Skip to Main Content

Knowledge Center

Underwriting and Claims

Global Claims Views: India - fraud detection tools

Insurance Fraud

Insurance fraud can present in many guises. Some are basic and unimaginative, while some are far more subtle and ingenious. The fundamental principle of commission of fraud tends to be the same – deliberate non-disclosure or misrepresentation of material information with the intention to obtain unauthorized benefits.

The extent of the problem

According to Indiaforensic Research, India’s insurance sector loses INR 300 billion (US$6 billion) every year due to fraud, representing a loss of 8.5% of total industry revenue. Additionally, six times more fraud is seen within the life insurance sector (which accounts for 86% of total insurance fraud) than in the non-life / general insurance sector.

Indiaforensic also found in 2011 that mis-selling of insurance policies was responsible for 36% of fraud, and fake documentation for 33% of fraud in the life insurance sector.

RGA India recently conducted a survey with our clients which compared the incidence of fraud in 2012 with 2011. The survey, scheduled to be published in late 2013, found that 41% of the participants said mis-selling has decreased due to proactive measures taken at the proposal stage to control fraud. However, 56.5% of the participants said submission of fake documents has risen by 7.3% and incidence of non-disclosure has increased by 7%. Additionally, more than half of the survey participants believe this recent increase in fraudulent activity has contributed at least 3% to the cost of insurance, with some participants believing the cost increase may be as high as 20%.

How can we address the problem?

Managing fraud presents a great challenge for the insurance industry. Insurers are under constant pressure to cover new risks and develop original products. This pressure, combined with the fast-evolving business and technology landscapes, makes for a favorable environment for fraudsters to use to come up with innovative ways to stay one step ahead of fraud detection. Often the insurance industry is left trying to play catch-up when managing the ever-changing nature of fraud and abuse.

Fraud is unlikely to ever be eradicated completely, but there are steps we can take to control it effectively.

The first step in the process to control fraud is, of course, to detect fraud. Fraud detection tools and techniques, which can be used to identify actual as well as potential fraud, fall into two primary categories: traditional/manual and artificial intelligence.

Read More +

Traditional/manual techniques

Traditional techniques of detecting fraud include:

  • Manual assessments and desktop investigations of targeted claims.
  • Manual data processing techniques, both to validate claims and to detect claims that are suspicious and could be fraudulent.
  • Internal audits and post payment claims audits, to detect suspicious claims settled due to lack of evidence and flag them in the event any future claims are made by the same claimant. A person who has defrauded an insurer once will often attempt to do so again, and usually using the same methods. Large patterns of fraud can be unearthed using this method.

Other traditional techniques include ‘Random Welcome’ calls to prospective or new policyholders to confirm no mis-selling, mystery shopping to detect provider fraud, and having a dedicated risk control unit (RCU).

The traditional, manual approaches of detecting insurance fraud are costly and inconsistent for insurance companies. Close to 50% of the respondents in our recent fraud survey believe that experience analysis, having an RCU and using random welcome calling are the most effective tools for detecting fraudulent activity. However, in isolation, these are not adequate to control fraud.

Artificial Intelligence

Industry uses of artificial intelligence may include:

  • Data mining and experience analysis: This is the automatic (or semi-automatic) analysis of large quantities of data or groups of data records to extract previously unknown patterns. Data mining uses information from past data to analyse the outcomes of specific problems or situation that may arise. Data mining can also be used to determine functional strategies and develop new underwriting and claims guidelines.
  • Automated red flag systems: These systems use specific criteria to identify claims with suspicious trend items. Such systems need regular monitoring and periodic evaluation to verify that the cases being flagged deserve scrutiny.
  • Profiling systems: These systems may also help detect trends and abnormal patterns of behaviour among an insured’s claimants, advisors, providers, etc., to enable identification of the nature of the fraud being perpetrated.
  • Predictive modelling: The methods listed above are often disadvantaged by the fact that instances of fraud can present similarly in content and appearance to genuine claims. However, they are not usually identical. The techniques covered so far may point to actual or potential fraud, but predictive modelling can help to unearth future fraud. This is a process whereby current facts, historical facts and abnormal patterns of behaviour are used to develop predictions about future events and behaviours. Predictive modelling is rapidly gaining attention in the insurance industry as more structured data becomes available, allowing sophisticated analytics to advance underwriting, claims and risk assessment knowledge. It can be used to identify claims that are most likely to be fraudulent in nature and also to triage claims, allowing assessors to focus on claims most likely to have the biggest impact on an insurer’s bottom line.

In the RGA fraud survey, which was completed by 24 participants from 20 companies across India, only two respondents indicated they are currently using artificial intelligence in fraud detection, but 50% of the remaining respondents stated they plan to develop the capability for using artificial intelligence in the future.

Expertise gathered during manual assessments and audits will help to develop system frameworks and rules engines. Artificial intelligence tools will not only provide insights into past trends but will also help to create predictions about future events and behaviors which could improve the industry’s ability to detect and manage fraud exposure. The future is likely to lie in blending the use of artificial intelligence with traditional fraud detection methods.



© 2013 Reinsurance Group of America, Incorporated (RGA). All rights reserved. No part of this publication may be reproduced in any form without the prior permission of the publisher. For requests to reproduce in part or entirely, please contact publications@rgare.com.

RGA has made all reasonable efforts to ensure that the information provided in this publication is accurate at the time of inclusion and accepts no liability for any inaccuracies or omissions.

Download "Global Claims Views"

The Authors

  • Rahul Gupta

    Senior Analyst
    Claims Services

    RGA Middle East
    Send email >

  • Peter Barrett
    LLB, FCII, FAMS
    Global Head of Claims, 
    RGA UK
    Send email > 

Summary

The fundamental principle of commission of insurance fraud tends to be the same – deliberate non-disclosure or misrepresentation of material information with the intention to obtain unauthorized benefits. How do we address this problem? 

 

Download "Global Claims Views"
  • big data
  • asymmetric
  • insurance fraud
  • insurance fraud prevention