Psst.... Can You Keep a Secret?

Please don’t tell the Consumer Federation of America.  In an important new report (see item) from Moneysupermarket.com Sagittarians (Nov. 23 – Dec. 21st) have lower insurance claims.

Compared with the average driver, Sagittarians file seven per cent fewer motor insurance claims and receive eight per cent fewer convictions, Moneysupermarket discovered.

If the CFA gets a hold of this info, it will try to prohibit insurers from using birth dates to underwrite!

via  Cosmic Actuarial Outpost.

Comments

[R.J. Lehmann]

Hey, I'm a Sagittarian! I don't buy into astrology, but if it's an actuarially credible factor, then give me my discount, dammit.

[James Bowman]

That's all well and dandy, but you can't just give one twelfth of the market a blanket discount.

We won't know to underwrite until we have astrological tables, complete with houses, ascendants and knowing whether Saturn passes through Virgo, which would obviously be a disaster.

[James Taylor]

Very funny! In all seriousness though, how should one decide which factors should be allowed? Is it enough that a factor has a strong correlation? Must it also pass a reasonableness test for causality? What if it has correlation and causality but is discriminatory (like the recent zip code discussion in CA)? Should you have to show that something correlates, has likely causality and is the least discriminatory way to analyze something?

[martin]

I think people would say the zip code issue is actually related to risk: risk of living in a higher density area. There are more chances for fender benders as well as theft. As I understand it, most states' laws require that there be a statistical relationship between the risk classification and the risk premium. The idea behind risk classification is to cause discrimination between high and low risks. Competition will weed out bad risk categories and promote useful ones. California's approach is just a way to give a price cut to the majority urban residents, but it comes at the expense of the lower risk rural residents.

The state laws do not require causality because this is almost impossible to prove using statistics. Thus, we have to rely on logic and that also causes problems as people do not see a logical relationship between insurance credit scores and risk of loss either.