Cracking the Code

As fraudsters are becoming more sophisticated, insurance providers around the globe are looking for new ways to step up efforts to protect customers and decrease fraud losses. The conventional rules-based approach is easily implemented and adjusted by insurers, but it is also limited because it cannot proactively search for new areas of fraud.

It is, in fact, limited to the rules.

Predictive analytics are becoming more prevalent because they can uncover more fraud and create sophisticated defence methods against unknown schemes that rules do not catch. The optimal fraud-detection approach combines elements of both analytics and rules.

INSURERS ANTICIPATE UPHILL BATTLE

It is widely known to insurers that the industry is losing substantial sums to fraud. Recent statistics from the Coalition Against Insurance Fraud estimate that US$80 billion in fraudulent claims are made annually in the United States. And Canada is no exception.

The Insurance Bureau of Canada has repeatedly estimated insurance crimes cost Canadians several billions of dollars in insurance premiums – with 10% to 15% of insurance premiums being used to cover the cost of fraud.

Consumer awareness of this issue has increased as well, with a Pollara poll issued in March 2012 finding that more than 80% of respondents in Ontario believe insurance fraud is a frequent and regular occurrence.

Research released in September by FICO, a global provider of predictive analytics and decision-management technology, found that one in three North American insurers polled does not feel adequately protected against fraud. Focusing on insurance claims fraud, the survey revealed that insurers feel most vulnerable in the areas of premium leakage and new applications, when policyholders often underestimate or leave out such information as annual auto mileage that would have an adverse effect on the cost of the policy.

Respondents also said they expect the biggest fraud loss increases to hit personal property, workers’ compensation and auto insurance. In terms of fraud by individual policyholders, 58% of surveyed insurers forecast an increase in personal property fraud, 69% forecast an increase in workers’ compensation fraud, and 56% forecast a rise in personal auto fraud.

Geographically in Canada, 42% of those surveyed foresee Quebec as being the hardest hit by personal lines fraud, and 39% anticipate Ontario being the hardest hit.

USING ANALYTICS TO BEEF UP BUSINESS RULES

Across all segments, and using various tools, the insurance industry is constantly working to stay ahead of fraudsters. Many companies today are still heavily reliant on business rules-based management systems to fight fraud. These systems test and measure each claim against a predefined set of business rules and examine the claims that look suspicious as a result of their combined scores.

Although traditional rules-based solutions can be effective and relatively easy to implement, they only tackle the tip of the iceberg. Over time, fraudsters can readjust their methods and manipulate the business rules in their favour.

As well, to date, much of the effort to combat fraud has focused on the laborious process of trying to recover money from false claims by investigating suspicious claims after they have been paid. On average, this pay-and-chase approach takes one to two years – and in some cases much longer to recoup payments.

A more effective and efficient method is to identify false claims before they are paid. To do this, leading companies are now supplementing their business rules with sophisticated predictive analytics and link analysis to detect more kinds of fraud, waste and abuse.

Whether in auto, homeowner, property and casualty, health or any other category, predictive analytics make rules-based approaches more dynamic and effective. Analytics can uncover more fraud and create sophisticated defence methods against unknown schemes that rules do not identify – and are among the most transformative methods of increasing speed to detection and improved decision-making.

They can help fight fraud by distinguishing patterns in claims that may point to fraudulent activity, as well as by understanding payers’ transactional and relationship data to reveal wider instances of fraud.

In addition, analytics enable the efficient detection of more types of fraud than any other method. The power of an analytic model is that it takes multiple features into account simultaneously, and uses them to identify cases with a very high likelihood of fraud.

For example, an unusually high number of fraudulent accidents or incidents occur very close to the date of inception or expiration of the policy. However, not every claim filed near the inception or expiration date is fraud.

Taking these features into account in an analytic model in combination with multiple dimensions simultaneously can provide an accurate indication of the likelihood of fraud.

Additionally, advanced analytic techniques such as link analysis detect suspicious patterns based on previous claims data. It takes data and considers the larger picture by examining relationships among organizations, people and transactions, or among providers, members and claims. In combatting insurance fraud, link analysis works by flushing out the associated claims or providers who may not always appear to be related in an obvious manner. 

WHERE FRAUD DETECTION DELIVERS

There are three key points in the insurance process where a focused assault on fraud using a combination of business rules and analytics can deliver substantial results. These appear below:

• Point of sale (POS) – Avoid doing business with fraudsters and reduce premium leakage by more accurately classifying and pricing risk. Business rules verify that applications are complete and correct, accessing internal and external data sources as necessary. Analytics detect suspicious behaviour patterns and score applications for risk of premium leakage or fraud. Based on scoring thresholds, rules then determine what happens next (e.g., accept, reject or refer to a specific analyst queue).

• First notice of loss (FNOL) through claims adjudication – Avoid paying fraudulent claims, and identify fraudsters and fraud rings. Business rules guide claims representatives and automate online claims-filing applications. Rules-powered intelligent forms ask only for the information required based on the claim and coverage. They access data sources as necessary, invoke analytic models to score the claim for risk of fraud, then recommend or automate actions based on the fraud score. During the adjustment process, rules can monitor claims for suspicious new information and revoke models to re-score claims for fraud risk.

• After payment – Find more fraud while gaining deeper behavioural insights to improve POS and FNOL detection. Analytics examine large volumes of data to detect patterns of fraudulent behaviour not evident in smaller data sets, and discover the complex and often subtle connections pointing to organized fraud rings.

Insurers can benefit greatly by injecting into their rules-driven processes the appropriate type of analytics to address fraud risk at POS. They could subsequently add analytics for fraud detection at FNOL and throughout claims adjudication, and later strengthen their after-payment fraud detection.

GREATER PRECISION EARLIER

To reduce fraud losses, insurers must detect suspicious activity with greater precision at the earliest possible moment. Achieving this real-time intelligence on potentially fraudulent claims via sophisticated combinations of rules and analytics is the next frontier of significant bottom-line improvement in the insurance industry.

How much value analytics can deliver depends on how effectively they are used. The challenge for insurers is to know at what point they are getting the desired return on investment – how much f raud is being uncovered, how much loss avoided and how much time saved.

Establishing the right strategy to reduce fraud stands to dramatically improve the customer experience and increase competitive advantage by potentially enabling insurance providers to lower their costs of doing business and offer more competitive premiums.