OVER the year, we have seen a number of high profile cases of insurance fraud which, thanks to the efforts of the insurance companies concerned were successfully identified and brought to light.
Sadly, this is only a tiny fraction of the full extent of fraudulent activities that insurers are subjected to – most fraud incidents are never discovered. While there are varying estimates on the financial effects of insurance fraud, these are at best just guesses and I will not take the time to discuss them. What is certain however is that thanks to insurance fraudsters, we all end up paying more for insurance than we should and have to endure long delays waiting for our claims to be processed. The good news is that actuaries have developed ways to detect insurance fraud.
First, let us take a look at the two main types of insurance fraud.
The first is opportunistic fraud. An example of this is a person who legitimately gets into an accident to which he is entitled to an insurance pay-out. This person might go on to also claim that a dent on another part of his vehicle was due to that accident, when in actual fact that dent was formed by an unrelated event – probably before insurance cover was put in place.
The second type is organised crime where criminals operate fake or staged accidents in order to claim from the insurance company. A much more heinous example is when these organised criminals operate car ‘chop shops’ to completely disassemble cars thereby wiping their existence from the face of the earth – making it easier to claim that a car was stolen. In Marine insurance, ship owners have been discovered to contract pirates to steal their ships in order for the owners to claim insurance money to buy new ships.
Using statistical techniques, it is possible to flag certain insurance claims as being more likely to be fraudulent than others. To better understand how this is possible, consider what happens when you initially get insurance cover. When you take out an insurance policy, you supply your age, gender, occupation, place of residence, salary and so on. These factors are used by actuaries to determine the chances of you having an accident – in the case of a motor policy, or your chances of dying – in the case of a funeral policy. Such statistical techniques have passed the test of time and have been shown to be quite accurate especially for a large collection of policyholders.
The same techniques can be applied at the point of making a claim. At the point of claim underwriting the claimant provides details of the accident and we get to record various data attributes such as the delay between the time of accident and the point of reporting of the claim event. Other pertinent information such as the time of occurence of the claim and the remoteness of the claim location can all be used to calculate the chances that a claim is fraudulent. All in all, over a dozen data attributes can be fed into the model so as to calculate the chances of a claim being fraudulent.
Currently, when a claim is lodged with an insurance company, it has to go through a rigorous verification system which not only increases costs, but causes unnecessary delays for policyholders. Insurers who have developed predictive models such as the one I have described above, have enjoyed a decrease in claim management fees since they only focus the bulk of their investigative attention to those claims which would have been flagged as highly likely to be fraudulent.
With the worsening of the economic environment, I believe that insurers need to step up their defences against insurance fraud. More and more people are likely to view insurance companies as a low hanging fruit that they can take advantage of. However, greater cooperation is needed throughout the insurance industry if meaningful progress is to be made. One way of doing this is establishing a centralised fraud database that will enable the quick identification of policyholders who for example, switch insurance companies to stay undetected from the fraud radar or would have been caught trying to defraud or inflate claims from multiple insurance companies.
Thomas Sithole is an Actuarial Analyst and the current Head of Enterprise Risk Management (ERM) Solutions at Bluecroft Actuarial Solutions. If you have any comments or questions concerning any of the matters discussed on this platform, please do contact him on email@example.com or his twitter handle @ZimboActuary.
Follow us on Twitter on @FingazLive and on Facebook – The Financial Gazette