In an attempt to contain fabrication of claims rampant in the third party arm of motor insurance, National Insurance Academy has proposed a “scientific method” which would facilitate insurers and the courts to point out any anomaly in the claims.The method is based on seven processes, four preventive and three retrospective tools. Preventive tools, which would come in handy while an insurer is settling a claim, comprise stress analysis, which would gauge the stress in a claimants voice to figure out if he is giving a real account of things. The other is “red flagging”, which essentially means to report bogus claims thereby creating some sort of a bank so the next time a surveyor visits a particular case he can dip into the bank to help him identify a pattern.Predictive modelling is another tool by which an insurer will have hands on information on the type of vehicles making the type of claims in the type of area. Says K C Mishra, director, National Insurance Academy: “Like in the US, the claims data for every vehicle is analysed to identify a pattern. For example, a big car is more accident prone in hilly areas.” The last is data base searching, adds Mishra: “We know of a place which has been infamous for raising third party claims by throwing dead bodies in front of speeding trucks. Having records would help an insurer to be extra cautious to settle claims in that area.”Retrospective tools, which would help an insurer find out if the claim was bogus after it has been settled, comprise “exception report” which would point out any exception in the number of claims in various branches of the insurer. It would also include “online analytic processing”, under which a person and his records would be maintained to keep track on the claims he has made.The third retrospective tool would be “link analysis” where a link would be searched for similar type of accidents in different parts of the country. Says Mishra: “If a fraudulent claimant raises claims in different parts of the country to avoid suspicion, by our records we would be able to establish a link and be able to investigate his claim.”But these methods are still waiting adoption as motor third party is still under a lot of discussion after the motor insurance was de-tariffed on January 1, this year. Third party, however, retained its tariff and was made to be handled under a motor pool comprising all insurers. The premium and claims were to be handled by all insurers collectively on April 1, though private vehicles were pulled out of the pool, and tariffs still exist.Mishra feels that insurers desperately need some respite in third party as it is highly loss making and methods to control fabricated claims should be put in place as soon as possible. “Right now our proposal is bookish since it is only theoretical, but to make it practical the insurance industry has to look in this direction. We have floated the proposal but we are still waiting for motor to start functioning smoothly to invite discussions.”Says Shiv Kumar, national head personal line, Tata AIG General Insurance: “The loss ratio in motor is 200 per cent. Of this, third party constitutes a large chunk. And since claims in third party are unlimited, insurers have also paid out Rs 10 crore on a premium of Rs 700.”Clearly, the need is there not only for insurers but for consumers too, who end up facing higher premiums because of such frauds.7 steps to assess fraud-free claimsPREVENTIVE MODELLING • Stress analysis: Recording the voice of the claimant while taking an account of the accident, to establish the authenticity in his statement.• Red flagging: Reporting of bogus claims by a surveyor to create a record bank so that a pattern of bogus claims can be identified.• Predictive modelling: To have hands on information on various vehicles and their types to assess whether the claim is realistic.• Data base searching: Keeping records of various regions so that a claim raised in a region infamous for fraudulent claims can be dealt with.RETROSPECTIVE MODELLING• Exception report: Keeping a record of the average number of claims raised by a particular branch so that if the average is exceeded, anomaly is detected.• Online analytic processing: Keeping a tab on the number of times an insured raised a claim to see if there is a possibility of fraud.• Link analysis: Two similar accidents occurring in two different parts of the country may have some link. By whom was the claim raised, how did the accident happen, would help determine if the claims were fraud.
