Basing Car Insurance Risk on the Individual

One of the long-running complaints among people who buy automobile insurance is the way carriers lump drivers together based on general characteristics -- age, sex, where they live and so on -- in setting prices.

These categories are insurers' attempts to predict the riskiness of the drivers they insure. Since they can't evaluate everyone individually, they employ generalizations. Carriers' experience tells them that these generalizations are broadly valid, but they acknowledge that they aren't perfect, and as a result some good drivers pay more than they should and some bad drivers pay less.

Some insurers have turned to other factors, such as credit scores, in an attempt to refine their ratings, triggering even more complaints.

But for about the past 10 years, one large carrier, Ohio-based Progressive Corp., has been trying out technology-based systems that carry risk analysis right down to the actual driver, or at least to the actual car. In the late 1990s in Texas, the company experimented with global positioning devices and cellular telephones mounted onboard consenting customers' cars to monitor when they drove, where they drove and how far they drove.

Now, Progressive is beginning a second round of experimentation. It is offering to provide 5,000 Minnesota customers with gizmos that plug into their cars' onboard diagnostic ports -- which most newer cars have -- and capture data on how far the cars are driven, at what speed and at what time of day. The device also gathers information about rapid acceleration and hard braking, but, lacking a GPS component, it does not record where a car goes.

The goal, as it was in Texas, is to get an exact reading of the driver's behavior. Based on that, Progressive is offering discounts to those who appear to be driving more safely than average. The discounts are based on mileage, speed and time of day, but not on braking and acceleration, which Progressive plans to analyze for "predictive value."

The experiments are "based on our belief that we can use technology in innovative ways to help people save money on their car insurance," said Mark Connally, online marketing manager at Progressive.

And if participants feel inhibited by the fact that their behavior is being recorded -- feeling forced in a way to become safer drivers -- both they and Progressive will benefit, he added.

The Texas experiment, known as Autograph and concluded in 2001, showed that "it is technically feasible to do something around usage-based insurance, and to get real data about the driver and the vehicle from the vehicle itself," Connally said. It also showed that "consumers are pretty enthusiastic about the concept," he said.

But Autograph used expensive technology, for which Progressive charged a fee, and because it sent data automatically to Progressive, it raised some questions about privacy.