Combating Fraud in Insurance with Data
Well, it is International Fraud Awareness Week, focused on promoting fraud prevention and education. A fantastic initiative! Maybe I am naïve but I feel a bit sad that there is a need for “fraud week”. The insurance industry has a long and intimate relationship with fraud in many different ways. Insurance fraud can take place at a process or business function level, most notably in claims or underwriting. It can also take place at a source or organizational level – internal employees, external people, criminal organizations, or networks. It can be digital (edited photos) or physical (misuse of claims checks processing). Finally, it can be opportunistic or deliberate and organized. The different venues to commit fraud against an insurer are mind-boggling, with serious financial consequences. Insurance fraud is costly for insurers and indirectly their customers and society at large:
- Just in the U.S., the total cost of insurance fraud (non-health insurance) is estimated by the FBI to be more than $40 billion per year. This means Insurance fraud costs the average U.S. family between $400 and $700 per year in the form of increased premiums. Yet, fraud in the U.S. is relatively low compared to other fast-growing markets such as APAC.
- Globally, RGA research shows that 3.58% of all claims are fraudulent, and underwriting fraud is reported in 1.38% of all underwriting cases.
The sizable impact from fraud on the insurance market is increasingly being addressed by fraud detection, prevention, and mitigation technology tools and services, creating a substantial fraud detection market. According to ResearchAndMarkets, the global insurance fraud detection market size is expected to grow from $ 2.5 million in 2019 to $7.9 million by 2024, a CAGR of 25.8% from 2019 to 2024.
Unfortunately, fraudsters will continue to look for new opportunities and will also seek to leverage new technologies. Therefore, it is critical to advance our fraud prevention techniques and solutions with agility to stay ahead.
Use the Data Available
Besides the financial impact on an insurer’s results, society at large expects insurers to manage our collective premiums prudently including detecting and preventing fraud as much as possible. This requires a lot of data, a variety of data, and advanced analytic capabilities. Network analysis (quite often in knowledge graphs) is critical to identify criminal networks and organizations. Third-party data such as location, social media, obituaries, repair costs, and others help in faster identifying suspicious claims or applications. And finally, newer technologies (such as Cloudera’s) that facilitate cloud computing, machine learning, and streaming data, enable us to integrate and use structured, unstructured and third-party data to identify and address possible fraud earlier in the attempt, hopefully preventing fraud altogether. Fraud is one of the use cases where technology has made a huge, measurable difference: focusing on fraud detection and prevention, Emerj reports insurers have seen ROIs of up to 400% on their investments in fraud technology.
At Cloudera, we observe with our insurance customers that given current market developments such as COVID-19, hurricanes, wildfires, economic downturns, and low-interest rates, there is an increased risk of fraud. As a result, we’ve seen a recent reprioritization of fraud-related use cases on our platform. Since preventing a fraudulent claim results in a direct positive impact on the bottom line, fraud use cases tend to pay off quickly and create support for new fraud-fighting initiatives. It is very satisfying to be able to help prevent fraud, although I sincerely hope sometime in the next few years we don’t need a “fraud week” anymore in insurance. Until then, let’s continue to work on those technology initiatives and fight fight fight!
To learn more about some techniques and strategies for fighting fraud, visit our Fraud Prevention Resource Kit.