Insurance is a strange business, because unit costs are not known ahead of time and are different for every customer. Sometimes these costs are not fully known for decades, such as when asbestos claims surfaced in the 1970s and resulted in hundreds of billions of dollars in liability for manufacturers and insurance firms.
Not only are costs not known ahead of time, but they are different for each customer. In some cases the factors most heavily influencing claims are random. Who can know which coastal town in Texas is more likely to be hit by the eye of a hurricane? For car insurance, though, the customer is in much more control. While there is certainly still lots of randomness to car insurance claims, an aggressive driver has a much different risk profile than a conservative one.
Since this factor -- how someone drives -- is both heavily correlated to insurance risk and knowable ahead of time, car insurance companies make a big effort to quantify it. In the early days of auto insurance, this was pretty tough. Insurance companies had to make an educated guess based on age, gender, and driving history (tickets, accidents, insurance claims). Over time, though, insurance companies have gotten more sophisticated at collecting data.
What this is
Informal thoughts on stocks and markets from our CIO, Evan Tindell.