Insurance fraud detection

According to the information available, almost every seventh submitted Insurance claim is fraudulent. This not only includes reports of insurance claims that never happened, but also cases in which indicated sums of loss were much higher than in reality. Insurance fraud causes huge losses in the millions of euros per year for insurance companies.

There are several prescribed steps used for detecting insurance fraud in the liquidation process. The liquidation process should not be unfairly prolonged because the time taken from reporting the insurance claim until its refund is a very sensitive issue for clients. That is why the effective selection of suspicious insurance claims is very important for dealing with insurance fraud.

It is standard for adjusters to focus on insurance claims with higher cover value or insurance claims with the date of submission listed as immediately after its closing, etc. These claims are very simply selected out of the production system.

There are also cases of insurance fraud that are difficult to detect in the system. These cases consist of different coincidences in the dates, accumulation of several claims connected with one person (insured, agent), accumulation of paid-out benefits to one bank account or an insured’s address. These and many other suspicious aspects of claims are possible to solve by Business Intelligence tools and sort them according to the rate of suspiciousness. Moreover it is possible to create combinations for suspicious aspects out of partial determinations which create a more complex sample of fraudulent activity. As an example, there are claims in the case of a car accident when the police are not involved, the loss is reported for an older car and less than one month has elapsed from the sign-up date of the insurance contract. Aspects of insurance claims that are selected and evaluated in this way create a ranking of suspicious insurance claims which should be solved by Insurance companies primarily.

Using such a mechanism, it is possible to bring all the information about suspicious or confirmed cases of insurance fraud to the primary evidence of insurance companies in their production system. From this point they can subsequentlybeused in processes that prevent the liquidation itself – by reporting a claim (by asking detailed questions about the claim by means of the call centre) or by the assignment of an insurance contract (using the so-called black list of suspected persons).

Business Intelligence tools allow even further monitoring.

Using methods of data mining, it is possible to detect other scenarios of fraudulent action based on data of confirmed insurance fraud, using only statistical methods. The scenarios identified in this way might be used for detection of insurance fraud in the future. With these solutions, it is possible to make the system of insurance fraud detection more effective.

To summarize, we can say that Business Intelligence tools facilitate the processing of a huge amount of data and effectively select suspicious insurance claims which should be investigated primarily by the insurance companies. At the same time, total time for verifying insurance claims that are not suspicious can be kept at the necessary minimum for processing claims. This is a significant competitive advantage for Insurance companies.

Meeting requirements following from the Solvency II

The structure of Solvency II is based on the principle of three pillars consisting of requirements and processes from three main areas.

The first pillar is primary focused on quantitative processes and requirements – calculation of capital adequacy by the standard model for definition of minimal (MCR) and solvency capital requirements (SCR), or by using the checked, calibrated and approved (by a regulator) internal model.

Following IT needs follow from the requirements leading mainly to estimation of assets and liabilities taking into account risk, calculation of minimal and solvency capital in compliance with the standard model, or with the checked and approved internal model:

• Data integration and consolidation from internal and external systems
• Designing actuarial models and product evaluation, or integration with existing actuarial solutions, calculation of individual risks
• Calculation of SCR and MCR according to the standard model
• Designing an internal model
• Model stress testing
• Statistical processing and analysis of model quality

Each IT need from the first pillar could be solved by standard Business Intelligence tools as well as by BI solutions specialized in risk management. But there is to say, that especially requirements for risk calculations in the area of 1st pillar are mostly covered by standard actuarial tools used within insurance calculations. The Directive Solvency II puts the maximal stress on consolidation of data base and data auditability used in calculations and that is very useful to solve by Business Intelligence tools.

In the second pillar it is expected introduction of a complex process of own risk and solvency assessment within a financial institution. The complex process contains processes of risk management and its control, qualitative assessment of risk capital as well as elaboration of risk management principles into all key processes. It could be supposed that reassessment and change of key processes and probably also introduction of new process-integration systems need to come up in some insurance companies.

Requirements from the second pillar lead mostly to estimation and management of operational risk and then to its minimization by automation of internal processes and elaboration of risk assessment principles into all key activities of the financial institution. The requirements can be summarized as follows:

• Tracking, documentation and management of operational risk
• Estimation of operational risk
• Tools for documentation of processes
• Workflow and tools for process integration
• Capital classification, qualitative assessment of risk capital

Business Intelligence tools within the second pillar are primary useful in the area of operational risk estimation and capital classification. Other areas of the second pillar are especially the domain of integration and BPM platforms.

The third pillar concentrates on report preparation and report publishing for a regulator and public, having the goal to provide objective and entire information on status of a financial institution and risks connected with offered products and services. These requirements from the IT point of view concentrate mainly on following:

• Data analysis and visualization,
• Sharing analytical information within financial institution,
• Automation of creation of output documentation and its publishing,
• Designing predefined and ad-hoc reports by users,
• Approval and check of report for regulator.

All the matters of reporting required within the third pillar have been a natural domain of Business Intelligence for a long time, but in the case of requirements of the third pillar Solvency II it is needed to solve also the support of individual process of strict approval and check of reports for regulator and public.

Conclusion

To implement projects in listed areas successfully it is needed to take aim not only at selection of Business Intelligence tools having required functionalities and features, but also at selection of right implementation partner which has rich experience in the area of Business Intelligence, as well as in the area of insurance and realization of complex integration and BPM projects.