Insurance companies are putting more money and resources behind big data and analytics every day with the hope that greater insight will help their companies grow strong books of business. But carriers must tread carefully and ensure that the data and models they use pass the scrutiny of the public and regulators and are viewed as fair and relevant to the insurance product.
That’s one way to interpret the message sent by 18 state insurance departments over the past several months through bans on the practice of “price optimization” in those states. At issue is the use of certain data points in setting rates at renewal time. The Consumer Federation of America, among others, contend that insurance companies are using certain kinds of data, not related to policyholders’ actual risk profile, to identify the absolute highest rate a customer can be charged before they will switch carriers.
The upshot, according to J. Robert Hunter, director of insurance for the Consumer Federation of America, is that a customer’s rate could go up not based on their actual risk, but based on their appetite or ability to shop for insurance.
“Price elasticity of demand is being applied to each individual person,” Hunter says. “You start with two people who are equal risks, but now they have different rates.”
As the CFA won victories in several states, insurers began to worry that the definition of price optimization and the practices that were being taken off the table would run into other things as well, and the industry pushed back looking for clarification.
“We were concerned about that [the bulletins] were kind of sloppy in terms of what they were talking about and identifying the concern and the fix, and that carried over a couple states,” says Alex Hageli, director of personal lines policy for the Property Casualty Insurers Association of America (PCI).
(
Last year, the National Association of Insurance Commissioners convened hearings and commissioned a whitepaper from its Casualty Actuarial and Statistical Task Force to attempt to bring some uniformity to the process. The whitepaper identified four data points as “inconsistent with statutory requirements that rates shall not be … unfairly discriminatory’”:
- Price elasticity of demand.
- Propensity to shop for insurance.
- Retention adjustment at an individual level.
- A policyholder’s propensity to ask questions or file complaints
The NAIC also issued a draft bulletin that has been used by the states who have banned the practice since the whitepaper was released in November 2015. Hageli says that the NAIC brought “much needed uniformity to terminology and regulatory outcomes.”
Impact on Big Data Overall
The Center for Economic Justice, a partner group of the CFA, has termed this practice as “big data run amok.” Despite this battle, insurers aren’t slowing down their efforts to refine processes using data, but the precedent is set for scrutiny of new approaches to ratemaking.
Insurers’ appetite for more and better tools to quantify risk means that they are drawn to any sort of technology that promises that outcome, says Ben Seessel, an attorney with Carlton Fields.
“Vendors are coming up with new technologies, and insurance companies are eager to get a competitive edge,” he says. “There’s a recognition that big data is here to stay, but the insurers need to prepare with caution -- the regulators are looking at it and could act quickly.”
Many observers liken the fight over price optimization to the fight over credit-based insurance scoring, which has been called the first major use of big data in insurance. Even today, opinions on the credibility of credit scores range from Seessel’s view that “it’s been validated and much more widely accepted” to Hunter’s contention that “they still don’t know what they’re measuring.”
However, not all insurance big data is equal, Hunter says. For example, with an appropriate amount of transparency, data such as that collected from usage-based insurance programs is perfectly OK with his group.
“We like the idea of moving away from things like credit scoring and into risk-based things,” he says. “We like the idea of collecting data that would help refine rates so people who get in more accidents pay more than people who don’t. If it’s tied directly to the risk, we can support it, as long as it’s transparent.”
That’s likely to be where the battle lines are drawn for insurers as big data proliferates throughout the industry. Hageli says that companies he’s talked to are still more than willing to innovate around big data.
“Insurers are indeed looking to maximize and utilize big data to make their process more efficient,” he says. “It seemed to me that this [price optimization] was an issue outside that.”
