Why Report Period Assignment Matters
A claim may be reported days, months, or even years after the event that caused the loss. A report-period view organizes claims by when the insurer first learned of them. It therefore answers a different question from a view organized by when the loss occurred. The distinction matters whenever a claim is reported late: the claim belongs to a later report period, even though its occurrence date remains earlier.
For consistent analysis, the reporting-date rule must be clear. In the examples below, the report date is the date the insurer first receives notice of the claim. A claim is counted once, using its unique claim identifier and its first-notice date. Later claim transactions or supplemental information do not create additional claims.
A report period can be a year, quarter, month, or another defined interval. Its boundaries must be set in advance and applied consistently. For example, if quarterly periods are used, the first quarter is January 1 through March 31, inclusive. The key assignment rule does not change with the interval: use the first-report date, not the occurrence date, payment date, or date a later transaction is processed.
The Assignment Rule
Suppose \(r_j\) is the first-report date for claim \(j\), and \(P_k\) is a specified report period. Assign claim \(j\) to \(P_k\) exactly when \(r_j\) falls within that period. Define \(I(r_j\in P_k)\) to equal 1 when the condition is true and 0 otherwise. The number of claims assigned to the period is:
Here, \(m\) is the number of distinct claims in the data set. The formula counts each distinct claim once. If the periods do not overlap and cover every report date in the data set, each claim contributes to exactly one period total.
This assignment rule classifies claims by reporting time; it does not move the underlying event date. If a claim arises from an event in one year but is first reported in the next, the report-period assignment uses the later year. The accident date stays as recorded. Report-period counts can consequently differ from counts grouped by occurrence or accident date.
Worked Example: Assigning Claims to Annual and Quarterly Report Periods
A hypothetical insurer has five distinct claims. Its data definition uses the date the insurer first receives notice. Assign the claims to report years and, where useful, report quarters. Count each claim once.
| Claim | Accident date | First-notice date |
|---|---|---|
| C101 | December 29, 2023 | January 8, 2024 |
| C102 | January 31, 2024 | February 2, 2024 |
| C103 | November 28, 2024 | November 30, 2024 |
| C104 | December 31, 2024 | January 3, 2025 |
| C105 | September 14, 2024 | December 31, 2024 |
Method: first-report-date assignment. For each claim, use its first-notice date, then place it in the period containing that date.
C101 was first reported January 8, 2024, so it belongs to 2024 Q1. C102 was first reported February 2, 2024, so it also belongs to 2024 Q1. C103 was first reported November 30, 2024, so it belongs to 2024 Q4. C105 was first reported December 31, 2024, so it also belongs to 2024 Q4. C104 was first reported January 3, 2025, so it belongs to 2025 Q1.
The report-year totals are \(4\) claims in 2024 and \(1\) claim in 2025. The quarter totals shown are \(2\) in 2024 Q1, \(2\) in 2024 Q4, and \(1\) in 2025 Q1; the other quarters have \(0\) claims in this data set.
Check: The annual totals sum to \(4+1=5\), matching the five distinct claim identifiers. C101 and C104 have accident dates in an earlier year than their report year, but that does not change their report-period assignments.
Common mistake avoided: Grouping by accident date would put C101 in 2023 and C104 in 2024. That answers an occurrence-date question, not the report-period question. The records here are deterministic; no interest, mortality, or probability-distribution assumption is needed.
Late Reports and Unchanged Accident Dates
A late report shifts when a claim appears in report-period counts, not when the loss occurred. For a fixed set of claims, adding the first-notice dates can reveal that some claims belong to report periods after their accident dates. If a later data extraction includes claims that had not yet been reported at the earlier extraction date, the earlier report-period totals do not retroactively include those claims: their first report occurred later.
It is useful to keep two facts separate for each claim: the occurrence or accident date describes when the loss event happened, while the first-report date determines its report-period assignment. A late-reported claim preserves its occurrence date and is assigned to the period containing its first-report date. This does not imply that its report date is a substitute for the accident date; they answer different questions.
Worked Example: Late Reports Across a Year Boundary
A hypothetical book contains four distinct claims, all arising from accidents in 2023. The insurer later obtains the following first-notice dates. Determine the report-year counts and explain what changes when the two late reports arrive.
| Claim | Accident date | First-notice date |
|---|---|---|
| R1 | December 18, 2023 | December 26, 2023 |
| R2 | December 29, 2023 | December 30, 2023 |
| R3 | November 14, 2023 | January 6, 2024 |
| R4 | October 2, 2023 | March 12, 2024 |
Method: assign using each claim's first-notice year. R1 and R2 were reported in 2023, so the 2023 report count is \(2\). R3 and R4 were first reported in 2024, so the 2024 report count is also \(2\). The total is \(2+2=4\) distinct claims.
At the end of 2023, only R1 and R2 had been reported, so the report count for that year was \(2\). When R3 and R4 were reported in 2024, they increased the 2024 report count by \(2\); they did not become 2023 reports merely because their accidents occurred in 2023. All four accident dates remain in 2023.
Check: The final report-period counts sum to \(4\), the number of claims. The late reports change the period in which those two claims first appear, while leaving every listed accident date unchanged.
Common mistake avoided: Placing R3 and R4 in the 2023 report period because their accidents occurred in 2023 confuses event timing with reporting timing. No interest, mortality, or distribution assumptions are relevant to this record-classification calculation.
First Notice, Data Entry, and Duplicate Activity
Operational databases may contain multiple dates associated with one claim. The insurer might receive notice on one date, enter the information into a system later, and receive additional documents after that. The selected report-date definition must match the purpose and available data. In the examples here, notice received by the insurer controls. If the source records only system-entry dates, those dates may be used only under a clearly stated and consistently applied definition; they should not silently be treated as first-notice dates.
Claim identifiers also matter. A claim can have many transactions, such as a payment, a reserve adjustment, or a supplemental report. Those activities are not automatically new claims. For report-period claim counts, use one record per distinct claim and the date of its first report. If a claim is reopened or supplemented, a separate analysis may be needed to define what is being counted, but the original claim should not be counted again as a newly reported claim merely because it has new activity.
Worked Example: Applying the Receipt-Date Rule at Year-End
A hypothetical insurer defines a claim's report date as the date it first receives notice, rather than the date the notice is entered into its system. The records below cover three distinct claims. One claim also has later supplemental activity. Assign each claim to a report year and compare the result with the incorrect use of entry dates.
| Claim | First notice received | Initial system entry | Later activity |
|---|---|---|---|
| P71 | December 31, 2025 | January 2, 2026 | Supplement received January 10, 2026 |
| P72 | January 1, 2026 | January 1, 2026 | None |
| P73 | December 30, 2025 | January 4, 2026 | None |
Step 1: Apply the stated date definition. P71 was first reported on December 31, 2025, so assign it to 2025. P72 was first reported on January 1, 2026, so assign it to 2026. P73 was first reported on December 30, 2025, so assign it to 2025.
Step 2: Count distinct claims. The correct report counts are \(2\) claims in 2025 and \(1\) claim in 2026. The supplement for P71 does not create another claim and does not change its first-report date.
Step 3: Identify the effect of the tempting alternative. If initial system-entry dates were incorrectly used, all three claims would appear in 2026: the entries were January 2, January 1, and January 4. That would produce a mistaken count of \(0\) in 2025 and \(3\) in 2026, contrary to the stated receipt-date definition.
Check: Under the correct rule, \(2+1=3\), which matches the three distinct claims. The alternative also totals three, but assigns them to the wrong years. The calculation uses administrative dates only; no interest, mortality, or probability-distribution assumptions apply.
Common mistake avoided: Treating entry date as report date without checking the data definition shifts claims across a period boundary. Counting P71's supplemental report as a new claim would also incorrectly inflate the claim count.
Common Errors to Avoid
- Using the accident date. That groups claims by when losses occurred, not when they were first reported.
- Using the payment date. Payment activity may occur well after the first report and is not the assignment basis defined here.
- Using system-entry dates automatically. First receipt and system entry can occur in different periods. Follow the stated report-date definition.
- Counting claim transactions as claims. Identify distinct claims first, then assign each claim once by its first-report date.
- Changing period boundaries between records. Apply the same specified annual, quarterly, or other intervals to every claim.
- Moving the accident date when a report is late. Late reporting changes report-period classification, not the date of the underlying event.
Report-period assignment is a data-classification step. It is aligned with the CAS Exam 5 (Basic Ratemaking and Estimated Claim Liabilities) syllabus because period definitions are essential to organizing insurance experience consistently. The practical discipline is straightforward: define the interval, define the first-report date, use one record per distinct claim, and keep the event date separate.
Check Your Understanding
For each problem, use the stated report-date rule and count distinct claims only.
- A hypothetical insurer uses first-notice dates and annual periods. Claim A was first reported December 30, 2024; Claim B on January 1, 2025; and Claim C on December 31, 2024. What are the report-year counts? (Answer: 2024: 2 claims; 2025: 1 claim.)
- Three distinct claims have first-notice dates of May 4, 2025, June 18, 2025, and February 3, 2026. Using calendar quarters, how many belong to 2025 Q2 and 2026 Q1? (Answer: 2025 Q2: 2 claims; 2026 Q1: 1 claim.)
- A claim resulted from an accident on November 29, 2024, but the insurer first received notice on February 6, 2025. To which report year is it assigned, and what happens to its accident date? (Answer: It is assigned to report year 2025; its accident date remains November 29, 2024.)
- A claim was first reported December 28, 2025, entered into the system January 3, 2026, and supplemented January 15, 2026. The stated rule uses first receipt and counts distinct claims. To which report year does it belong, and how many claims does it contribute? (Answer: 2025; 1 claim.)
- A hypothetical data set has five distinct claims. Two were first reported in 2023 and three in 2024; all five have accident dates in 2023. What are the report-year counts, and do the later first reports change the accident-year information? (Answer: 2023: 2 claims; 2024: 3 claims; no, the five accident dates remain in 2023.)