Scale of insurance fraud in Europe
Global insurance fraud (undetected and detected) could amount to 10% of all claims expenditures in Europe according to a 2013 report by Insurance Europe, the European association of national insurance federations.
A report by the Association of British Insurers in 2012 explains that on average 15 claims an hour in the UK are exposed to be of a fraudulent nature. The British industry in 2011 detected ?1bn worth of fraudulent claims; it was also estimated that undetected fraud could represent as much as ?2bn additional that year.
In France, the association against insurance fraud (Association de Lutte contre la Fraude ? lAssurance) estimated that in 2015 insurance fraud amounted to 2.5bn, the equivalent of 5% of premiums. On this amount, 265m could be claimed back by insurers (21% more than in 2014, considering the number of detected fraudulent claims increased by 8% in a year, amounting to 46255 cases).
In Germany, the Gesamtverband der Deutschen Versicherungswirtschaft (GVD) estimated that fraudulent claims amount to 4bn.
Public attitude regarding insurance fraud
Insurance fraud represents a real loss for insurers, but also has a considerable impact on consumers, as the cost of detection and of expected loss is passed on within premiums. Nevertheless, insurance fraud is not considered a particularly vile crime, and may even be considered fair game by a large portion of consumers. In its 2013 edition of Insurance Fraud: A Public View, the Insurance Research Council found that 24% of Americans considered that exaggerating a claim to make up for deductibles is an acceptable behavior, while 18% considered that exaggerating a claim in order to make up for years of paid premiums without claims on their part is acceptable. The study also showed that young men are more tolerant when it comes to insurance fraud than their female counterparts (23% against 8% considering padding claims is acceptable). While tolerance of insurance fraud seems to be declining (compared to 33% in 2002), the percentage of Americans tolerating that kind of behavior remains paradoxically high compared to the 86% of Americans agreeing with the fact that insurance fraud leads to higher premiums for everyone.
Considering those numbers and assuming they are likely to be representative of a general attitude towards insurance fraud as well in Europe, several challenges for the insurance industry can be identified. Behind numbers related to the scale of fraud, it is interesting to have a quick look at the crime itself.
Types of insurance fraud
Generally, insurance fraud can be divided in 3 categories, related to the perpetrator (insider or outsider), the stage of the fraud (underwriting or claim) and the nature of the fraud (soft or hard). For the purpose of this paper, we will focus on insurance fraud perpetrated by applicants, policy-holders, claimants (insured) or third-party stakeholders. This paper aims to analyze the correlation between insurance fraud and the public perception of insurance companies as well as drawing conclusions on the ways insurers can combat fraud; in this respect, insiders fraud has little to do with this perception and should therefore be excluded from this overview. It is however interesting to note that insiders fraud is by far the most prevalent type of fraud; according to the FBI, premium diversion is the most common type of insurance fraud in the US. A study by EY shows that the risk exposure of insurance companies to premium-related fraud is estimated at 21.2%. In addition to that, employee-related fraud is estimated at 20.5% and vendor-related at 13.6%. Overall, insurance companies are more likely to be deceived by their own employees or business insiders than by their own customers. The same study evaluates claims-related fraud to 27.3% and application-related fraud to 17.4%.
The stage at which the fraud occurs is of some relative importance. While frauds related to underwriting underline intent on the part of the customer, who may conceal information for instance an already existing policy, or deliberately hiding facts which may have increased the cost of the premium, fraudulent claims are more likely to be opportunistic in nature. Two different mechanisms are at play here; they however both relate to the perception individuals have of the insurance sector.
When it comes to premeditated fraudulent acts related to underwriting, such as the concealment of information or misinformation, motivation to fraud appears to be straight forward: the end goal remains the perspective of misleading the underwriter in order to obtain a better deal on an insurance policy. However, this economic reality does not seem to be the only factor contributing to fraud. The image of the insurance industry in general as well as the products it sells seem to play a role.
Motivation to commit fraud
Beyond the fact that an insurance product is by nature intangible, and that a policy holder may knowingly be buying something he or she will never need, the perception of the industry as a whole is a negative factor which may influence cheating. Tennyson published a series of articles relating to insurance fraud and the factors affecting its occurrence. In 1997, the results of her research led her to conclude that a negative perception of the industry had an influence on the tendency to cheat on insurance contracts. Insurers are regarded by a significant proportion of their clients as making undue profits by selling an intangible product with no immediate return on investment and making the settlement of claims a difficult and painful process for claimants who by definition have already lost something. The general public does not have a good image of insurance companies, and the scandals which came with the financial crisis did nothing to improve it.
Other studies show that a previous negative experience with a claim significantly increases the propensity to fraud. A survey conducted by Accenture in 2010 in the US and found that 55% of interrogated people believe that poor service on the part of the insurance company makes a customer more likely to commit fraud. If the phrasing does not explicitly state it, it implies a majority of Americans would consider fraud relatively justifiable if the insurance company has provided poor service to its customers.
These findings indicate a correlation between the perception of the insurance industry and of its products and the motivation to commit fraud. If the achievement of an overall better image might seem a complicated goal to reach within a reasonable time-frame for insurers, improving their services and their relation with clientele could potentially deter fraud at least when it comes to opportunistic fraud by customers.
Prevention of insurance fraud:
Fraud occurs when opportunity and motivation meet. Fraud prevention should therefore focus on reducing opportunity and motivation. If working to reduce the window of opportunity seems rather vain, insurers may find ways to hinder motivation to fraud, by adapting and changing their offer and improving the service they provide.
When it comes to motivation, but also to some extent to the opportunity to commit fraud, the type of contract used seems to be of relevance. A study by Gabaldon, Vazquez-Hernandez and Watt of 2009 shows that the type of contract used has an impact on fraudulent activities. Through the conduct of experiments, the study concludes that if there is no significant difference in terms of fraud when a pure audit or a bonus-malus contract is used, the experience of a bonus-malus contract has a tempering effect upon the amount of fraud committed in contracts with audits and a fixed premium. While this does not mean that bonus-malus type of contract is less likely to induce fraud, the experience of such contract has a positive effect.
Regarding contracts with deductibles, Miyazaki in 2009 demonstrates that for higher deductibles, individuals consider claim build-up to be more justified. A study by Dionne and Gagne in 2001 shows that the level of the deductible is a factor of high significance for the submission of the claim. Finally, von Bieberstein and Schiller demonstrated in 2014 that in comparison to a bonus-malus contract, individuals commit more fraud when given a deductible contract.
Considering the above results, it is clear that the type of contract has an impact on fraud. In this respect, those findings should push insurers to revise their offers when it comes to contract format. While it may not have an immediate deterrent effect, reshaping contract offer could prove efficient in the long-term.
To increase the potential positive effect of contract reformatting, insurers could also take a deeper look at what their customers expect, both in terms of initial offer and general service. A survey conducted in 2009 by IFOP in France showed that 78% of questioned individuals considered that insurers made little or no effort to adapt their offers to customers needs. This striking number reflects the general opinion that people get insurance because it is required, sometimes legally required, and that this requirement does not push insurers to innovate and propose adapted offers to their needs.
Obviously, one of the prevalent factors inducing fraud remains the cost of insurance. Thinking you are overpaying for a service you do not value (for reasons explained in the previous paragraphs) can explain the temptation to commit fraud, and especially to exaggerate claims. In this respect, improving insurance pricing and adapting coverage to specific needs could help hinder fraud. Insurance policies are often considered unreadable by customers, with clauses that they may not understand. The lack of clarity and transparency in contracts is an issue insurers have to take into consideration when shaping their offers, on top of pricing issues. Customers are no longer ready to engage financially in products they cannot understand or see little use for, which is why tailor-made policies could partially help to prevent fraud. Some insurers have already made some steps to provide on demand insurance, with the possibility to switch your cover on and off, paying only for what you are actually using. This type of offer, closer to the actual needs of the customer, allows the customer to envisage the product in a much more tangible way and enables them to have better control over their spending, making the offer overall more attractive.
Pricing can also be improved through the use of data. For vehicle insurance, using telematics enables insurance companies to price their products depending on individual behavior. In this case, benefits are quite clear, both for the insurance company and the customer. Bad drivers will be less likely to use such type of insurance, and if they will, the pressure to modify their behavior will be much higher. Good drivers are rewarded by paying less for their own insurance.
Technology, beyond the reshaping of offers, can be used by insurers to improve their overall service. Understanding customers expectations is key to prevent fraud; it is particularly important when it comes to a younger demographic, one that is statistically more likely to commit fraud than any other. Offers can be improved by increasing the already on-going digitalization process, to allow a better, faster and more transparent access to insurance services, from underwriting to claim settlement. Indeed, a survey by Accenture shows that the speed and transparency of the claim settlement process is paramount for customers. 90% of customers also consider important or very important the real-time update of the status of the claim and the possibility to check that status at any time. This is where digitalization first impacts customers satisfaction accessibility. The influence of real-time access to your claim on your motivation to commit fraud should be investigated; it seems only natural to think that being able to see the status of your claim and expect a timely result can give the customer the feeling his/her claim is actually being taken care of by the insurance company.
Another aspect insurers should take into consideration is social media use. The same Accenture survey shows that 43% of interrogated people have or plan to read other customers posts about their experience in claims settlement. Access to online information and the almost customary practice to share opinions regarding a particular service provider should push those insurers to take better care of their clients and improve their services. If data on customers satisfaction greatly varies depending on the area surveyed, a study by EY shows that insurers are less trusted by the general public than banks or even pharmaceutical companies. In Europe, only 60% of polled people stated they completely or moderately trust insurers. This shows the insurance industry does suffer from a relatively bad reputation. To improve their image, insurers will eventually have to reconsider their approach to customer service in the long-term, beyond digitalization and automatization.
As insurance fraud is a real issue for the industry, it seems only fair that insurance companies invest in means to detect fraud. There are numerous ways insurers can try to detect ex post, from creating fraud detection units dedicated to the analysis of claims to the use of data analytics and other more automated systems. Insurers can also invest in the specific training of its claim-handling staff, or participate in industry-wide initiatives such as national fraudsters database like the Insurance Fraud Register in the UK.
Without going into details, those measures are absolutely necessary as the more data is gathered on fraud, the easier it gets to detect new fraud or fraud attempts. Nevertheless, this ex post process has most likely no impact whatsoever on the occurrence of fraud: insurers can detect fraudsters, they cannot prevent customers from becoming one. Considering the financial investment those measures represent, it seems that insurers would benefit from balancing their efforts to fight fraud by working harder on prevention rather than detection.
Collecting an exponentially increasing amount of data on customers and connecting big data (correlation) with small data (causality), using network analysis for instance, can be considered a reasonable use of technology to identify fraud. When it comes to prevention, the use of big and small data is however less obvious. Currently, technology can be used to automate flagging of potential misinformation when signing a new customer and therefore help insurers identify which future claims to investigate more thoroughly. The impact on the number of fraud attempts is difficult to measure here, but potentially marginal. In this respect, insurers have to rely on other types of solutions if they want to improve their fraud prevention strategy.
Image is one part the issue insurers can have a direct impact on, but they can also consider implementing psychological incentives into the design of their product. Insurtechs often do not possess the data they would need to combat fraud in a traditional way; they had to design alternative models to try to prevent it. Lemonade for instance gives unclaimed money to charity at the end of each year, and Friendsurance simply returns unused cash to policy holders.
To conclude, insurers should keep in mind that reliance on technology only to solve a problem as complex as insurance fraud is unlikely to work in the long-run. Combining what technology can bring in terms of computing and analytics with a deeper understanding of the psychological aspect of fraud may however prove more efficient for insurance companies.
 The impact of insurance fraud , Insurance Europe, 2013
 No Hiding Place : Insurance Fraud Exposed, Association of British Insurers, 2012
 Insurance fraud : issues and challenges, The International Association for the Study of Insurance Economics, 2014
 Insurance Consumer Fraud Survey, Accenture, 2010
 Why claims service matters, Accenture, 2014
 Reimagining customer relationships, EY, 2014