On January 18, 2019, the New York Department of Financial Services (NYDFS) issued a circular letter to all insurers authorized to write life insurance in New York State setting out the Department’s views concerning the use of external data sources in underwriting for life insurance.¹

The circular letter acknowledges that the use of external data sources has the potential to benefit both insurers and consumers by simplifying and expediting life insurance sales and underwriting processes, and can result in more accurate underwriting and pricing. The letter nonetheless expresses “significant concerns” about the potential for negative impact on consumers, insurers and the New York life insurance marketplace due to the variances in the accuracy and reliability of external data sources and the fact that many external data sources are not subject to regulatory oversight and consumer protections. Notably excluded from the scope of the circular letter, however, are external data sources of information directly related to the medical condition of an applicant.

The circular letter focuses on two areas of “immediate concern” -- unlawful discrimination and transparency to consumers. NYDFS then sets out its views as to specific actions insurers are obligated to take under existing law to avoid illegal discrimination and meet customer disclosure and transparency requirements when using external data sources, algorithms and predictive models.

Unlawful discrimination. The letter states that NYDFS has determined that insurers’ use of external data sources in underwriting has the strong potential to mask forms of discrimination prohibited by various state and federal anti-discrimination laws.² In particular, it notes many of these sources use geographical data (including community-level mortality, addiction and smoking data), homeownership data, credit information, educational attainment, licensure, civil judgments and court records, “which have the potential to reflect disguised and illegal race-based underwriting” in violation of New York law. It also notes how other models and algorithms purport to make predictions about a consumer’s health status based on other sources of information that “may either lack a sufficient rationale or actuarial basis and may also have a strong potential to have a disparate impact on the protected classes identified in New York and federal law.”³ The circular letter defines a number of “principles” insurers should use as guidance in using external data sources in underwriting:

First, the insurer should determine that the external tools or data sources do not in any way collect or utilize prohibited criteria. In doing so, an insurer may not simply rely on a vendor’s claims of non-discrimination or the proprietary nature of third-party processes to justify failing to independently determine compliance with anti-discrimination laws.

Second, insurers should not use external data sources, algorithms or predictive models in underwriting or rating unless they establish that they are not unfairly discriminatory, after considering the following questions:

• Is it supported by generally accepted actuarial principles or actual or reasonably anticipated experience that justifies different results for similarly situated applicants?
• Is there a valid explanation or rationale for the differential treatment of similarly situated applicants that is derived from external data sources or information?

The letter declares there must be valid rationale or explanation supporting the differential treatment of otherwise like risks even if statistical data is interpreted to support an underwriting or rating guideline. This is “particularly important where there is no demonstrable causal link between the classification and increased mortality and where an underwriting or rating guideline has a disparate impact on protected classes.”

Consumer disclosure and transparency. The circular letter states NYIL 4224(a)(2) requires insurers to notify insureds or potential insureds of the right to receive the specific reasons for a declination, limitation, rate differential or other adverse underwriting decision. According to the circular letter, an adverse underwriting decision would include “the inability of an applicant to utilize an expedited, accelerated or algorithmic underwriting process in lieu of a traditional medical underwriting.” Additionally, where an insurer is using external data sources or predictive models, the reasons provided to the insured or potential insured for any declination, limitation, rate differential or adverse underwriting decision must include details about the specific source of the information upon which the insurer based its adverse underwriting decision.

The letter states insurers must also provide notice to and obtain consent from consumers to access external data, where required by law or regulation. NYDFS states that the failure to adequately disclose the material elements of an accelerated or algorithmic underwriting process, and the external data sources upon which it relies, may constitute an unfair trade practice prohibited under the New York Insurance Law, even where the insurer has relied on a vendor’s proprietary algorithm.

Observations. Circular letters are written communications to entities regulated by the Department that call attention to statutory and/or regulatory requirements and positions being taken by the Department regarding their interpretation and enforcement.⁴ Unlike regulations, which set rules to implement broad policies embodied in legislation, they do not establish new legal requirements and do not have the force of law, but instead remind regulated entities of the requirements of new laws or regulations or set out the Department’s interpretation of the requirements of existing laws and regulations.⁵

This circular letter arises out of the Department’s inquiries on accelerated underwriting, algorithmic underwriting and the use of big data following a data call NYDFS issued to licensed insurers in June 2017. During a recent Q&A session with industry representatives, NYDFS’s senior staff indicated they spoke with a variety of stakeholders before issuing the guidance, including insurance company representatives and consumer advocates. However, it did not undergo the formal notice and comment processes that are required for promulgating new regulations.

NYDFS’s focus on causality and the specifics of data sources will likely pose a challenge for many life insurers seeking to use non-traditional data sources as part of underwriting. Sound actuarial principles require actuaries to make assumptions about a person’s life expectancy based on probabilities. Actuaries use the law of large numbers and complex statistical tools, including models constructed from data from multiple sources, to draw statistical correlations and inferences and to develop reliable mortality estimates. Actuarial science does not deal in causalities, but rather in correlations, to draw assumptions on life expectancies.

Similarly, life insurers will likely also be challenged by the burden placed on insurers that they “determine” that third parties do not utilize prohibited criteria in either their data collection or algorithmic processes, with NYDFS explicitly stating that insurers may not rely on a vendor’s representations in doing so. It is unclear how insurers are to undertake due diligence on their vendors’ data sources, data aggregation practices and algorithmic methods in order to make these determinations, especially as national debates ensue on transparency in algorithms and the regulation of data brokers.⁶ NYDFS’s stance arguably shifts to insurers responsibilities that other policy makers, including the NAIC and the FTC, have been reluctant to impose absent explicit legislative authority.

Finally, NYDFS’s treatment of the inability of an applicant to utilize an expedited, accelerated or algorithmic underwriting process as an adverse underwriting decision will likely also require that insurers think carefully about how they implement new underwriting processes that incorporate non-traditional data.

The positions taken by NYDFS in this circular letter offer one of the first looks at how regulators will approach the use of non-traditional data sources in underwriting, and one that is likely to generate considerable discussion and concern on how to appropriately and practicably address the issues raised by NYDFS. Insurers seeking to use external data sources in underwriting are likely to continue facing novel regulatory concerns. The letter notes NYDFS does not believe that the concerns it raised provide an all-inclusive list of potential issues that arise from the use of external data sources and is not intended to suggest that an insurer’s due diligence in assessing external data sources should be limited to these two concerns.

____

¹ “External data” is defined for purposes of the circular letter to include “any data or information sources not directly related to the medical condition of the applicant, used - in whole or in part - to supplement traditional medical underwriting, as a proxy for traditional medical underwriting or to establish ‘lifestyle indicators’ that may contribute to an underwriting assessment of an applicant for life insurance.” The letter excludes from the meaning of “external data” known consumer reporting agencies, such as MIB Group, Inc. member information exchange service, motor vehicle reports and criminal history searches.

² The circular letter notes legal prohibitions on the use of race, color, creed, national origin, status as a victim of domestic violence, past lawful travel or sexual orientation in underwriting, and limitations on differential coverage because of physical or mental disability, impairment or disease, prior history of disability or disease except where permitted by law or regulation and based on sound actuarial principles or related to actual or reasonably anticipated experience.

³ The examples cited are models and algorithms that use retail purchase history, social media, internet or mobile activity, geographic location tracking, the condition or type of an applicant’s electronic devises and any systems or applications operating thereon, or as on how the consumer appear in a photograph.

⁴See W.B. Dunham, New Appleman New York Insurance Law 2.05[1] (2018)

⁵Id.

⁶ “Tim Cook Calls for “Data Broker Clearinghouse” in Push for Privacy Rules,” Bloomberg News (Jan.17, 2019) (Apple CEO calls on FTC to “establish a data-broker clearinghouse, requiring all data brokers to register, enabling consumers to track the transactions that have bundled and sold their data from place to place, and giving users the power to delete their data on demand”); “Toward Algorithmic Transparency and Accountability,” Communications of the American Association of Computing Machinery (Sept. 2017).

[View source.]