Explorations into Insurance Fraud - Research Paper Example

Explorations into Insurance Fraud EXPLORATIONS INTO INSURANCE FRAUD General Discussion Questions Introduction Although it has beenestimated that insurance fraud costs the UK economy £2.1 billion every year, academic research into how claims investigators make decisions and the profiles of fraudsters, is limited. Indeed, there has been no research attempting to characterize prevalent properties of fraudulent claims or even to compare these with fraud specialists’ judgments in the insurance industry. In addition, it has been challenging to come up with reliable indicators of fraudulent claims activity due to a dearth of empirical evidence covering effectiveness of techniques to intervene in insurance fraud. This is despite the UK insurance industry being the 3rd largest in Europe, as well as forming a critical pillar of the UK economy. Moreover, significant amounts of the industry’s net revenues are sourced from foreign markets. Insurance fraud is referent to various illegal activities that enable individuals to acquire material gain from insurance companies through dishonest methods and can be divided into hard and soft fraud. Whereas hard fraud refers to attempts to fake theft, personal injury, accidents, and arson so as to make insurance claims as stipulated under the insurance policy, soft fraud refers to overstatement of genuine claims by exaggerating one’s loss for material or financial gain. Fraud as an offence in the UK is covered under the Fraud Act of 2006. To counter fraud, insurance companies, have come together to assess fraudulent activity impacts through intelligence bodies, such as the IFB and CIFAS, while the London Police has the IFED that is funded by insurers. The current research seeks to: examine the effectiveness of existing methodologies for fraud detection interventions; establish the different types of fraud events being dealt with by insurance fraud investigators; conceptualise specific typologies applied by investigators; and Establish whether a framework of behavioural characteristics and themes representative of insurance fraudsters can be born of the fraud data. General Discussion Effectiveness of Existing Methodologies for Fraud Detection Interventions The research begins by conducting a literature review to examine the effectiveness of fraud detection methodologies currently in use to intervene in insurance fraud. It contends that the insurance industry over the recent past has become more concerned about how to detect fraudulent behaviour and prevent it, which has not proven an easy task. In order to be more successful, they have employed various interventions aimed at controlling fraud. One of these methods is prevention, which involves the reduction of risks related to misconduct and fraud occurring. Another method is detection that involves the discovery of misconduct and fraud once it occurs. Finally, the third method is response, which involves the taking of corrective actions to remedy the fallout after misconduct and fraud have occurred. In addition, they have also turned to psychologists, who have contended that fraudsters normally change their story with time, which is what the detection systems must seek to identify. Insurance fraud interventions can also involve auditing techniques, which assumes that a process of audit can discern whether claims are fraudulent or honest. Data matching techniques can also be used to detect insurance fraud through data matching techniques, which involves the combination of data from various sources and their comparison. To finish, data mining techniques can also be used; this involves the use of computer software analytics to detect insurance fraud. These techniques can be used to reduce losses by detecting and preventing fraud, as well as reducing compensation time for legitimate claims. However, because of the dynamic nature of fraud and a lack of resources, the research contends that these detection systems could prove unreliable and misleading, which necessitates the selection of the most effective solutions due to the high investments required. This chapter sought to conduct a systematic review methodology in identifying intervention descriptions to assess evidence that supports the effectiveness of interventions used in combating insurance fraud. The literature search initially yielded 645 studies that were whittled down to 15 after some were excluded on the basis of duplicity, titles, unavailability of full titles, abstracts, and finally the full text. This researcher identified that there existed a wide and extensive literature gap with regards to the effectiveness interventions aimed at combating domestic insurance fraud. This was explained by the possible difficulties in detecting fraud, as well as the low attention paid to fraud by academics and scholars. Different Types of Fraud Events Fraud Investigators Deal With The research then turned to establishing different forms of fraud that insurance fraud investigators have to deal with, as well as the characteristics of the fraudsters who present these fraudulent claims. It begins by studying the characteristics of fraudsters in motor and household insurance claims. It found that the mean age of those committing fraud in the household sample was 41 years of age with the most common range between 34-43 years. In the motor fraud sample, the most frequent offenders were aged between 28 and 29, while the mean age of fraudsters was pegged at 37 years within a range of 19-65 years. With regards to marital status, the household sample showed that 60.3% of fraudsters were married. In the motor fraud sample, 41.2% of the claimants reported they were single, compared to 39.2% who reported being married. The geographic locations with the most number of fraudulent household claims were in West Yorkshire and West Midlands. The entire Yorkshire collective accounted for 18% of fraudulent claims. With regards to motor fraudulent claims, most claimants came from London and Lancashire. For the household claimants providing contact numbers, most provided home numbers compared to mobile numbers or both cell and home numbers, while most motor insurance claimants also provided mobile phone numbers. The importance of contact details in this research was because most fraudsters have been claimed to only provide mobile numbers, especially due to their disposable and interchangeable nature. Turning to previous claims history, most claimants in the household sample were found to have made prior claims. While only 25 of the 78 sampled participants admitted to having made previous claims, retrieval of data from the Claims Underwriting database showed that 76 of them had made prior claims. On cross-tabulating previous household claims and disclosure of the same by policy holders, 20 of them admitted to a single claim, while 4 and 1 admitted to 2 and 3 previous claims, although the latter participant was found to have made 5 claims. 9 cases also involved claimants admitting to having made prior claims despite a lack of claims history. This makes the reliability of most insurance industry tools questionable with regards to soundness and validity. Investigations into the cause of loss for household claimants found that most of the claims were as a result of knock/drop and spillage. However, claims that could be verified using forensics analysis were lower, especially power failure, fire, and water escape. For household items that were most likely to be subject to claims, audio-visual items made up the largest percentage followed by computing equipment. Turing to the documentation used for these claims, engineer reports, were found to be used in only 7.8% of the cases, although it is difficult to discern whether these documents were requested to begin with, which may explain the low rate of presentation. Other documents, such as household receipts, household inventory, and household photographs were not received in most of the claims made. Coming to household reserve data, it was found that costs ranged from £32 to £7,111. With regards to motor insurance claims, the most prevalent cause of loss was theft followed by road traffic accidents, personal injury alongside collision, and malicious damage. The vehicle that was most prevalent in motor insurance claims was the Vauxhall followed by Ford, Nissan, BMW, and Toyota. This is despite most of these car manufacturers being members of the ISR that seeks to deter theft, cloning, and theft, which makes them less likely to be stolen due to registration and marking of the cars. It is important to note, however, that Vauxhall is not part of the ISR. For this car manufacturer, the Astra was the model most reported to have been stolen, followed by Vectra and Cavalier. Ford Escort and Ford Fiesta were the most vulnerable Ford models, while Nissan Micra, Almera, and Skyline are the most vulnerable for the Nissan model. Finally, for BMW, the 3 Series and the 7 Series were most vulnerable. Regarding the recovery of these vehicles, it was found that, 23.8% of the claims were recovered, compared to 49.2% of cars that were never recovered. Ford and Vauxhall models make up the biggest percentage of cars that were never recovered, which means that the ISR that covers Ford models is not as effective as it should be. Again, however, it is important to take into consideration that this sample is made up of fraudulent claims, which means that the policy holder’s knowledge about the theft is questionable at best. For the condition of the cars that were recovered, it was shown that only 3 of the 15 cars reported stolen were recovered with limited or no damage, which supports the contention that fraudulent claimants tend to hide cars in conditions that allow them to resell the cars later. From this sample, one may conclude that fraudsters prefer to completely damage or burn their cars, which is essential in getting rid of forensic evidence that could prove fraudulent activity. Specific Typologies Applied By Investigators Now turning to consider the perceptions that insurance experts have about fraudsters, the researcher collected beliefs and perceptions of industry experts towards the traits of insurance fraudsters and compared them to realities of fraud and its perpetration. When considering the gender of insurance fraudsters in the motor and household insurance sub-sectors, 63% of motor insurance experts contended that 50-75% of fraudsters were likely to be male, while 20% of them contended that 75-100% of them were male. This compares positively to the results in the earlier section that held most motor insurance fraudsters to be male. For the gender of household insurance fraudsters, 54% believed that the fraudsters would most likely be male in 50-75% of the cases. This correlated positively with the results from the earlier section that research that concluded males were more likely to be involved in household fraud. When asked about the percentage of motor insurance fraudsters they believed to be married, most of the experts reported they expected most of the fraudsters to be married. With regards to household insurance fraud, experts reported that they expected most of the fraudsters to be married. These results correlate positively with results from the earlier section, which held that most motor and household insurance fraudsters were married. However, almost half of the participants also expected motor insurance fraudsters to be co-habiting, while a similar number expected household insurance fraudsters to be co-habiting. However, this perception is much higher than that found in the previous section for both motor and household insurance fraud. When asked whether they expected fraudsters to be single, 46% motor insurance fraud experts reported that 25-50% of motor fraudsters were single, while 44% of household insurance fraud experts believed that 0-25% of fraudsters would be single. The difference in results for motor and household insurance fraud could be attributed to gender stereotypes. However, the results were consistent with those from the previous section. When asked to assign a percentage to ages they expected most fraudsters to fit into, most motor insurance fraud experts contended that they expected them to be between 27 and 31 years of age, while household insurance fraud experts expected fraudsters to be between 32 and 36 years of age. However, compared to data from the previous section about the age of fraudsters, both groups of experts made wrong assessments by suggesting the fraudsters were younger than they actually were. When questioned about the geographic location they expected the fraudsters to come from, motor insurance fraud experts answered that they expected the most fraudsters to live in London followed by the West Midlands. Household insurance fraud experts reported that they expected the most fraudsters to live in London, followed by the North West. For the motor insurance experts, the responses correlated positively with data from the previous section that held that most fraudsters were from London. However, household insurance experts were incorrect since previous data shows that most fraudsters would in the West Midlands and West Yorkshire. When requested to report on the number of claims a fraudster would most likely have made in the past, most motor insurance experts said that they would have made 1 or 2 claims in the past. Most household insurance experts contended that fraudsters were most likely to have made 3 or 4 claims in the past. Compared to results from the earlier section, it is seen that motor insurance experts misconceived the number of claims, as had household insurance experts. Majority of motor experts also expect fraudsters not to disclose any prior claims they had made when asked, which is similar to most household insurance experts who also expected the same. When questioned about the most common reason for loss in fraudulent claims, motor insurance experts felt that accidents and theft were the most likely. Household insurance experts felt that spills and knock/drop were the most common. Finally, when questioned about their basis for these perceptions, motor insurance experts based their perceptions on work experience, while household insurance experts based their perceptions on the same. Framework of Behavioural Characteristics and Themes Representative of Insurance Fraudsters from Fraud Data Finally, the researcher also sought to find out whether a framework of behavioural themes and characteristics of fraudsters could be extracted from data on fraud in order to enhance early “at risk” claims identification. The study found that 65% of all fraudulent claims had core behaviours, which are considered central to making fraudulent claims. The most prevalent behaviour was cooperative that was present in 80% of all claims, which was important because being disagreeable or obstructive would bring more focus on the fraudster’s claims. 69% of all fraudulent claims were also found to possess time gaps, whereby the fraudster was unwilling to provide the entire cycle of events related to their claims. The third core behaviour was chronological structured variable in 67% of claims, in which chronologically structured accounts indicate fraud since it is impossible to narrate real life stories in a perfect chronological manner all the time due to the way the brain orders information retrieval. The second level fraud detection involves corroboratable facts, which are variables that relate to the informational theme that can either be corroborated or disproved. In 63% of all fraudulent claims, the item present variable was found to be present, whereby the claimant argued that the damaged item was no longer available for forensic analysis. Previous claims form another variable under this level of fraud detection present in 55% of all fraudulent claims, which can be found in the CUE database discussed above. The notion that a claimant has made a claim prior to the present one could predict fraud, although the research does not conclude whether prior claims were fraudulent of genuine. Another variable is the provision of a description that is rich and specific in detail about a claimed item. This variable exists in at least 55% of all fraudulent claims since fraudsters provide fewer details about their claim than genuine claimants do, which makes their claims less complete and their talk time shorter. Another theme established by the researcher to detect fraudulent claims is narrative detail that involves variables specific to a claimant’s account or story. Admit previous claims predominates in 44% of all fraudulent claims with these claimants believing that withholding of previous claims information could act to benefit their fraudulent claims. Another variable under this level of detection is spontaneous correction, which, although only present in 38% of fraudulent claims, could indicate fraudulent claimant accounts of the loss. Claims assessed under this variable, however, could contain some elements of genuine claims. Finally, performing of personal investigations by the claimant is another variable that is present in 26% of all fraudulent claims. In most fraudulent claims, the fraudster does not suggest any attempts to limit losses in any way, meaning that this variable could be a useful detector of fraudulent claims, acting as a differentiator, rather than a clear pointer of fraud. Finally, claim specific details are the fourth level of fraud detection, occurring in 35% of all fraudulent claims. However, these behaviours and themes are not as clear-cut as the rest and are further from the core. Since 27% of all claimants contend that they can provide photographs of their claims, it can deduced that 77% of all fraudulent insurance claims do not possess any photographic evidence to prove their claims. In addition, it is also expected that claimants acting on insurance fraud present their claims shortly after purchasing an insurance policy in 9% of the cases. This is not consistent with the belief that fraudulent claims are submitted three months after the policy inception, making this variable an unreliable predictor in detecting fraud. In addition, where the insurer does not suspect fraudulent activity, this variable is limited in differentiating genuine and potentially fraudulent claims. Read More