Insuring higher rates...

Eugene Volokh is excited by the prospect of pay-as-you-go insurance. I'm not.

The thing about insurance is that the more data you get, the closer you can estimate the likelyhood that someone is going to file a claim and charge accordingly. But if you could perfectly estimate the likelyhood of claims, there would be no reason to have insurance, because the cost of insurance would equal the cost of claims.

I'm no statistical expert, but my guess is that the stats that insurance companies use are pretty broad - for every person that they charge higher rates to who files a claim, there are probably a bunch who pay higher rates but don't file claims.

I'm also really apprehensive about the idea of using speed as a means of judging people's insurance rates. I've gotten exactly one speeding ticket in my life (while driving a collge owned 15 passenger van to a debate competition when I was in college, on RT 29 in Virginia). My insurance rates didn't go up, because it was my first 2 point ticket, but I did lose a safe-customer discount when I moved and switched insurance companies. If you aren't familiar with 29 South, it's a giant speed trap and revenue maker for the little towns along the way. At least 3 other people at the debate also got tickets.

But if someone else had been driving the van that day, I would have a perfectly clean driving record. Speeding tickets are a horrible judge of driver safety - they are more a judgement of the unluckyness of the person who got it.

I suppose one could argue that knowing the driver's speed via the proposed devices would be fairer. But I think it would just result in higher rates to everyone.

I drove most of the way to work today at 80 miles an hour. I don't think that was an unsafe speed at 7:15 in the morning on a 6 lane interstate, with almost no traffic around - but I bet Progressive would disagree.

The funny thing is that Progressive has tried this before. Patrick Bedard looked at this 8 months ago and has some excellent analysis.