Study: Auto Insurers Underpay On Claims

For researcher Keith J. Crocker, the fact that insurance companies would not endorse his findings is no surprise. "I didn't think any insurance company would come out and say they are purposely underpaying claims, even though from an economic and realistic standpoint they do," he said.

Mr. Crocker, a professor of business economics and public policy at University of Michigan Business School in Ann Arbor, Mich., worked on the study with Sharon Tennyson, associate professor of policy analysis and management at Cornell University in Ithaca, N.Y.

The pair used data from a 1987 Insurance Research Council survey in which 34 auto insurers reported on auto injury claims closed in a two-week period with data from accidents all over the United States. Mr. Crocker said he did not use data from no-fault states. Thirteen thousand claims were examined.

Mr. Crocker and Ms. Tennyson, who researched their scholarly 45-page paper for publication later this year in the Journal of Law and Economics, even reduced their theory to a mathematical formula expressing the insurer cost-minimizing strategy with E representing the "falsification cost parameter."

"The miserly proclivities of insurers when settling claims is legendary, and occupies a place in the pantheon of business stereotypes along with the sharp horse trader and the obdurate banker," the authors wrote. "And, while it may appear that a dollar saved through reduced claims payments amounts to a dollar earned, insurers engaging in a strategy of systematic underpayment of claims do so at their own peril. At the very least, underpayment is likely to generate administrative costs to deal with aggrieved claimants, and the most egregious shortfalls may spawn protracted episodes of litigation and can result in substantial penalties."

According to the researchers, their findings support the idea that insurers use claims payment strategies designed to mitigate claimants' incentives to exaggerate losses–in other words, they underpay claims to reduce the incidence of fraud.

The claimants in the bodily injury liability claims studied were eligible for financial losses due to injury damages (medical bills, lost wages and rehabilitation expenses) and general damages (other losses associated with the injury, such as "pain and suffering"). About 75 percent of the claims involved a sprain injury, while roughly a third included a claim for lost wages.

"Automobile insurance claiming is an area where fraud is legendary, and automobile liability insurance claims are thought to be particularly prone to exaggeration due to the possibility of compensation for pain and suffering in addition to economic losses," the researchers wrote.

At the American Insurance Association, Dave Snyder, assistant general counsel for the Washington- based group said: "We would not agree there is a conscious strategy to underpay."

What is shown, he said, is that "serious injuries are, as one would expect, being generously compensated, and less serious injuries and injuries subject to fraud don't receive the same level of compensation."

The data, according to Mr. Snyder, shows that insurers "are acting responsibly within the law, balancing need to fight fraud with the need to compensate legitimate claims."

According to the study: "The severity of the injuries often encountered in automobile accidents, such as sprains or other soft-tissue injuries, may be inherently unverifiable, and some of the types of claims filed, such as those for lost wages, may be easily manipulated by claimants."

The researchers found that claims with low falsification costs (such as those claims that are easy and inexpensive for claimants to fabricate) received, on average, lower payouts in injury-related financial damages. Easy-to-detect, non-sprain claims (for example, contusions, amputations, fractures and burns) were paid at a marginal rate of 78 cents on the dollar, while those involving harder-to-diagnose sprain injuries received about 71 cents.

Likewise, non-wage claims were compensated at a marginal rate of 80 cents on the dollar, while claims that involved wage losses (which usually require minimal documentation and can be inflated by a claimant's malingering in returning to work) paid at a marginal rate of 71 cents, the researchers found.

"Since the amount of general damages awarded is usually linked to the amount of direct financial loss experienced by the claimant, there is an incentive for claimants to exaggerate the amount of their financial loss," the researchers wrote. "General damage awards are often argued to be the primary motivating factor for liability claims fraud."

The researchers found that insurers can combat a great deal of fraud simply by investigating and denying claims in which the potential for deception is great. But when cases lack observable physical markers that can signal fraudulent behavior, they said insurers believe they have little choice but to resort to a strategy of underpayment.

Mr. Crocker said that while there might be some legitimate claims that get shorted, the situation is not unlike that of insurance risk classifications which charge all 24-year-old males a high auto insurance rate, even though some may be excellent drivers with blemish-free records.

"Economically it makes sense," he said, to underpay the unverifiable claims and keep rates low for the majority of policyholders.

"When audits of claims cannot uncover the true loss amount, the optimal strategy of an insurer is to reduce, at the margin, the settlement payment as a function of the claimed amount, thereby mitigating the incentives facing claimants to expend resources on claims inflation," he wrote.

The paper noted that "underpayment of claims, however, may generate costs to the insurer from claims negotiations and potential litigation, and can result in civil damages if the insurer is found to have engaged in bad faith. As a result, insurers must balance optimally the effects of claims underpayment on reducing the incentives for falsification, on the one hand, against the costs of underpayment, on the other."