumstances.
For example, new policy applications received in June might not be subject to the same level of scrutiny as those received in the first weeks of January, especially if the policyholder submits a claim shortly after his application was approved. This could suggest that the loss event happened before he had cover in place, or that his claim was rejected by another insurer and he’s now taking a chance with a different provider – what we call cross-carrier fraud.
In the past, it would take weeks to settle an insurance claim, with the onus often on the policyholder to show why the claim should be paid. But competition in the short-term insurance market has forced providers to settle claims quickly to retain customers. Insurers need to strike a balance between managing the customer experience by processing legitimate claims as soon as possible and limiting fraud.
Having the ability to segment claims using analytics allows insurers to quickly separate the genuine claims from the suspicious ones. Insurers that use analytics on an ad hoc basis are being curtailed in their ability to respond quickly and could lose customers – and more money to fraud – as a result.
It is estimated that fraud costs the short-term insurance industry anything between R6-billion and R8-billion annually. Having the ability to weed out fraud means insurers won’t have to compensate for losses by hiking premiums, resulting in happier, more loyal customers.