Data Call Q&A

Data Call Q&A – updated April 12, 2023

1. Should cancellation and non-renewal transactions be included in the Policy Data portion of the data call?

On the Policies tab we only want to know about policies that were actually purchased (or renewed). We want to know about the policy while it was in force, even though we don't care about the cancellation transaction itself. For example, suppose that a given policy was renewed on 2/22/2019, then a few months later in a separate transaction it was canceled. We want to have a record of the policy as renewed on 2/22/2019, but we don't need a separate record of the cancellation transaction.

2. Does “start date”, per the excerpt from the Excel spreadsheet, “Scope: Policies with start date on or after 1/1/2019 and end date on or before 12/31/2021” mean effective date and, similarly, “end date” mean expiration date. Therefore, the applicable policies and claims would be pulled for policies effective on or after 1/1/2019 through 12/31/2021.

Start date and end date mean the effective date and expiration date. It would be the effective date for both new issues and renewals during that period.

3. What companies and NAIC codes should be included in the data call?

The data call covers all companies in the insurance group issuing private passenger auto policies in the District of Columbia.

4. Will the confidentiality requirements under DC Code § 31-1404 apply to this data call including contractors engaged by the Department?

Yes, we are doing the data call under our market conduct examination authority to ensure the highest confidentiality standard we have would be applied to the review which are the protections under §31-1404.

5. Should we include private passenger auto assigned risk business in the data call?

Policy and claims data for assigned risk plans should be included but quote data (if it exists) should not be included.

6. Can insurers with a small volume of business be excluded from the data call?

No, all insurers with private passenger auto insurance business written in the District between 2019 and 2021 should be included regardless of the volume.

7. What is considered a “specialty car” that can be excluded from the data call?

Antique cars and similar vehicles that may have different uses and different insurance concerns should be excluded. If you are unsure about a particular class of vehicles, please ask or include information about the class of vehicles excluded in the narrative questions.

8. Are customized vehicles to be included in the data? A customized vehicle consists of individually adapted items built (or fit) in or upon vans, minivans, autos and pickup trucks installed or customized by dealers or vehicle owners. For example, additional air conditioning units, carpeting, furnishings, etc.

Yes

9. In Section 3 of the Qualitative Questions question 1 (about the data dictionary) includes “coverage_status” as one of elements, but we are unable to locate that variable. Can you address?

That is a typo. Previous drafts included a variable called “coverage_status” but it was removed. Please ignore the reference to this variable.

10. The scope/definitions explain that each claim should have a row per type of coverage. Taking this into consideration for the claim status is the state expecting to see the status of the claim or the coverage? It is very plausible for a particular coverage on a claim to close before other coverages do and the claim itself remains open until all coverages have been closed. The same question applies for the claim resolved date and the loss paid, is the state expecting this to be for the claim or coverage?

All claims reporting should be at the level of coverage. For instance, suppose a single car accident leads to a claim including a physical damage component and a bodily injury component. Then in the row corresponding to the physical damage component of that claim, please provide the details of just that component, not the whole claim.

11. There are a couple of fields which we do not have available to systematically pull but could discern by manually reviewing the claim if the state selected (subro reimbursed amount and fraud flag). May we present these as AOR (available on request)?

We do want to know subro_reimbursed_amount and fraud_flag for each claim. Please make your best effort to properly identify those items.

12.  ​​​​Will there be any contractual terms shared between the carriers and the DISB on how the data can be used and with whom it can be shared?

• The review is being conducted under our market conduct examination authority to provide the maximum confidentiality protection that we can. As with all market conduct examinations, the contractor and department can use the data only for the intended purpose of the examination.

13. Is data minimization possible? For example, does DISB need to know that it was Jane Doe who got the quote on x day or can we anonymize some of the data?

• Data minimization would likely remove some of the data necessary for the BISFG methodology, so data minimization to anonymize the data should not be done.

14. When a vehicle is deleted from a policy, how should that be handled on the policy template?

• Just reflect what the policy looks like afterwards.

15. Should excluded and unlicensed drivers be excluded from the template?

• Excluded and unlicensed drivers should be excluded from the template.

16. For beginning of term conditions, for number of rated drivers, would that be at renewal or as things are modified (policy id)?

• As things are modified. In general, the idea is for each transaction (e.g., renewal, modification) to capture a snapshot of key policy characteristics post-transaction and the resulting premium.

17. On the “DISB Data Call Template” spreadsheet, the “Policies – Scopes and definition” and “Claims – Scope and definitions” tabs ask for data from 1/1/2019 – 12/31/2021 under the “Scope” section. However, the “Quote – Scope and definitions” tab asks for “quotes initiated since 1/1/2019 (or earliest available date, if later than 1/1/2019) up to today”. What does “today” mean?

• Today is the date of the examination warrant – February 17, 2023.

18. The scope section of the claims tab states “All claims related to policies with start date on or after 1/1/2019 and end date on or before 12/31/2021 (i.e., all the policies in Policies dataset)”. We want to double check that this should include all the claims (since inception) that are part of the policies from the policies dataset. This would include claims that have loss date as early as 1989. Or are you only looking for claims with a loss date between the dates of 1-1-19 and 12-31-21? It’s unclear to us because there are no date parameters in the Claim Loss Date description.

• We don't want claims with a loss date of 1989. For policies that have been renewed many times, we only want claims for the policy terms with start date (meaning either initial start date or renewal date) on or after 1/1/2019 and end date on or before 12/31/2021.

19. Should the data call apply to named non-owner policies since there are no vehicles?

• No

20. Our system does not provide quoted premium amounts for every combination of driver-vehicle on a prospective policy; we can provide the total policy premium, or if preferred, the premium for each vehicle. In either case, the total premium reported for a quote (when you add up the 6 records as given in the instructions for a 2 vehicle-3 driver policy) the sum will be greater than total policy (quote) premium. Please describe how you would like the premium to be reported.

• We would prefer premium for each vehicle. On a multi-vehicle policy we can add up the vehicles' premiums to get the total policy premium without double-counting in the case of multiple drivers that you mentioned.

21. The Examination Warrant 2023-1, dated February 17, 2023, indicates that O’Neil Regulatory Consulting and Algorithmic Auditing (ORCAA) was engaged to participate in the examination ordered in the warrant as an examiner. The Commissioner specifically reserves the right in the warrant to assign examiners in addition to ORCAA. What examiner title should a company subject to the warrant use when referring to ORCAA as part of the company’s data submission?

• To this point, we have not had the need to engage other contractors under the warrant. Under the law on examinations, it indicates the “Mayor” is allowed to retain professionals and specialists as examiners, so you may refer to ORCAA as an examiner.

22. We appreciate that the report will be opining on aggregate metrics and not digging into individual companies. What controls are in place to ensure company-identifying information isn’t disseminated through mechanisms other than the final report?

• We will send out a draft report to each company with information that is not relevant to their company redacted to give the companies a chance to review any information that could be considered related to their company. This will give each company the opportunity to identify any matters that they would like us to not make public. We have done this previously on a multi-company exam and it worked well. I don’t expect much information in the report to be related to particular company information because of the nature of what the report will be addressing. I am overly cautious about not inadvertently sharing information and certainly open to other ideas.

23. We have some questions around the CPE data fields requested for this MCE:

Collision and comp, both include a limit that covers custom parts and equipment. However, additional CPE coverage can be purchased. Are you looking only for the additional CPE limit/deductible as they are essentially their own coverages, or do you want to know what the covered limit is for CPE as part of the base coll/comp coverages?

• The fields "CPE_limit" and "CPE_deductible" should only be populated if the consumer is buying additional CPE coverage, beyond what is included with any collision and/or comprehensive coverage they are buying.

24. CPE is covered under both collision and comprehensive, meaning there can potentially by 2 limits and 2 deductibles to report. However, the template only has room for 1 coverage (2 columns: one for limit, one for deductible). Is it ok if we add 2 more columns for the second coverage? If not, how would you like us to report the CPE limit/deductible between the 2 coverages?

• Presuming the included coverage is already reflected in the price of the collision and/or comprehensive coverages, we don't need to hear about it separately.

25. How would like the situation for which a coverage has multiple vehicles and parties to show on the one line of data expected?

• The vehicle from this policy that was involved in this multivehicle / multiparty claim. In other words, we're assuming this claim involved one vehicle that is part of the policy, and other vehicles that are not part of the policy. We want the former.

26. We are pulling data from a few different systems that might have different data gaps. Would you prefer that we combine data from the systems (i.e., appending data from a smaller PPA program/system to the bottom of data from a larger PPA program/system) or would you like us to split the data for each program separate submissions/tabs?

• We would like the data combined together rather than separate.

Data Template Questions

Quotes – Scope and Definitions Tab

Row	Request	Company Question/Clarification	D.C. Response
11	rowID –Unique identifier for this row	Can we eliminate this field since Excel has a built in unique rowID?   We won’t be able to properly sort/merge/etc. which can impact our validation and merge efforts, so we appreciate any consideration to removing this requirement.	No, we need a unique rowID as its own field.
45, 51		When would "Quote_firstprice" and "Quote_seriousprice" have different values?	The values would be the same if all 4 requirements of “serious” are met before an initial quote is given. To clarify, the carrier should provide both fields even though they will always have the same value for a given quote.
46	quote_serious – is the quote serious	Under what circumstances would this field be “0”?	If a quote is provided prior to all four elements of a serious quote being met.
47	uw_approval – is the quote approved by underwrting	Could you please define what is meant by "underwriting approval"? Does this mean approval made directly by an underwriter?	No, it doesn’t need to be an underwriter (human); it could be an underwriting algorithm. If an application is reviewed for placement and is not declined/rejected but rather placed into a particular company, tier, and/or classification – i.e., some insurance will be offered to this applicant, at some price – then an underwriting approval has occurred.
48, 49, 50		Under what circumstances would “decline” be the appropriate response to these questions?	Decline would be the appropriate response when the quote process is terminated by the insurance company rather than the person requesting the quote.
50	uw_tier_standardized – standardized tiers	Could you please provide definitions for super preferred, etc.?	The carrier should do this mapping using their judgment, and provide an explanation.
56	dob – the driver’s date of birth	Row 13 states 'No personally-identifying information please.' Should we redact the DAY of the birthday and leave the month and year?	Month and year is ok.
58	yrs_licensed – how many years licensed	We are able to provide the date licensed but not the years licensed. Is this acceptable? Can we calculate the difference between the rated operator's license date and the policy effective date.	This is ok. Since rated operator’s license date is the date of initial license, this is acceptable.
67	vehicle_new – is the vehicle new	Since we don't have an indicator for new vehicles, we are planning to match the vehicle model year with the year from the policy effective date to indicate which ones are new. We do not have this information for older products, should we use “0”?	Yes, this is an acceptable approach. If this information is not available, use “n/a”

 

Policies – Scope and Definitions Tab

Row	Request	Company Question/Clarification	D.C. Response
11	rowID– Unique identifier for this row	Can we eliminate this field since Excel has a built in unique rowID?   We won’t be able to properly sort/merge/etc. which can impact our validation and merge efforts, so we appreciate any consideration to removing this requirement.	No, we need a unique rowID as its own field.
12	policyID – unique identifier for the policy	For policyID - first five positions will be the same for all new, ren and mod records for one policy. The next two positions will represent the policy term. The next two will be transaction within the term. Vehicle and Driver combinations will not be unique. Does this approach fulfill the Department's request?	Yes
14	policy_termination_date – last day of coverage under the policy	This will reflect the policy term expiration date for all policyIDs within the policy term. Cancellation date will not be reflected in this field. Is this correct?	Yes
17	firstname –First name of this driver	1)  Confirm this is the first named insured only. 2)  Our system maintains the first and last name of the driver in one field.  Can this field include both first and last name in one field? We have had difficulty separating the names in the past, especially when they have more than three parts to their name. 3)  The systems that Quotes data and Claims data is able to separate the names. If one field is used for the name for policies, should we do the same for Quotes and Claims or separate them?	1) No, DISB is requesting the names of all drivers on the policy. 2) Respondent must separate the name and the first name and surname separately. For complex names, please use your best judgment to separate the names. 3) In this data call response, always use separate fields for first name and surname.
18	surname– Surname of this driver	See row 17 comment/question	[see above]
44	contract premium – total premium on the final policy	How should contract premium be represented for each policy id? Option 1 -Do you want to see the end result for the policy term (written premium)? Option 2- Or do you want to see end result as of each financial transaction (policy id)? Please note that we do not have private passenger only policies. Policies can have/are eligible to include other types of vehicles. Is the expectation that contract premium represents the full policy or should we only provide the premium associated with private passenger vehicles? Should the Claim data only be associated with private passenger vehicles?	Follow Option 2 please. In general, the idea is for each transaction (e.g. renewal, modification) to capture a snapshot of key policy characteristics post-transaction and the resulting premium. Ideally you could provide for each policy just the premium amount associated with the private passenger vehicles on the policy. If this level of detail is not available then you may provide the policy term premium. In this case please include a new field in your Policy data response: “nonPP_vehicle”. This will be a binary variable, =1 if the policy described in this row includes a type of vehicle other than a private passenger vehicle, and =0 otherwise. This will allow us to identify the policies for which policy term premium reflects a vehicle other than the private passenger vehicles depicted in the data provided. Depends on the previous question; the scope of Claim data provided should match the scope of premium reported in the Policy data response. Specifically, if you can provide just the premium amount associated with private passenger vehicles on the policy, then provide Claim data only associated with private passenger vehicles. Otherwise, provide Claim data associated with all vehicles on the policy.
46	dob – This driver’s date of birth	Row 16 states 'No personally-identifying information please.'  In order to protect PII, a standard business practice is to redact the DAY of birth. Is this acceptable?	Month and year is ok.
48	yrs_licensed –For how many years has this driver been licensed to drive?	We are able to provide the date licensed but not the years licensed. Is this acceptable? Can we calculate the difference between the rated operator's license date and the policy effective date.	This is ok. Since rated operator’s license date is the date of initial license, this is acceptable.
53	lapse_in_coverage – documented lapse incoverage	We do not have this information at a driver level, named insured only and it would be for prior coverage with a different insurer. Should we 'use a zero on all rows?	Please include this information for named insured and leave blank for others.
57	vehicle_new – is the vehicle new	Since we don't have an indicator for new vehicles, we are planning to match the vehicle model year with the year from the policy effective date to indicate which ones are new. We do not have this information for older products, should we use “0”?	Yes, this is an acceptable approach. If this information is not available, use “n/a”

 

Claims – Scope and Definitions Tab

Row	Request	Company Question/Clarification	D.C. Response
14	policyID – policy id for the claim	The policyID that will be will potentially match to multiple policy records, is this correct?	Yes
22	loss_paid – Total dollar amount of claim paid as of 12/31/2022	Is the loss_paid data to includes ALAE?	No
22 & 23	loss_paid– Total dollar amount of claim paid as of 12/31/2022 & loss_reserve– Total dollar amount reserved for future losses on this claim as of 12/31/2022	Do we provide the total dollar amount for each coverage or at the entire claim level? See example on page 4 below. Example 1: loss amounts at each coverage which will add up to the entire claim level.   Example 2: the total amount of the losses at the claim level will populate the same for each coverage.	Please provide the information as shown in Example 1: Coverage level.
24 & 25	ALAE_paid – Total dollar amount of ALAE paid as of 12/31/2022 allocated to this claim & ALAE_reserve – Total dollar amount reserved as of 12/31/2022 for future ALAE on this claim	Do we provide the total dollar amount for eac coverage or at the entire claim level? See example on page 4 below. Example 1: loss amounts at each coverage which will add up to the entire claim level. Example 2: the total amount of the losses at the claim level will populate the same for each coverage.	Please provide the information as shown in Example 1: Coverage level.
22 & 23 & 24 & 25	loss_paid & Loss_reserve & ALAE_paid & ALAE_reserve	For loss_paid, loss_reserve, alae_paid, and alae_reserve data. Are these supposed to be gross or net of subrogation and salvage?	These values should be gross.