Colorado law limits the use of data in setting insurance rates
The law will require that insurance companies test their data systems, algorithms and models to identify unfairness.
New legislation in Colorado, which the governor and state insurance commissioner have signed, will require insurance companies to demonstrate that their use of external data and complicated algorithms do not discriminate on the basis of certain classes and will help guarantee that citizens do not overpay for coverage or become unfairly targeted by insurance companies because of who they intrinsically are.
The law aims to hamper practices that disproportionately harm communities of color or people on the basis of sex, sexual orientation or gender identity,
The law, SB 169, identifies the classes that “must be protected from discrimination in the marketing, underwriting, pricing, utilization management, and reimbursement methodologies, and claims management in the transaction of insurance.” The law instructs the Insurance Commissioner to take input from companies and consumers and adopt rules to ensure that insurers’ use of data and models does not result in harmful discrimination. The rules must also allow companies to mitigate any biases found in their algorithms.
It is an industry practice for auto insurers to use non-driving characteristics like credit history, gender, education, and marital status to calculate premiums. Using these characteristics has disproportionately impacted people of color and other protected classes. These practices also perpetuate systematic bias.
Although this bill does not prohibit the use of credit history in setting insurance premiums, it prevents insurers from using any external consumer data, information, algorithms, or predictive models that disproportionately harm members of any protected class. In order to use factors such as credit history in the future, insurers will need to ensure those factors do not disproportionately harm people of color and take steps to mitigate unfair practices.
A study by the Consumer Federation of America (CFA) using premium quotes for a 35-year-old with a perfect driving record from ten of the largest insurers in Colorado for every ZIP code in the state found that with excellent credit scores, a driver can expect to pay an annual average premium of $592.11. In comparison, a driver with fair credit can expect to pay an annual average premium of $785.67, and a driver with a poor credit score can see the average premium increase by 72% to $1,019 for the same coverage even with a perfect driving record.
“This bill holds insurers accountable for systemic biases built into company practices by requiring that they test their data systems, algorithms, and models to identify unfairness that has historically been ignored and accepted,” Michael DeLong, CFA’s insurance advocate, said in a release. “The bill creates a process to ensure that insurance markets are more equitable. With its enactment, Colorado has an opportunity to be a leader in the national effort to reduce systemic bias in insurance and other financial services.”
Related: