Spring 2020
March 06, 2020
How do insurers determine how much to charge you for your auto insurance?
According to state regulators, premiums should be based on risk, or how likely it is that a customer will be in an accident and file a claim, plus a little extra (private insurers are for-profit businesses, after all). So, the higher the perceived risk, the higher the premium and vice versa, right?
Maybe not.
The article Why You May Be Paying Too Much for Your Car Insurance (Consumer Reports, February 25, 2020) takes a deep dive into the highly secretive world of auto insurance pricing and the challenges of regulators charged with protecting consumers from discriminatory practices.
Insurance providers closely guard as “trade secrets” the details of how they set prices for individual customers. Even state regulators are typically in the dark about—or, at least, prohibited from revealing to the public—the data and algorithms, or methods, that insurers use to set premiums. The algorithms should, however, use data related to risk to determine the likelihood of a customer being involved in an accident.
In practice, however, insurance providers may also be looking at something entirely unrelated to risk.
A rate adjustment proposed by Allstate in Maryland in 2013 offered a rare, detailed look at pricing algorithms and customer data. Analysis of the information revealed the application of personalized pricing—setting individual customers’ prices based on how much companies believe each customer is willing to pay.
Allstate claimed this “customer retention model” was needed to prevent “sticker shock” that might lead customers to drop their policies in the face of the price increases some would encounter as a result of a new risk analysis algorithm. In reality, it identified customers who were already paying high premiums and were unlikely to drop their policies because of rate increases. These customers’ premiums would be raised by up to 20%, while customers who were currently already paying low premiums would see increases of no more than 5%. On the flip side, customers whose rates should have gone down based on the new risk algorithm would never have seen the discounts they were due. Rate reductions were capped at a meager 0.5%.
Maryland ultimately rejected Allstate’s proposal, but what about other states? Public records indicate Allstate mentioning using customer retention models in at least 10 states, including Illinois. And it is unlikely that Allstate is alone among auto insurance providers in its use of personalized pricing. Hampered by secrecy, complexity, and pressure exerted by the insurance industry, insurance regulators are frequently unaware of how prices are set for the residents of their states.