Auto Fraud

Fraud Management in Auto Insurance A Thought Paper Introduction F raud, in its basic form, would involve Approximately 90 percent of the costs of insurance fraud are the result of what is called “claims padding” – adding damage, injuries, and fictitious passengers to insurance claims. According to a survey by the Insurance Research Council, 35 percent of Americans think it’s all right to pad their insurance claims. The other 10 percent of insurance fraud comes from organized accident staging. Innocent victims (private motorists, truck drivers carrying chickens, etc.) are targeted by organized auto-accident rings. These rings make an “accident” happen by setting up an innocent person for a collision. Not only is this illegal, but tragedy can result. misrepresentation / suppression of material facts which would ‘alter’ the decision of an insurance company. Insurance fraud can be either “hard” or “soft”. Hard fraud occurs when someone deliberately fabricates claims or fakes an accident / event. Soft fraud (also called as opportunistic fraud) occurs when normally honest people inflate legitimate claim amounts (“padding”) for life / health insurance, or intentionally understate the number of miles they drive each year (in case of auto insurance), or list fewer employees or misrepresent the activities (in case of business owners / worker’s compensation) to get lower premium. Some lines of insurance are more vulnerable to fraud than others. Health care, worker’s compensation, and Most of the insurance companies have adapted to custom-built, rule based technologies to identify and prevent fraudulent claims from happening. However, lack of a comprehensive centralized database of all automobile insurance are more ‘fraud-prone’ as compared to other lines of insurance. Insurance Fraud can happen at any stage in the insurance transaction by different parties – the fraudulent activities in the industry and “intelligent appropriate technologies that are applicants for insurance, policyholders, third-party providers / claimants and / or any other enough” to prevent fraud is one major issue, which insurance companies need to address to curb the ever-increasing losses. professionals involved with the policy. Based on the intent and the information provided by the insured, an insurance application (for coverage/claims) may be classified either as fraudulent or genuine. If undetected, fraudulent information would result in huge losses. Insurance fraud costs the insurance industry an estimated $30 billion each year and the average American household approximately $300 each year in extra insurance premiums. This paper focuses on the incidence of fraud at two different processes in the auto insurance company namely, underwriting and claims settlement and suggest some probable techniques to mitigate losses (since elimination of fraud is almost impossible task!). 2 Fraud Incidence during Underwriting During the underwriting of an automobile insurance application, the prospect may resort to fraud to get a ‘favorable’ underwriting decision, either providing “false information” or by Systems designed by using user-defined rules and business logic based algorithms can define multiple scenarios to assist the underwriters arrive at an accurate decision by identifying the information inaccuracies and ‘flagging’ them as potentially fraudulent. An example of one such flag would be a simple business rule which notifies if the client has purchased multiple insurance from company ‘A’, but is approaching a different insurer company ‘B’ for auto insurance cover only Such a behavior suggests that the client has certain undesirable history (pertinent to automobile insurance) and is confident about Company ‘A’ NOT providing him the auto insurance cover. He would therefore approach a different insurer, Company ‘B’ and apply for automobile insurance without disclosing The illustration below shows that information required to monitor the fraudulent intent is available from both external and internal sources. While most of the information is available internally or from the insurance application, monitoring of fraud also involves interfacing with external agencies like the MIB or the MVR for obtaining certain sensitive information. An underwriter has to use his/her due diligence by applying analytical and judgmental capabilities before concluding on the decision to either accept or not accept the application for ‘Auto Insurance’ coverage based on the ‘face-value’ of information provided. the adverse information. In such a scenario, a simple business rule designed to capture and compare all the personal insurance details of the client should be able to alert the underwriter of company ‘B’. This rule would be designed to interact with a “centralized database” of the insurance industry and map the prospects “incomplete information”. Either way, the insurance company is at a disadvantage since both these acts by the prospect would result in improper risk assumption by the insurer. Typically the information that is incompletely provided includes (a) the residential address, (b) personal & family medical history, and (c) the occupational details. Information pertaining to (a) previous / concurrent insurance information, and (b) the vehicle details like state of registration, the VIN number, the mileage etc., may be incomplete. information to the available details. This database would be very similar to the Medical Information Bureau (MIB) database often used by the life insurance companies. Such rules are very basic, easy to understand and implement. Since there are a plethora of possible combinations for verifying various information, availability of the relevant data Insurance companies can, however, identify is very important. inappropriate information most of the times by using expert underwriting systems or technology based solutions that result in minimizing the losses. Predictive analytics based decision tree models are another technology that an underwriter can depend on to make precise and consistent decision about the prospects. These models are easily understood and are based on a complex set of business rules that produce a fraud score. This technology incorporates compiled information from multiple sources rather than relying on a simple red-flag system to provide an insight on future customer behavior. These models are created from predefined data elements where the ‘outcomes’ are known. Any new information can therefore be ‘run’ using this model to view the probable outcome of the decision being taken. Internal Sources Personal Information • Age / Gender • Residence Location • Work Location • SSN / Credit History Vehicle Information Applications Interface MVA/ Dealers Insurance Information • Database • Claims History • Other Company’s Cover • Sales Person Medical Information Applications Interface Physician / MIB External Sources Fraud Incidence during Claims Settlement The claims handling process of (automobile) insurance policies 1 compared to underwriting since it is the final stage of an insurance policy resulting in payment to the policyholders. Once an insurance claim payment is is more prone to fraud as There are basically six different types of coverages that are usually available under an auto insurance policy. Some may be required by law. Others are optional. These coverages are: (a) Bodily injury liability - for injuries the policyholder causes to someone else, (b) Medical payments or Personal Injury Protection (PIP) - for treatment of injuries to the driver and passengers of the policyholder’s car, (c) Property damage liability - for damage the policyholder causes to someone else’s property, (d) Collision - for damage to the policyholder’s car from a collision, (e) Comprehensive - for damage to the policyholder’s car not involving a collision with another car (including damage from fire, explosions, earthquakes, floods, riots, and theft) and (f) Uninsured motorists coverage - for costs resulting from 1 made, the claim becomes ‘legitimate’ and the money paid is lost ‘forever’. Research has shown that the magnitude and frequency of occurrence of an insurance fraud is greater at the claims stage as an accident involving a hit-and-run driver or a driver who does not have insurance 4 compared to the time of applying for a policy. Fraud may be committed at different points in the insurance claims transaction and different parties may be involved – the insured, third-party claimants and professionals who provide services to the claimants. Common frauds include: • • • • “Padding” or inflating the actual claims Misrepresenting of facts on the claim application Submitting claims for injuries or damages that never occurred “Staging” accidents 2 • Injuries (payable under PIP) claimed to have happened as a result of a certain accident are not in tune with the severity of the accident • • • Considerable delay between date of the accident and the date of filing the claim Diagnostic tests suggested by the physician does not justify the intensity of the accident In case of claims arising from a specific area, reference is made repeatedly, to a specific physician, or a specific garage located in that area • Repeat claims from a specific / group of individuals at regular intervals Claims’ department is typically the first line of defense an insurance company has, to combat fraud. Manual detection processes used by the claims departments in most of the companies are ineffective and most of the fraud slips-through undetected. Even the best claims adjuster cannot perform the detailed analysis needed to find the complex patterns that indicate fraud. It is also estimated that industry wide, claims adjusters handle upto 250 claims at any given time. These loads only add to compounding the problem further. Based on the claims handling experiences of insurance industry, some indicators to suggest a potential fraudulent intention have been identified. • Loss Value does not justify the location, time and nature of the accident involved Knowledge of subject, experience and attention to detail are essential in detection of fraudulent claims. While human experts are capable of identifying some of the red flags and simple fraud patterns, sophisticated modeling techniques are required to find more complex patterns of fraud. Advancements made in technologies like data mining and neural networks can help in the accelerated decisionmaking abilities of systems and empower the insurers to do better risk assessment. Extensive Data modeling, data warehousing, and data mining capabilities are essential for establishing trends and performing analysis. However, all these techniques are data intensive and the availability of client relevant information is very critical for their success. 2 Staged accidents usually follow some basic schemes like (a) Stopping suddenly for no apparent reason, (b) Disregarding the right-of-way intentionally, or giving up the right-of-way on purpose, (c) Reporting passengers who were not in the vehicle at the time of the accident, and/or witnesses that were not actually on the scene, (d) Claiming excessive bodily injury not commensurate with vehicle damage, (e) Claiming property that was not in the car at the time of loss, (f) Falsely reporting a car as stolen, (g) Fake hit-and-run claims, and (h) Destroying a vehicle by setting it on fire 5 Use of Technology in Handling Insurance Fraud Many of the actual and potential indicators of fraud and misconduct reside within the insurers financial, operational, and transactional data and can be identified using appropriate data analysis tools and techniques. Fraud detection systems operate by consistently and dependably relying on diverse mathematical techniques like probability, statistics, and artificial intelligence. The propensity of an applicant to involve in fraud at a later date may be ‘guesstimated’ using data analytics techniques. An effective solution to insurance fraud and abuse relies on a comprehensive approach and not limited to a single technology. These techniques can vary from a simple business rule based “red-flag” triggering system to very complex pro-active data analysis systems that use sophisticated analytical tests, computer-based cross matching, and nonobvious relationship identification techniques to highlight potential fraud and misconduct unnoticed by management for years. Among other things these techniques provide for: • Identification of hidden relationships between people, organizations, and events to analyze suspicious transactions • • The potential to continually monitor fraud threats and vulnerabilities. The ability to consider and analyze thousands of transactions in less time, more efficiently, and cost-effectively than using more traditional forensic sampling techniques Some of the emerging technologies that can be considered by the insurance industry to handle the claims fraud include • • • Transaction Based Profiling – The ‘creditcard’ vendors, for tracking the spending behavior of their customers, very effectively employ this technology. This would involve collecting, analyzing and transforming client data into a set of features to describe the behavior of insured. The insurance history details of an insurance policy holder like the number and type of policies held, the number of claims made etc., when mapped to the financial profile of the client may present a clear picture regarding the intent of the insured. Access to a centralized data repository having the ability to provide the said information and containing the transactional history of (all) the insureds is essential for ‘tracking’ the profile of the insured. Neural Networks – Neural Networking Models replicate the functionality of human brains and help in stimulating the situational behaviors of the insureds, based on certain preset “stress factors” or claim parameters. The output is usually in the form of scores ranging from 0 to 1000 where a higher score indicates a greater likelihood of the malafide intent. These systems are particularly suitable for claims processing since the systems can adjust according to the capacity of the investigative units. Scoring methods employed by these technologies help in identifying the claims with highest fraud propensity. Data Mining technologies – This technology employs the mining (searching) of all the data sources for insurance information and correlating the results with known fraudulent claims to indicate relationships that can be used to monitor future claims or enrollments. The volume of data that needs to be searched depends on ‘predefined’ set of parameters and the reason for conducting the search. . As in the case of transaction based profiling, access to overwhelming amounts of claims data is required to identify the required information. A simplified model based on the above three techniques and useful for identification of potential frauds is presented below Input 1 Logic Rule 1 Input 2 Logic Rule 2 Input 2 Debit 3 Logic Rule n Debit 1 Debit 2 = Fraud Score Risk Identification Input n Business / Mathematical Logic Data Input User data inputs are captured and used by the business/mathematical logic layer. Each node in the mathematical layer has embedded business logic and indicates a unique combination of the inputs and the expected output. This logic could be anything like a simple ‘if…then…else…’ type of business rule or a complex regression model. Suppose, the input ‘A’ is the residence location and input ‘B’ is the work location, and a business logic pertaining to distance (parameter) would probably state that if the distance between residence and work is more than a given certain value (say 25 kms) the probability of the accident is 30%. Results obtained from this activity are then compared to the (a) analytical data emanating from claims experience of insurance industry, (b) the % of claims involving the said distance & locations (c) the policies sold vs. the claims experienced and (d) the risk appetite of the insurer. This process results in marking the ‘intelligent’ leads required to identify the possibility of a claim. A multiple regression algorithm written (instead of the rules) taking the said factors into consideration would then be used to arrive at a variable value (probable claim value). This number indicates the probability of an event happening. This approach is similar to the ‘exception reporting’ mechanism employed in the information technology processes. Multiple values can thus be generated using different parametric values (inputs) and stored in a database, to 7 establish a trend. Based on the risk appetite of the insurance a unique debit value can be assigned to each of the results. By running multiple stimulations of this model through large sets of data, a comprehensive database of the debit values can be built and a trend of input vs. output can also be generated. This database would form the basis for establishing the fraudulent intent (if any) on the basis of the information supplied by the insured. • On receipt of a fresh application, the values that are input are run through the model to calculate the debit value and compared to the data in the database to establish the “probability” of the case ending up as a claim. If the debit values fall under the category of suspected claims, the insurance company can decline cover and avoid a potential claim. • With little modification, the system can also be used to ascertain the “genuine-ness” of a claim also. Most of the claim related information is readily available with the insurance company and only certain parameters like place of accident and the parties involved, vehicle related information etc would differ from case-tocase. Stimulations can be run for such data and a database of the fraud values may be generated situations. The above model analytics, incorporates neural features of and for comparing with ‘real-time’ predictive networks regression-based models. Since there is minimal human intervention (except at data entry stage and execution of the application), results are expected to be accurate. However, elaborate programming and processing capabilities are required along with large amounts of data, for the design and development of the system. 8 Conclusion Insurance fraud is multifaceted. There is no single typical profile of the fraudulent claimant; Unfortunately, the absence of any centralized data repository (like the MIB) is only encouraging the fraudsters to design the ingenious ways of professional criminals as well as an ordinary citizen can commit fraud. The strategy to combat fraud must not only be reliable and effective in its prevention but also be capable of detecting fraudulent activities as and when they occur. A misrepresenting information from various insurance companies. Fighting fraud requires a commitment to process review, workflow automation, and auditing. Insurers must also maintain a constant vigil for new types of fraud and the evolution of existing fraud schemes A good strategy to combat fraud involves effective risk identification, analysis, and reporting activities and it requires the use of a combination of data validation and mining capabilities, visualization techniques and reporting tools to identify collaborative effort comprising of the insurance industry players, the regulators and the ‘policing’ authorities and all the relevant parties (attorneys, physicians, garage owners) may help in designing intelligent systems as a part of the risk mitigation strategies against losses that happen due to fraudulent claims. The performance of these systems depends on the availability of client data, which can be shared between various insurers. A proactive strategy enables insurers to remain vigilant against fraudsters protecting themselves against both small and large-scale claims frauds attacks. questionable behavior before a claim is paid. Replacing the traditional manual processes that involves sorting though large amounts providers data and claims information is not effective in establishing potentially fraudulent, wasteful, abusive or questionable behavior. About NIIT Technologies NIIT Technologies is a global IT Solutions major servicing customers in North America, Europe, Asia and Australia. The company offers services in Application Development and Maintenance, Enterprise Solutions including Managed Services and Business Process Management to enterprises in the Insurance, Banking & Financial Services, Travel &Transportation, Manufacturing and Retail sectors. NIIT Technologies’ subsidiaries NIIT SmartServe Limited and NIIT GIS Limited offer Business Process Management and GIS services respectively. 9