At last week’s Premier Business Leadership Series conference in Orlando (which, by the way, was sunny and warm – unlike the snow I returned to in Connecticut) three insurance thought leaders gave a presentation innocuously titled “The Latest Advances in Fraud Detection and Prevention for Insurance.” Now, for the non-insurance people of the world, this topic may sound a little dry, but I came away with a whole new way of thinking about fraud, its impact on me as a consumer of insurance, and (as consumers) our responsibility and expectations around combatting fraud.
While the true dollar impact of fraudulent claims is almost impossible to measure, the Insurance Information Institute estimates that it represents approximately 10 percent of all property/casualty insurance claims: In the 2005-2009 period, the dollar impact of that fraud totals $30 billion. So now you’re saying – fine, the insurance company just writes that off and it doesn’t impact me. Wrong. Fraudulent claim costs are in many ways passed down to all policyholders in the form of increased premium: You’re paying for the bad behavior of others. So, it’s in everyone’s best interests to limit fraudulent activity: Insurers, regulators and policyholders.
In his introduction, Allstate’s Frank Llende told us that personal lines insurance fraud is on the rise: “I’ve never seen anything like the economy we’re in today. I don’t know if it’s increased Medicare or Medicaid fraud, the soft economy or more sophisticated crime rings … but our customers are being attacked and we have to be good stewards of those [premium] dollars. Our customers have a high expectation that we are guarding their dollars. Our customers expect us to be financially sound.”
And CNA’s Tim Wolfe says that “Fraud is hot due to a downturn in the economy…[in commercial lines insurance, where the claim severity is much higher] a couple of big claims could put an insurer out of business. What we’re seeing is more grandiose schemes [from providers], especially around medical billing, equipment, services and transportation. The fraudsters are slick in the way they set up the billing – they know the insurance company’s weaknesses.”
As fraud becomes more sophisticated, insurers are changing their game as well. Here are some of the ways that the industry is evolving their capabilities:
- Data-driven decision making: At Allstate, “Over the past couple of years, we’ve transformed our home office into an innovation environment. We’ve started to look at data and how it can influence decisions. For us, we always used data in the underwriting process, but we hadn’t looked at how we could use data to manage resources [related to the claims process]. We used to hire ex-military and police, but now we’re hiring modelers [in our SIU]. What we’ve been able to do for the enterprise has been phenomenal. We can show the business where our customers’ dollars are being attacked and allow them to make better decisions about the business: There is no room for fraud.”
- Empowering the field to make smarter decisions: At CNA, “Typically, the SIUs have provided training to adjusters on the [fraudulent] “flags” when they’re in the field. We hope that adjusters recognize the flags and inform the SIU – but what we found was that most of the referrals into the SIU came from a tiny group of adjusters; they weren’t consistently used. We’re making it easier by providing the information that they need to make better decisions and referrals. We’ve started to implement predictive models and…[we can use those models to] help educate the adjuster in following their instincts. We’re finding that our mindset needs to change” from going to a reactive to a proactive model.
But it’s not all about finding the bad. Llende reminds us that “if you can identify folks quickly that aren’t associated with fraud, you can give them a better customer experience.”
The evolution of these SIUs from “gumshoe detectives” to analytic leaders can be a challenging journey because the influencers of fraud impact every functional area within the organization. “There are clear and distinct data points that help us identify fraud: We can identify these data points and [operationalize them] early in the process, be more predictive, and use more analytic tools to better manage resources. We want to think about a different way of collecting information at the start of the claims process. We want to work with our product partners to see if there are products that are attracting more fraud and ultimately influence product design.” Both insurers echo that fraud detection is not all analytics: “We still need feet on the street. The technology [and analytics] help you identify cases with fraud detection, but you still need people out there doing the hard work and asking the right questions.”
And a final thought from Llende: “I never thought I would find myself recruiting modelers, but there is intense internal competition for these jobs…because the problem is so interesting to solve.”
*NOTE: Originally published in The Analytic Insurer.