The Picture Becomes Clearer as Colorado Holds Second Stakeholder Meeting
The second stakeholder meeting focused on:
- Providing insights on the testing required by the CO law.
- Filtering through different ways of developing definitions for the key terms "external consumer data and information sources" and "traditional underwriting practices."
The CO DOI also announced that it will develop a data call to assess the industry's current practices.
1. Testing
The CO law requires that insurers test their use of external consumer data, information sources, algorithms, and predictive models to ensure that the use results in a balanced picture so consumers are not harmed by unfairly discriminatory results. The CO DOI noted it is not suggesting that issues exist. Rather, it expects that insurers test their use of external consumer data, information sources, algorithms, and predictive models before any snapshots of actual consumers are taken so any issues can be exposed and resolved before the image is transferred to the negative.
The CO DOI expects that all data points used will be included in the required testing even though the CO law excludes "traditional underwriting factors" from testing. The CO DOI cast a light on the fact that the exclusion does not apply when those factors are "included in an algorithm that also uses external consumer data and information sources." The COI DOI posited that this is occurring.
The CO DOI invited ORCAA to present on the Bayesian Improved Surname and Geocoding (BISG) inference methodology that may allow insurers to test for racial bias without having to collect racial information from applicants. The method uses census data to infer a person's race and ethnicity just from his or her first name, last name, and address, which are all data points that insurers already collect from applicants.
2. Definitions
The COI DOI noted that a major goal for the stakeholder meetings is to expose the need to refine the definitions of the key terms "external consumer data and information sources" and "traditional underwriting practices."
The CO law currently defines external consumer data and information sources as:
A data or an information source that is used by an insurer to supplement traditional underwriting or other insurance practices or to establish lifestyle indicators that are used in insurance practices. "External consumer data and information source" includes credit scores, social media habits, locations, purchasing habits, home ownership, educational attainment, occupation, licensures, civil judgments, and court records.
The CO law does not currently define traditional underwriting practices.
At the meeting, it was discussed whether in adjusting the focus of these definitions, the CO DOI should superimpose the prior work of:
- The NAIC Accelerated Underwriting Working Group, which drafted an educational report to put a frame around the discussion regarding accelerated underwriting;
- The New York State Department of Financial Services, which issued Circular Letter No. 1 regarding insurers’ use of external consumer data and information sources that distinguishes between medical and non-medical consumer data; and
- The NAIC Market Conduct Annual Statement Blanks Working Group, which distinguishes data directly provided by the consumer in an application from data otherwise obtained by the insurer from external sources.
Commenters suggested that the CO DOI provide examples of data or data sources that clearly fit into one category or another, and examples of data or data sources that fall in a gray area.
The CO DOI recognized that any rules implementing the CO law could not be effective before January 1, 2023. However, the CO DOI stressed that since Colorado is the first state to take action in this area, it wants to have "lots of conversations" so it can "get this right."
The next stakeholder meeting is expected to occur at the end of May or early June. We will continue to monitor the activities of the CO DOI and the rulemaking process for the CO law.
