There is no doubt that South Africa is currently experiencing an economic slump accompanied by significant financial stress for most citizens. The arrival of the Covid-19 pandemic has also added to this burden. And insurance fraud is never more top-of-mind than during times of economic and financial crisis. Financial hardship is often the root cause of insurance fraud, and insurers should consider adopting mechanisms to mitigate the impact of fraud on their financial results.
Accurate fraud statistics are hard to find, since not all cases of fraud are detected and insurers aren’t required to report on fraud. Some estimates place the cost of fraud to the South African insurance industry as high as R6 billion to R8 billion per year. South Africa has a markedly higher fraud incidence; it is often reported that up to 30% of claims in South Africa contain some element of fraud. In contrast, the United States reportedly shows estimated claims fraud losses of only 10% of claims.
To combat fraud, insurers may turn to various technologies. Large volumes of data, increased processing speed and broad acceptance of machine learning algorithms place powerful fraud detection capabilities at insurers’ disposal. Munich Re works with fraud detection firm FRISS to arm clients with customisable software aimed at streamlining the fraud detection process.
There are also softer approaches to combat fraud, many of which are influenced by findings from the field of behavioural science. Taking a cue from such findings, an insurer may for example consider including duplicate (or cross-check) questions in the insurance application process. In this way, inaccurate information may more easily be detected. Researchers have also found a measure of success by rearranging the structure of forms: requiring an individual to confirm accuracy of details and sign on the first page, prior to any questions being asked and answered. In short, asking a person to confirm that he/she is honest makes them act more honestly thereafter.
Incorporating a visual claims process may also reduce fraud. Indeed, some behaviourally driven approaches make use of video to discourage fraudulent claim declarations. Requiring the policyholder to submit a recorded video description of the loss event may lead to reduced claims fraud. While technology hasn’t quite evolved to the point of reliably detecting the facial cues of a lie, most individuals are certainly less likely to lie while on camera.
One shouldn’t discount the benefits of modern data analytics to come to grips with insurance fraud. One application of this, which has already been mentioned is the use of machine learning algorithms to build a fraud rules set, allowing the claims handler to determine when a suspicious claim should be referred for investigation. Data analytics software also allows for much more dynamic management information than ever before. Munich Re offers its clients the Non-Life Analytics Platform, which provides rich online reporting capabilities underpinned by our global expertise.
There is no doubt that our country, and indeed our industry, will have to adapt to a very different post-pandemic world. Now is the time to ensure we adapt our responses to fraud by making use of new research as well as the capabilities and technologies available to us. For more on this and other evolving insurance topics, do reach out to us for advice and assistance.
By Erika Bothma (Senior Consultant – Munich Re Global Consulting)
