As I have been investigating price optimization (PO), I learned that there are two primary concerns: The first concern is whether the application of PO is a move away from risk based rating. The second concern is whether PO results in rates that are unfairly discriminatory, particularly to low income consumers. In other words, would two identical risks get different rates if PO techniques are applied. These are the two primary concerns that the MD, OH and CA insurance departments are trying to address, albeit in different ways.
Adding to the complexity is that each state (and just about everyone you talk to) has a different definition of PO; and PO techniques can be applied both at the class and individual policyholder level. The later approach is what is commonly done in unregulated markets but there is an accusation that some insurers might have adopted this approach in the regulated markets as well.
The Ohio and Maryland insurance departments issued bulletins that strive to bring an end to using PO at the individual policyholder level, while California issued a Notice that applies to both classes and individuals. The California Department of Insurance “Notice Regarding Unfair Discrimination in Rating: Price Optimization (dated February 18, 2015) states that “any method of taking into account an individual’s or class’s willingness to pay a higher premium relative to other individuals or classes is unfairly discriminatory because it does not seek to arrive at an actuarially sound estimate of the risk of loss….”
In a letter Meryl Golden, General Manager at Earnix, sent to all insurance commissioners, she explains “the term ‘price optimization’ as the use of advanced statistical and analytical techniques to inform an insurer’s judgment when setting its rates within a competitive environment. It is not about ‘profit maximization’ and is not utilized to charge insureds the highest price the market will bear. Instead, price optimization is utilized as a decision‐support tool which suggests variances to an insurer’s indicated cost-based factors taking into account, among other things, its competitive position and business goals while considering its new business conversion and retention ratios. These adjustments are not applied at an individual risk level. Instead, our software applies adjustments to an insurer’s selected factors at a rating class level across the entire rate schedule to suggest an ‘optimized’ result within predetermined constraints, including actuarial standards and regulatory requirements.”
Another seasoned insurance product professional has explained “price optimization” as merely doing what product managers have always done by gut feel, decide on which class of risks is more attractive for their strategic growth plans as compared to another class of risks, and making a broad adjustment across the class so pricing for that preferred class is more competitive. This use of “price optimization” does not group two insureds with similar risk profiles differently. This use influences the determination of factors within the rating class. Class plan rating using this tool stays within the bounds of the loss cost estimate.
Meryl agrees that Insurers should not be allowed to unfairly discriminate between two otherwise identical risks within the same rating class or “redline” protected or lower income classes. She promotes use of a price optimization technique that does not affect an insurer’s obligations under the law, using price optimization to suggest adjustments to an insurer’s cost-based factors to better meet its business goals in a competitive environment, but not used to develop an insurer’s cost-based factors.
Robert Hunter, of the Consumer Federation of America, has been publicly opposed to price optimization for some time and has reportedly taken his concerns directly to the regulators. He clearly opposes price optimization that is used on an individual policyholder basis. His letters and articles on PO suggest he would not consider any use of this technique to be acceptable.
Consulting actuary Michael Miller has commented that the results of a rate‐filer’s competitive judgment, whether or not this judgment was partially influenced by a price optimization model, is always embedded in the filed rates. He acknowledges that U.S. insurers use the output of a price optimization model as part of their judgment regarding rates that best fit the competitive marketplace, and that this is distinguishable from the type that are used to calculate premiums for individual insureds.
Some carriers have made rate filings that use individual or household pricing techniques that are prohibited by the Ohio and Maryland bulletins, and the California Notice challenges as unfairly discriminatory. Other carriers are just exploring how a properly used price optimization tool may be useful. It behooves any carrier investigating this statistical technique to understand the differences between permitted uses and the uses that constitute unfair discrimination. The Casualty Actuarial and Statistical Task Force of the NAIC has taken on the responsibility to examine and understand pricing optimization and is expected to produce a written report on their findings later this year.