How to Detect Auto Insurance Fraud

Investigators use public and private information to detect potential auto insurance fraud.

Claims adjusters and data experts at insurance companies and law enforcement organizations detect fraud in several ways. Suspicious claims are identified according to the company's proprietary statistical methods. Specialists review suspicious claims for more clues. Private citizens, usually unrelated to the insured, might suspect fraud and report it to police, fraud tip lines or fraud-focused organizations, such as the National Insurance Crime Bureau and the Coalition Against Insurance Fraud. Fraud raises auto insurance costs by at least 16 percent, according to author Saul W. Seidman in his book "Trillion Dollar Scam: Exploding Health Care Fraud."

Instructions

1. Use sophisticated computer programs and computing methods to detect fraud, according to "Surveillance Technologies and Early Warning Systems: Data Mining Applications Methods for Risk Detection." Investigators also use data to identify suspicious relationships associated with the insured. They review financial statements for unusual cash flows. They look for associations with known insurance crime rings, according to author Pamela Meyer in "Liespotting: Proven Methods to Detect Deception."

2. Evaluate supervised and nonsupervised potential fraud claims, according to the "Handbook of Statistical Analysis and Data Mining Applications" by Robert Nisbet, John Elder, John Fletcher Elder and Gary Miner. Investigators use both supervised and nonsupervised methods to evaluate the suspicious claim. Insurers develop fraud-detection models based on demographic, attitudinal and business data information. Supervised claims involve analysis of historical claim values to the insured's claim values. A suspicious claim is compared to previously identified frauds. Unsupervised analysis involves statistical identification of unusual amounts, repairs, medical care and other red flags. Unsupervised analysis also connects abnormal values in current claims to previously known fraud cases. Neither method absolutely confirms fraud. Additional analysis helps to identify higher probabilities of insurance fraud.

3. Recognize common auto insurance fraud. According to author Saul W. Seidman, fraud costs other insureds higher insurance premiums. Fraud costs of $1.05 for faked thefts, $2.15 for previous damages, $2.20 for overcharges from body repair shops and $3.00 in staged car accidents for every $100 in paid claims. Auto theft continues to increase in the United States. According to the National Insurance Crime Bureau, more than 57 percent of stolen cars disappear. Professional crime rings ship stolen cars overseas or sell the car for parts. However, some stolen cars are resold in the U.S. to unsuspecting consumers. The NICB's VINCheck helps prospective owners to check vehicle identification numbers for free.

References

• "Trillion Dollar Scam: Exploding Health Care Fraud"; Saul William Seidman; 2008; Page 43
• "Surveillance Technologies and Early Warning Systems"; Ali Serhan Koyuncugil, Nermin Ozgulbas; 2010
• "Liespotting: Proven Methods To Detect Deception"; Pamela Meyer; 2010
• "Handbook of Statistical Analysis..."; Robert Nisbet, John Elder, John Fletcher Elder, Gary Miner; 2009; Page 350
• National Insurance Crime Bureau: Auto Theft
• National Insurance Crime Bureau: VINCheck

Resources

• "The U.S. Experience with No-Fault..Insurance:"; J. M. Anderson, P. Heaton, S.J.Carroll; 2010
• Coalition Against Insurance Fraud
• "Fighting Fraud: How To Establish and Manage an Anti-Fraud Program"; Gerald L. Kovacich; 2007
• "Criminal Investigation"; Wayne W. Bennett, Kären M. Hess, Christine M. H. Orthmann; 2006
• "Delay, Deny, Defend: Why Insurance Companies Don't Pay Claims"; Jay M. Feinman; 2010
• "Data Mining and Knowledge Discovery Handbook"; A.H. Eschenfelder, Oded Maimon; 2010