On September 27, 2023, the Colorado Division of Insurance (“DOI”) released a draft proposed regulation on Quantitative Testing of External Consumer Data and Information Sources, Algorithms, and Predictive Models Used for Life Insurance Underwriting for Unfairly Discriminatory Outcomes (“Draft Testing Regulation”). Comments on the Draft Testing Regulation can be submitted to the DOI until October 26.
The Draft Testing Regulation is part of the broader DOI effort to develop regulations under a bill enacted by the Colorado legislature in 2021 (SB 21-169), which addresses unfair discrimination on the basis of race, color, national or ethnic origin, religion, sex, sexual orientation, disability, gender identity, or gender expression in the use by insurers of external consumer data and information sources (“ECDIS”) or algorithms or predictive models that use ECDIS. The Draft Testing Regulation follows on the heels of the DOI’s September 21 adoption of Regulation 10-1-1 Governance and Risk Management Framework Requirements for Life Insurers' Use of External Consumer Data and Information Sources, Algorithms, and Predictive Models.
The DOI has scheduled a stakeholder meeting to discuss the Draft Testing Regulation on October 19. In addition, the DOI recently announced a stakeholder meeting regarding Unfair Discrimination in Insurance Practices, focusing on underwriting practices in private passenger auto insurance, which will be held on November 2.
Testing Focused on Identifying Unfairly Discriminatory Decisions based on Race or Ethnicity
Under the Draft Testing Regulation, life insurers authorized in Colorado that use ECDIS or algorithms or predictive models that use ECDIS would be required to perform annual quantitative testing to evaluate whether the insurers’ use of ECDIS or algorithms or predictive models that use ECDIS are unfairly discriminatory based on the race or ethnicity of proposed insureds. Such testing shall be used to evaluate potential unfair discrimination with respect to the following two areas:
- whether Hispanic, Black, and Asian/Pacific Islander applicants are declined for coverage at a statistically different rate relative to White applicants; and
- whether Hispanic, Black, and Asian/Pacific Islander insureds are charged a statistically different premium rate per $1,000 of the face amount of their policies, relative to White insureds.
If the first-level testing described above indicates a statistical difference of 5% or more with respect to either of the areas identified above, insurers would be required to conduct second-level variable testing of the ECDIS or algorithms or predictive models that use ECDIS to identify the specific variable(s) contributing to the difference.
Estimation of the race or ethnicity of applicants and/or insureds must be done by utilizing Bayesian Improved First Name Surname Geocoding (BIFSG)—the statistical methodology developed by the RAND Corporation for estimating race and ethnicity—and the applicant’s and/or insured’s name and geolocation information included in such person’s application for life insurance. The following categories are to be used for these purposes: Hispanic, Black, Asian/Pacific Islander, and White.
Under the Draft Testing Regulation, life insurers would be required to provide an annual report to the DOI starting on April 1, 2024 with information regarding the application approval decision testing, the premium rate testing and any second-level variable testing. The report would be based on data up to December 31 of the prior year.
Our team will continue to monitor such regulatory developments in Colorado and other states for future developments.