American Academy of Actuaries: P/C Equity & Fairness Committee Comments to DISB

WASHINGTON, Jan. 21 (TNStalk) -- The American Academy of Actuaries issued the following letter on Jan. 20, 2023:

* * *

To: Philip Barlow, Associate Commissioner for Insurance, District of Columbia, Department of Insurance, Securities and Banking (DISB), 1050 First Street NE, 801, Washington, DC 20002

Re: Request for Comment - Draft Data Call on Unintentional Bias in Automobile Insurance

Dear Associate Commissioner Barlow:

On behalf of the Property/ Casualty Committee on Equity and Fairness of the American Academy of Actuaries ("the committee"),/1 I appreciate the opportunity to submit these comments regarding DISB's Request for Comment--Draft Data Call on Unintentional Bias in Automobile Insurance. As previously indicated in the Racial Equity Task Force (now Property/Casualty Committee on Equity and Fairness) verbal and written testimony, the committee supports DISB efforts to study the issue of unintentional bias in personal auto insurance. Clearly, any such study will be data-intensive, and the exposure of your draft data call for comments will facilitate moving forward with your study in an efficient manner. The committee's comments on the data call, as follows, are offered as considerations for you as you move forward. The committee has kept comments at a high level and would welcome the further opportunity to assist and advance this important work.

Comments on Complexities of Insurance Data

Like most data, insurance data is "messy," and analyzing insurance data is complex. Actuaries are guided by Actuarial Standard of Practice (ASOP) No. 23, Data Quality. ASOPs provide guidance on the techniques, applications, procedures, and methods that reflect appropriate actuarial practices in the United States. DISB may find that this ASOP provides useful information as it moves forward with its data call.

Generally, private passenger automobile insurance premiums are based on characteristics about each driver and vehicle on the insurance policy as well as characteristics about the policy itself. Information about all drivers and vehicles need to be considered in order to understand the premium or the relationship between the premium and the losses. For example, the current data call requests years of experience for the first named insured. For policies that include multiple family members, the number of years of experience of the named insured will likely be biased since this will often reflect the experience of a parent and not capture the lower years of experience for the children on the policy.

A policy's premium is sometimes changed during the term of the policy, for example when a new vehicle or a new driver is added to the policy. Because of the number of data elements that are typically considered to understand the premium (easily in excess of 35 rating variables, on average), the volume of data is large for each policy. The committee suggests therefore that DISB collect information so that there is at least one row of data for each policy term. Over a three-year period, this will result in at least three rows of data per policy, more if transactions that change the premium mid-term are considered or if the policy term is six months. Loss data is also difficult because claims are often paid long after the occurrence of the accident and the actual amount of the loss is sometimes not known for years.

A further complication to insurance data collection is that insurance companies do not all rate or quote policies the same and do not always use the same variables or rating structures. Some companies will create tiers for preferred business and some companies may not have tiers but will give additional discounts for business they consider preferred. Also, even for companies with tiers, the definition of preferred business for one company is likely not the same as the definition of preferred for another company. Because companies use different variables and different tiering plans, aggregating data across companies or comparing companies can be difficult.

Other Comments for Consideration

Given the complexities of automobile insurance ratemaking and the related data, the DISB may wish to consider limiting the initial data call and analysis to one company or a small subset of companies. This could help to refine a subsequent data call for a broader number of companies. Further, the committee encourages the DISB to consider providing a designated person to answer questions from insurers regarding the data call--this may help insurers understand each requested item in the data call and may help the DISB better understand the data that it will receive.

As the DISB produces initial results from its analysis, it may also be worthwhile to solicit peer review and feedback from experts on automobile ratemaking. The committee would be interested in being of assistance if the DISB chooses to go through such a process.

~ ~ ~

The committee continues to support DISB's efforts in this work. On behalf of the committee, I hope that these comments and suggestions will prove helpful to you as you work through refining the data call. As indicated, the committee would welcome the opportunity to assist and advance this important work.

If you have any questions about this letter or seek additional information from the Academy, contact Rich Gibson, Senior Casualty Fellow, at [email protected].

Sincerely,

Lauren Cavanaugh, MAAA, FCAS

Chairperson, Property/Casualty Committee on Equity and Fairness

American Academy of Actuaries

Footnote:

1/ The American Academy of Actuaries is a 19,500-member professional association whose mission is to serve the public and the U.S. actuarial profession. For more than 50 years, the Academy has assisted public policymakers on all levels by providing leadership, objective expertise, and actuarial advice on risk and financial security issues. The Academy also sets qualification, practice, and professionalism standards for actuaries in the United States.

* * *

Original text here: https://www.actuary.org/sites/default/files/2023-01/Comment_Letter_to_DISB_Unintentiontal_Bias.pdf