Tag Archives: fraud

A Behavioral Science Scandal

A much-cited claim about how behavioral science can guide insurance has been exposed as fraudulent. The claim was made most prominently by Dan Ariely, a best-selling author and pioneer in the field of behavioral economics, who was Lemonade’s chief behavioral officer from 2015 through 2020. But the claim turns out to be based on fabricated data.

The claim was based on a study that Ariely and four co-authors published in 2012 in the Proceedings of the National Academy of Sciences. Ariely then cited the study at length in his 2013 book, The Honest Truth About Dishonesty, continuing his string of best-sellers that began with Predictably Irrational in 2008.

The study reported that people would be more honest if you asked them to promise to be truthful before providing information rather than having them provide the information and then certify that what they reported was accurate. In other words, you disrupt the usual process, in which people supply information and then just have to rationalize a bit of cheating afterward.

The study said it drew on nearly 13,500 customers of an auto insurer, half of whom signed a claim of truthfulness at the top of an application and half of whom signed at the bottom. The study reported that those who signed at the bottom said they drove about 10% fewer miles than those who signed at the top — and, of course, paid lower premiums as a result.

The conclusion was so appealing that the paper was cited more than 400 times in academic publications. Many organizations, including the IRS, began having at least some people attest to their honesty at the start of the process. I certainly fell for the idea. I couldn’t even tell you how many times I’ve cited the study.

More importantly, from the standpoint of insurance, Lemonade incorporated behavioral economics ideas into its initial business model that at least rhymed with the study’s conclusion, even if they didn’t specifically build on it. Lemonade took a set share of premium, to demonstrate to customers that it had no incentive to deny claims. Lemonade also said it would donate to specified causes if claims were below a set level — encouraging clients to minimize claims.

Other insurers surely built on the study, especially given Lemonade’s success (even though its use of behavioral economics seems to have mattered far less than its sleek customer experience and slick marketing).

The plot began to unravel as others tried and failed to duplicate the study’s results. Eventually, the authors published two retractions in 2020, in the Proceedings of the National Academy of Sciences and in Scientific American.

As part of the retractions, the authors published the original data — which is how it became apparent that the study was based on more than an honest mistake; the data had been manufactured.

Sleuths at Data Colada spotted what, in retrospect, were obvious problems. The data didn’t follow a Bell curve, as you’d expect. There weren’t some people who drove a little, some who drove a lot and a whole bunch who fell in the middle. Every division based on mileage had almost exactly the same number of people in it, from low mileage through high mileage, and not a single person out of nearly 13,500 drove more than 50,000 miles in a year. In addition, the mileage that people supposedly reported was accurate down to the mile, even though actual people would round off the numbers. The precision was a clear indication that a random number generator was being used.

There was more, too. In any case, when confronted with the Data Colada analysis, all the authors quickly agreed that the data had to have been faked.

At the moment, the focus seems to be on figuring out who to blame for the fraud. I confess to some personal confusion. I spent time with Ariely at a small, three-day conference where we both spoke in 2008 and found him to be extremely smart and thoroughly engaging, so I’d like to think that he wasn’t involved. (He has vigorously denied faking any data.) But he has said he was the only one of the five authors who dealt directly with the insurer that provided the data, and it’s not at all clear to me what the insurer would gain by faking the results. (While the company wasn’t initially named, it’s since been identified as The Hartford.) I’m also confused because he cited the study to me, personally, at that gathering in 2008 but didn’t publish the results for four years. Why wait so long with such an interesting result? (He’s on the record as having cited the study in a talk at Google in 2008, so he wasn’t just talking to me, either.)

But I’m more concerned with the broader point, which I think is this:

Behavioral economics is still a powerful tool for insurers despite this embarrassing fraud. We may like to think of customers as completely rational, but they aren’t, and we need to understand them as they are, not as we’d like them to be. That doesn’t mean accepting broad pronouncements about behavior, even from charismatic experts like Ariely. Understanding behavior means engaging with our own customers deeply, testing how they react to various actions on our part and then tailoring our interactions with them, foibles and all, to maximize benefits both for them and for us.

I realize that this is two weeks in a row where I’ve take a contrary view about technologies and techniques that are huge benefits to the insurance industry — last week’s was When AI Doesn’t Work. I’m sure these two Six Things commentaries aren’t the start of a trend. But I don’t believe that trees grow to the sky, so I don’t see the point in pretending they might. When there’s a problem, I’ll always try to point it out.



11 Keys to Predictive Analytics in 2021

According to Willis Towers Watson, more than two-thirds of insurers credit predictive analytics with reducing issues and underwriting expenses, and 60% say the resulting data has helped increase sales and profitability.

That figure is expected to grow significantly over the next year, as the inherent value of predictive analytics in insurance is showing itself in myriad applications.

Predictive analytics tools can now collect data from a variety of sources – both internal and external – to better understand and predict the behavior of insureds. Property and casualty insurance companies are collecting data from telematics, agent interactions, customer interactions, smart homes and even social media to better understand and manage their relationships, claims and underwriting.

Another closely related tool is predictive modeling in insurance, such as using “what-if” modeling, which allows insurers to prepare for the underwriting workload, produce data for filings and evaluate the impact of a change on an insurer’s book of business. The COVID-19 crisis has shown insurers that the ability to predict change is invaluable, and “what-if” modeling is a great tool for carriers that know they need to make changes but want to ensure they are doing it accurately. The right predictive modeling in insurance software can help define and deliver rate changes and new products more efficiently.

Using the plethora of data now available, here are 11 ways predictive analytics in P&C insurance will change the game in 2021.

Pricing and Risk Selection

This isn’t exactly a new use for predictive analytics in insurance, but pricing and risk selection will see improvement thanks to better data insights in 2021. Given the increased variety and sophistication of data sources, information collected by insurers will be more actionable.

Why do these data sets help predictive analytics improve pricing and risk selection? Because they are largely composed of first-hand information. Data and feedback collected from social media, smart devices and interactions between claims specialists and customers is straight from the source. Data that isn’t harvested through outside channels (such as the typical demographic material used in the past, like criminal records, credit history, etc.) is more direct and can provide valuable insights for P&C insurers.

But just how much data are insurers collecting from IoT-enabled devices? Some reports estimate it’s approximately 10 megabytes of data per household, per day, and that figure is expected to increase.

Identifying Customers at Risk of Cancellation

Predictive analytics in P&C insurance is going to help carriers identify many customers who require unique attention – for example, those likely to cancel or lower coverage. More advanced data insights will help insurers identify customers who may be unhappy with their coverage or their carrier.

Having this knowledge in hand will put carriers ahead of the game and allow them to reach out and provide personalized attention to alleviate potential issues. Without predictive analytics, insurers could miss credible warning signs and lose valuable time that could be used to remedy any issues.

Identifying Risk of Fraud

P&C insurance companies are always battling various instances of fraud and oftentimes aren’t as successful as they would like. The Coalition of Insurance Fraud estimates that $80 billion is lost annually from fraudulent claims in the U.S. alone. Additionally, fraud makes up 5% to 10% of claims costs for insurers in the U.S. and Canada.

Using predictive analytics, carriers can identify and prevent fraud or retroactively pursue corrective measures. Many insurers turn to social media for signs of fraudulent behavior, using data gathered after a claim is settled to monitor insureds’ online activity for red flags.

Insurers are also relying on insurance predictive modeling for fraud detection. “Where humans fail, big data and predictive modeling can identify mismatches between the insured party, third parties involved in the claim (e.g. repair shops) and even the insured party’s social media accounts and online activity,” according to SmartDataCollective.

See also: What Predictive Analytics Is Reshaping

Triaging Claims

Customers are always looking for fast, personalized service. In the P&C insurance industry, that can sometimes present a challenge. But with good predictive analytics systems, carriers will be able to prioritize certain claims to save time, money and resources – not to mention retain business and increase customer satisfaction.

Predictive analytics tools can anticipate an insured’s needs, alleviating their concerns and improving their relationship with their carrier. It can also contribute to tighter management of budgets by employing forecasted data regarding claims, giving insurers a strategic advantage.

Focusing on Customer Loyalty

Brand loyalty is important, no matter the product, and now insurers can use predictive analytics to focus on the history and behavior of loyal customers and anticipate what their needs may be. How important is brand loyalty? About half of customers have left a company for a competitor that better suited their needs. Also, this data can help insurers modify their current process or products.

Identifying Outlier Claims

Predictive analytics in insurance can help identify claims that unexpectedly become high-cost losses — often referred to as outlier claims. With proper analytics tools, P&C insurers can review previous claims for similarities – and send alerts to claims specialists – automatically. Advanced notice of potential losses or related complications can help insurers cut down on these outlier claims.

Predictive analytics for outlier claims don’t have to come into play only after a claim has been filed, either; insurance companies can also use lessons learned from outlier claim data preemptively to create plans for handling similar claims in the future.

Transforming the Claims Process

With predictive analytics, insurers can use data to determine events, information or other factors that could affect the outcome of claims. This can streamline the process – which traditionally took weeks and even months – and help the claims department mitigate risks. This also allows insurers to analyze their claims processes based on historical data and make informed decisions to enhance efficiency.

Advancements in artificial intelligence and other analytical tools have also become increasingly important in the claims process and are transforming how carriers do business.

Data Management and Modeling

Data is one of the most valuable assets an insurer can have, and predictive analytics have been helping businesses make the most of that data. From forecasting customer behavior to supporting underwriting processes, predictive analytics and data have been working together to provide valuable insights to insurers for years now.

However, making the most of your data is only possible with excellent data management and modeling capabilities. If data is scattered across disparate systems and there isn’t a strategic plan in place, all of that data is wasted. With data management solutions, predictive analytics tools can build a robust customer profile, provide cross-sell and upsell opportunities or even forecast potential customer profitability. And with insurance data models, insurers can deliver on-demand services to their customers via the cloud, using the data-driven insights gathered from their data management platforms.

Identifying Potential Markets

Predictive analytics in insurance can help insurers identify and target potential markets. Data can reveal behavior patterns and common demographics and characteristics, so insurers know where to target their marketing efforts.

Because there are 3.2 billion people on social media around the world, these platforms have become increasingly important when it comes to identifying potential markets. The platforms also influenced customer service: about 60% of Americans say that social media has made it easier to obtain answers and resolve problems.

Gain a 360-Degree View of Customers

TechTarget defines the 360-degree view of a customer as “the idea that companies can get a complete view of customers by aggregating data from the various touch points that a customer may use to contact a company to purchase products and receive service and support.”

Using predictive analytics, insurers can quickly and accurately consolidate data and generate insights that paint a more complete picture of a customer. What are their buying habits? What is their risk profile? How apt are they to buy new or expanded coverage? Before predictive analytics, insurers could estimate or take guesses at these questions, but now they are able to accurately and effectively service customers, which ultimately results in happier customers and increased revenue.

See also: How Analytics Can Tame ‘Social Inflation’

Providing a Personalized Experience

Many consumers value a customized experience – even when it comes to shopping for insurance. Predictive analytics in insurance provides the capability to comb through IoT-enabled data to understand the needs, desires and advice of their customers.

More and more insurers will use predictive analytics to help forecast events and gain actionable insights into all aspects of their businesses. Doing so provides a competitive advantage that saves time, money and resources, while helping carriers more effectively plan for a future defined by change. After all, data is only a strategic asset when you can actually put it to work.

3 Ways to Bar Fraud in Roadside Assistance

Unemployment is rising due to COVID-19, and some of the top risk management firms in the industry have indicated that fraud will also quickly increase. While claims fraud, inflated repair invoices and other common scams are probably the first that come to mind, roadside assistance fraud is another issue to which insurers should pay attention. It’s more common that one might think.

Especially as insurers increasingly offer ecosystem services such as roadside assistance to strengthen customer loyalty and generate additional revenue, it’s important that they ensure their roadside assistance partners are taking measures to protect against fraud, which can range from customers abusing the aid to get free gas, to tow operators sending fraudulent invoices. 

Here are some of the ways to protect against roadside assistance fraud:

Fast Payments Promote Trust

The adage, “An ounce of prevention is worth a pound of cure,” holds true in roadside assistance fraud. Instituting policies that reduce the incentive to commit fraud is far less painful than attempting to recover a loss, and one of these measures is to pay tow operators quickly. Especially in difficult times, those who issue payment within minutes instead of the standard Net-30 will foster loyalty that cuts down on fraud, especially while many companies are struggling to keep operations running. 

Tow operators balk at complex billing that deducts difficult-to-understand fees from their payment. The fees create unpleasant surprises that can make the difference between a profit and a loss. So, make sure that roadside assistance partners offer clear and transparent billing.

Transparency in Invoicing

With fraud on the rise, be on the lookout for roadside vendors that will attempt to bill you for “ghost” services. One way to mitigate this type of fraud is by asking your roadside vendor to provide transparency into completed services, ideally in real time. Require your roadside vendor to provide an unfiltered view into jobs, as well as customer confirmation that the job was completed. This kind of transparency makes it far less likely that you’ll be inaccurately billed or overcharged. Plus, this level of transparency gives you a better view into the customer experience. 

See also: 3 Ways AI, Telematics Revolutionize Claims  

Transparency in Operations

History tells us that, during times of high unemployment, we are likely to see more bad actors entering gig economy jobs. But fraud is often easily caught at the background check level, which can prevent bad actors from getting into the system in the first place. While it’s easy to provide a false name, it’s more difficult to provide a false Social Security number and matching drivers license. So, it’s important to have transparency in how tow operators are onboarded into your roadside assistance partner’s network and the methods they use to verify the identity of each driver and the person’s background. Ask about the types of checks they’re using to ensure identity verification, proper licensure and insurance compliance. Ideally, you want visibility all the way down to the driver level of who is servicing policyholders.

You need transparency because, while background checks have been an industry standard for years in roadside assistance, they may not be conducted at the appropriate level. For example, it’s common to accept a prior, third-party background check for a new contractor, a practice that leaves critical gaps in a contractor’s history and doesn’t necessarily report on charges or information relevant to the position. A “clear background check” usually does not tell you that the driver’s license is suspended, for instance. 

Background checks should be run annually at a minimum, but there are now next-generation background check services that will run in the background to provide live monitoring of arrest feeds, county reports and other proprietary information sites. This kind of continuous monitoring can flag events that could signal trouble, providing the opportunity to prevent fraud before it occurs.

The Importance of Analytics

Some policy holders may look to their roadside policy to help get what they see as “free fuel” as many times as possible. It’s a common scam, where drivers purposely avoid filling up and, when they run out of fuel, call the roadside assistance service to get some for free. 

This kind of fraud is most effectively detected through technology, specifically artificial intelligence, machine learning and analytics. Data analysis can identify previously overlooked trends to catch these kinds of issues and resolve them quickly. Insurers save money when machines and automation do the work instead of adding to headcount or finding problems only after the damage has been done. 

Even as fraud is anticipated to increase, roadside assistance many times has been overlooked. Don’t settle for passive fraud detection. Demand transparency and encourage the use of technology to mitigate risk, which will both reinforce your reputation and drive your bottom line.

What’s New in Fight Against Fraud?

Insurance fraud has been around since, well, the beginning of insurance.

The ancient Greeks created a form of maritime insurance to indemnify against potential losses incurred with the sinking of a commercial ship in transit. It became a common scheme for the boat owner to hide the boat in a foreign port and collect the insurance money. Even in those early times, special investigators were hired to determine if the boat had indeed sunk.

Fast-forward to the present, and, for the last few decades, the industry has been using increasingly sophisticated technology to address fraud. Now, several technologies can change the game for detection.

For example, machine learning, social media and aerial imagery can all contribute. All generate and rely on massive amounts of data, including many new data sources. Whether we are talking about opportunistic fraud or organized crime rings – these technology areas provide terrific opportunities to combat fraud.

Of course, fraud may occur during the underwriting OR the claims process. When a person or business is applying for insurance, there is always the potential to purposely supply incorrect information to get a lower rate. On the claims side, fraud may occur at many points during the lifecycle. In the case of staged accidents, it is occurring even before the accident occurs.

So how is the advance of technology aiding in fraud detection today? First, let’s look at new data sources.

Rate evasion can be more easily spotted today due to the wide variety of new data sources that can provide checks on the information provided by a customer or agent. For example, for auto, it is easier to spot true garaging locations or identify if a vehicle has been in a flood. For property, there is a wealth of data about the current characteristics of the property.

See also: Identifying Fraud in Workers’ Comp  

When it comes to machine learning, big data approaches with massive computing power and huge data sets can spot patterns and anomalies that it would be impossible for humans to spot – and do so with a lower rate of false positives. Social media has become a central tool for investigators and law enforcement, especially for workers’ comp fraud. We’ve all heard stories about individuals claiming disabling injuries then show up in Instagram pictures skiing or skydiving. The social media universe also yields a lot of information about connections between various individuals and businesses that can be mapped to identify fraud rings. Using aerial imagery, it becomes easier to compare before and after pictures of a property to determine if damage was caused by a particular weather event.

One of the biggest benefits of all this new capability is that technology allows fraud to be detected significantly earlier in a claim and with greater accuracy, so that Special Investigative Units (SIUs) and claims processes are more effective (compared with before, when SIUs or management found out about a fraud three to four weeks or longer after FNOL, by which point it was too late).

There is still much work to be done to find the right solution partners, integrate new solutions with existing systems and determine the optimum balance of technology and human expertise. But there is now greater potential to finally make significant headway in reducing fraud, especially the potential for earlier identification and more accurate outcomes. That’s what’s new and encouraging in this long-running battle!

Fighting Fraud With Data Analytics

The FBI reports that the total cost of insurance fraud is estimated to be more than $40 billion per year, costing the average U.S. family – in the form of increased premiums – between $400 and $700. A long-established and growing problem, insurance fraud has its many guises – ranging from tiny, one-off opportunistic cases to multimillion-dollar syndicates of customers and suppliers working together to routinely defraud insurers.

Luckily, digital enhancements within the insurance industry have been able to help companies lessen certain fraud risks – particularly when data analytics is brought into the mix.

To remedy insurance fraud using data analytics, individuals and businesses must be analyzed as they exist in the real world – as holistic, connected entities. To make these kinds of connections accurately, detection strategies must process high volumes of data in real time, be able to generate and constantly update a view of entities and apply a scoring model to the full picture. This allows companies to track and catch fraud, even across insurance lines and when multiple people are involved.

Fortunately, there are now technologies that are able to do just that – detect fraud and understand risk throughout a customer’s lifecycle. This will, in the long run, provide better claims processing and a healthier insurance system.

See also: Leveraging Data Science for Impact  

Quantexa, a data analytics firm that uses AI technology to piece together suspicious customer behavior, enables companies to make better decisions with their data. Their technology allows users to knit together vast and disparate data sets and derive actionable intelligence, a task that would normally take a human many months to complete. This technology can be focused on a single person and the many data points that are correlated to him or her, or larger entities such as corporations.

Technology like that of Quantexa’s can gather both claims and policies and build a network that provides three levels to which one can apply analysis:

The claim: This analyzes claim behavior over a long period. For instance, has a person filed for soft tissue damage multiple times? If so, how often and at what rate? This frequency could be a marker for fraud. There is also the ability to review if claims are filed close to when policies are taken out – another marker for fraud.

The entity: The entity can be either a claimant or, say, a medical provider; the analysis lies within the relationship between the two entities. Believe it or not, there are instances where medical providers have intentionally and habitually provided the wrong injury code; for example, if a claimant is in the hospital for an injured leg, the medical provider bills the insurance company for a more expensive procedure, such as a hysterectomy. Technology can detect and assess injury code discrepancies.

The network: This is based on the density of relationships and connections between claimants, witnesses, medical providers and beyond, and can stem from both claim information and transactional data. For instance, are multiple claims from “different” claimants all going to the same bank account? Factors can be pieced together to paint a larger picture on where fraud is originating.

See also: How Connected Data Can Help Stop Fraud  

Technology allows fraud to be detected much earlier on and across much larger schemes than humans ever could – a fact that should give thieves something to be concerned about, and all honest insurance policyholders something to rejoice about.