I saw this today - California Auto Bill Passes Assembly Insurance Committee - and it prompted me to ask some of my insurance analytic friends. Here's what they came up with:
- There is a clear opportunity for Predictive Analytics to help identify which factors are predictive of fair (or unfair) rate changes.
- Reality is that every year 90% of the drivers do subsidize the 10% that have claims.
- Because there are lots of votes in cities there is pressure from politicians to lower the rates there. However auto crime(vandalism, theft, fraud) is much higher in city areas driving costs of claims up.
- At the same time, many inner city folk are among those least able to pay for the real exposure.
- Better use of Predictive Analytics could identify which farmers are high risk and which inner city folks are low risk and so properly balance the risk and prices.
This is what multivariate analysis is all about. Just as with credit, some people with good credit scores are still high risk due to other factors and others with low credit scores are better risks because of their good factors.
The bill is right on about one thing - analysis, not politics, should be the driver.
Thanks to Wendell Larson for his expert advice.