Why Exposure Has More Than One Time View
An exposure record describes covered risk over time, but the year attached to that exposure depends on the question being asked. A policy-year view groups exposure by the year the policy became effective. A calendar-year view groups exposure by the calendar dates on which coverage was actually in force. The totals can differ for an individual year even when both views describe the same portfolio.
This distinction matters when preparing and checking experience summaries. A policy effective late in one year can contribute exposure to the next calendar year. Conversely, a calendar-year total can include portions of policies that began in several different effective years. The key is to identify the grouping rule before adding exposure.
For the examples, exposure is measured in annual exposure units: one covered risk in force for 12 months contributes 1.00 unit. Assume each month contributes one-twelfth of a unit, coverage is continuous during each stated month, and policies have the same exposure basis. These conventions make the time allocation transparent; a real data specification must use its stated exposure basis and date conventions.
A policy-year cohort is not the same thing as the exposure earned during that calendar year. The label “2024 policy year” identifies policies by effective year; the label “2024 calendar year” identifies covered time occurring between January 1 and December 31, 2024. Neither label, by itself, changes how much coverage a policy provided.
Allocating Policy-Term Exposure to Calendar Years
For a policy with annual exposure unit \(w_p\), let \(m_{p,y}\) be the number of covered months falling in calendar year \(y\). Under the monthly convention, its calendar-year exposure is \(w_p m_{p,y}/12\). For a standard annual policy with \(w_p=1\), this is simply the fraction of its 12 covered months in that calendar year.
Here \(e_{p,y}\) is policy \(p\)'s exposure assigned to calendar year \(y\). In the policy-year view, the same policy's complete-term exposure is assigned to the cohort matching its effective year. If a policy covers months in two calendar years, its calendar-year pieces add back to its full-term exposure, provided all covered months are included.
Worked Comparison
Worked Example: Three Annual Policies Crossing Calendar Boundaries
A hypothetical insurer has three annual policies, each covering one unit of risk for 12 months. Policy A is effective July 1, 2023; Policy B is effective January 1, 2024; and Policy C is effective July 1, 2024. Use the monthly convention to calculate the exposure by effective-year cohort and by calendar year. Check that the portfolio totals reconcile.
Method. First assign each complete term to its effective-year cohort. Then split the same covered months among the calendar years in which they occur.
Policy-year view. Policy A contributes 1.00 unit to the 2023 effective-year cohort. Policies B and C each contribute 1.00 unit to the 2024 cohort. Thus, the cohort totals are 1.00 for 2023 and \(1.00+1.00=2.00\) for 2024.
Calendar-year view. Policy A covers July through December 2023, or 6 months, and January through June 2024, or 6 months. Its allocations are \(6/12=0.50\) in 2023 and \(6/12=0.50\) in 2024. Policy B contributes 12 months, or 1.00 unit, to 2024. Policy C contributes 6 months, or 0.50 unit, to 2024, and 6 months, or 0.50 unit, to 2025.
Adding by calendar year gives 2023: \(0.50\); 2024: \(0.50+1.00+0.50=2.00\); and 2025: \(0.50\). The calendar-year grand total is \(0.50+2.00+0.50=3.00\) units. The policy-year grand total is \(1.00+2.00=3.00\) units, so the two views reconcile.
Check and common mistake. Each of the three policies contributes exactly 1.00 unit across its calendar-year pieces, so the total must be 3.00. A common error is to put all of Policy A's exposure in 2023 because that is its effective year. That would incorrectly give 2023 1.00 unit and omit A's 0.50 unit in 2024. The policy-year assignment is appropriate for the cohort view, not for the calendar-year allocation.
Why the Grand Totals Reconcile
The reconciliation follows from splitting each policy's covered time into non-overlapping calendar-year pieces. Suppose policy \(p\) has total exposure \(w_p\) over 12 covered months. If all those months are included, the counts \(m_{p,y}\) across calendar years sum to 12. Therefore:
Summing that identity over all policies shows that total calendar-year exposure equals total policy-year exposure, as long as both summaries contain the same policies and covered time. This is a bookkeeping identity, not a claim that the yearly subtotals are equal. If a calendar-year summary stops before a policy term ends, the omitted months will not appear in that summary, so the totals over the displayed years need not reconcile to the full-term policy-year total.
Worked Example: Comparing Effective-Year Cohorts with Calendar Totals
A hypothetical portfolio contains four annual policies, each contributing 1.00 unit over its complete term. Their effective dates are April 1, 2024; October 1, 2024; January 1, 2025; and July 1, 2025. Find the policy-year totals for 2024 and 2025 and the calendar-year totals for 2024, 2025, and 2026.
Method. Group the full annual terms by effective year for the policy-year result. For the calendar-year result, count the covered months of each term in each calendar year.
The first policy contributes 1.00 to the 2024 cohort. The second also contributes 1.00 to the 2024 cohort, so the 2024 policy-year total is \(1.00+1.00=2.00\). The third and fourth policies each have effective year 2025, so the 2025 policy-year total is \(1.00+1.00=2.00\).
For calendar year 2024, the April-effective policy contributes 9 months, or \(9/12=0.75\); the October-effective policy contributes 3 months, or \(3/12=0.25\). The total for 2024 is \(0.75+0.25=1.00\).
For calendar year 2025, the April-effective policy contributes its remaining 3 months, or 0.25, and the October-effective policy contributes its remaining 9 months, or 0.75. The January-effective policy contributes 12 months, or 1.00, and the July-effective policy contributes 6 months, or 0.50. The total for 2025 is \(0.25+0.75+1.00+0.50=2.50\).
For calendar year 2026, the July-effective policy contributes its remaining 6 months, or 0.50. The total is therefore 0.50. The calendar-year grand total is \(1.00+2.50+0.50=4.00\), matching the four complete policy terms, each with 1.00 unit.
Check and common mistake. The 2025 calendar total is 2.50, not the 2025 policy-year total of 2.00. Calendar 2025 includes the final 0.25 and 0.75 units from policies effective in 2024, while the July 2025 policy contributes only 0.50 during calendar 2025. Treating every policy effective in 2025 as a full calendar-year exposure would overstate the calendar total.
When a Policy Does Not Earn a Full Term
The same classification distinction applies when a policy does not remain in force for its entire scheduled term. Exposure should reflect covered time, not automatically the full annual term. The policy-year view still groups the earned exposure by effective year; the calendar-year view splits that earned exposure according to when the covered months occurred. The treatment of cancellations and reinstatements must follow the applicable exposure records and conventions.
Worked Example: A Canceled Policy Spanning Two Calendar Years
A hypothetical annual policy is effective November 1, 2024, and coverage ends after April 30, 2025. It provided six covered months in total. Under the monthly convention, determine its policy-year exposure and its calendar-year allocations.
Method. Count only the six months with coverage, then classify them in each view. The November and December months fall in calendar 2024; January through April fall in calendar 2025.
The policy's total earned exposure is \(6/12=0.50\) unit. Since its effective year is 2024, the 2024 policy-year cohort receives 0.50 unit. The 2024 calendar allocation is \(2/12=0.1667\) unit, and the 2025 calendar allocation is \(4/12=0.3333\) unit. Before rounding, the calendar pieces sum to \(2/12+4/12=6/12=0.50\), matching the policy-year amount. Rounded to four decimal places, \(0.1667+0.3333=0.5000\).
Check and common mistake. The term's full scheduled duration is 12 months, but only 6 months were covered. Assigning 1.00 unit to the 2024 cohort would ignore the cancellation. Assigning the entire 0.50 unit to calendar 2024 would also be wrong because four covered months occurred in 2025.
Interpreting and Checking Exposure Summaries
A useful comparison is to lay out policy-year and calendar-year totals separately rather than force them into a single table column. Each answers a different question: policy-year grouping connects exposure to the effective-year cohort, while calendar-year grouping measures the exposure occurring during a specified calendar year.
| View | Grouping rule | Typical implication |
|---|---|---|
| Policy year | Assign exposure to the policy effective year | A complete annual term is grouped together, even when it crosses December 31 |
| Calendar year | Assign exposure to the dates coverage is in force | A policy can contribute to two calendar years, or more for a longer term |
A sound reconciliation check is to sum all relevant policy-year cohorts and all relevant calendar-year totals over the same covered dates. If the totals differ, first check whether the date ranges are complete and whether both summaries use the same policy population and exposure units. Other possible causes include omitted coverage months, duplicated records, or inconsistent treatment of policy changes. A matching grand total is a useful control, but it does not prove that each policy was assigned to the correct year.
One frequent exam trap is to compare a single effective-year cohort directly with a single calendar-year total and expect equality. That expectation is valid only if the covered dates and policy grouping happen to make them equal. In general, the meaningful comparison is between grand totals across equivalent populations and complete time ranges; individual yearly subtotals are differently composed.
Check Your Understanding
Use the monthly convention and assume each stated policy represents one annual exposure unit unless the question says otherwise.
- A policy effective May 1, 2023, remains in force for its full 12 months. State its policy-year cohort and its calendar-year allocations. (Answer: 2023 policy-year exposure = 1.00; calendar 2023 = 8/12 = 0.6667 and calendar 2024 = 4/12 = 0.3333.)
- Two full annual policies are effective March 1, 2024, and September 1, 2024. Find the 2024 policy-year total and the calendar-year totals for 2024 and 2025. (Answer: 2024 policy-year total = 2.00; calendar 2024 = 10/12 + 4/12 = 1.1667; calendar 2025 = 2/12 + 8/12 = 0.8333.)
- A full annual policy effective November 1, 2025, contributes how much to the 2025 cohort and to each calendar year it touches? (Answer: 2025 policy-year cohort = 1.00; calendar 2025 = 2/12 = 0.1667; calendar 2026 = 10/12 = 0.8333.)
- A policy effective August 1, 2024, ends after five covered months, on December 31, 2024. Find its policy-year exposure and calendar-year exposure. (Answer: 2024 policy-year exposure = 5/12 = 0.4167; calendar 2024 exposure = 0.4167.)
- Three complete annual policies are effective February 1, 2025, June 1, 2025, and October 1, 2025. Find the 2025 policy-year total and calendar-year totals for 2025 and 2026; verify the grand total. (Answer: 2025 policy-year total = 3.00; calendar 2025 = 11/12 + 7/12 + 3/12 = 1.7500; calendar 2026 = 1/12 + 5/12 + 9/12 = 1.2500; both calendar totals sum to 3.00.)