Insurance fraud is rising fast. Indeed, statistics published in the Netherlands in 2013 by the Verbond Van Verzekeraars1 suggest it has increased by 25 per cent in the last five years, adding €150 to the average policy. Concerned by this growth, many insurers are looking to improve their fraud defences by reviewing their claims and new business processes and investing in enhanced technologies.
One key challenge they face is the sheer diversity of fraud. Perpetrators include opportunists who exaggerate claims to get better deals or to cover their excess, as well as deliberate fraudsters and members of organised gangs, who often invent claims or stage accidents in order to make money.
In tackling these fraud cases, the biggest obstacles many insurers face are the failings of their detection systems. Legacy solutions and difficulties integrating data from suppliers and third parties result in incomplete and unreliable data. Current data views rarely extend beyond a single customer, so it’s difficult to get a full picture of a claim and spot high-risk relationships.
A recent Europe-wide survey of insurers carried out by SAS revealed shortcomings in the way some insurers are fighting fraud. Only 13 per cent of respondents were making use of a comprehensive range of techniques including business rules, business analytics techniques and advanced analytics. Equally just 21 per cent monitor their fraud levels in real time.
One of the main problems with prevailing approaches is the large volume of false positives they generate. Over-reliance on a fraud detection approach based around red flags can generate vast volumes of alerts, which reduce operational efficiencies and drive up cost. These high rates can also negatively impact the client experience. The claims process is the shop window of any insurer and they need to avoid treating innocent customers as fraudulent.
A Second Line of Defence
In terms of detection, the first line of defence is and should remain the claims handlers themselves. Analytics should not be viewed as a replacement to this service, rather as an additional line of defence, helping these expert teams pinpoint additional fraud that they could not otherwise have detected.
Existing approaches often fall short because investigators look at claims in isolation. In addition, they fail to evaluate the history, view the network holistically or run comparisons with peers. Many insurers use systems focused on monitoring transactions, which may work well for individual claims fraud but are less effective at monitoring customer behaviour across multiple claims and lines of business in order to identify customers that appear normal on the surface, yet operate “below the radar.”
The latest breed of advanced analytics solutions provides an alternative approach to detection.
They typically start by extracting the data that exists across the insurance company and then clean it if necessary. They can then effectively connect the data to provide a holistic view across the organisation.
A hybrid approach can then be applied, incorporating everything from business rules, to anomaly detection models, to social network and text analysis in order to identify potentially suspicious cases. These can then be presented back for the business to accept into investigation or pushed back into the existing process.
This analytics-based approach benefits insurers by finding more fraud and by reducing the false positive rate. In the recent SAS survey, of those insurers using business analytics, 57 per cent had seen the amount of fraud they detected year-on-year increase by more than 4 per cent. In contrast, of those insurers with no automated solution or using only business rules, only 16 per cent had seen a similar increase.
Critically also, analytics can be deployed without disrupting the existing process for handling claims.
We expect this analytics-based approach to fraud to continue to grow. The larger insurers have been quick to adopt this approach, but interest from smaller insurance companies is growing. If they take the plunge and implement advanced analytics, they will see significant benefits, as the survey findings show.
Simply put, investing in a hybrid analytics approach, making use of multiple analytical techniques and combining the results, enables an insurer to proactively tackle the problem and to ultimately detect and prevent more fraud.Insurers today should not accept fraud as a cost of doing business. Instead they should be looking for the most efficient ways of detecting, preventing and managing claims fraud across all their business lines. Read more about how you can tap into the benefits of this approach.