Why Policy-Period Assignment Matters
An exposure record has a time interval, but an analysis may also need to know which policy term supplied the coverage during that interval. A renewal creates a new term even when the insured and coverage continue without a break. A replacement can create a new policy record partway through an earlier term. Assigning exposure to policy periods means connecting each portion of covered time to the contract term in force for it.
This assignment is not the same as deciding which calendar experience period contains the exposure. Here the unit of organization is the policy term: the term beginning at inception or renewal and ending at expiration, cancellation, or another valid termination date. Correct assignment preserves the connection between exposure and the policy that provided it, which is essential when calculating results by term or joining exposure to policy records.
Using half-open intervals makes adjacent terms unambiguous. If one term ends at the instant the next begins, the shared boundary belongs to neither both terms nor neither term: it is excluded from the first interval and included in the second. The convention describes instants; under a daily exposure basis, it means a day starting on a term’s effective date is assigned to that term.
Partitioning an Exposure Interval
An exposure interval may be contained in one policy term or cross one or more term boundaries. Partition it at each effective start or end date that changes the policy providing the coverage. Then calculate exposure for each resulting segment and attach that amount to its policy term. If the exposure basis is constant during a segment, the calculation can be made directly from the segment’s duration.
The rule is a duration allocation, not a rule for changing the underlying exposure basis. If the number of insured vehicles or the insured amount changes during a term, the segments must reflect those changes as well as policy-period boundaries. For this tutorial, assume a single, unchanged exposure basis within each segment.
Here \(d_{jk}\) is the number of covered days in segment \(k\) assigned to term \(j\). The sum is needed when a term has multiple separated covered segments. If the exposure basis has a scale factor, such as three insured units, multiply the time-based exposure by that factor.
A Direct Derivation and Assignment Procedure
The calculation follows from the definition of an annual exposure unit. Under the stated convention, one unit represents 365 covered days for one exposure basis. A segment of \(d\) covered days therefore represents the same fraction of a unit as \(d\) is of 365:
If the segment has \(n\) identical exposure units, each covered day contributes \(n/365\) annual units, so the segment exposure is \(nd/365\). This argument depends on exposure accruing uniformly by day and on the selected annual denominator. It does not apply unchanged to a basis that varies by day or to a convention that allocates exposure by another measure.
Confirm that the exposure interval refers to the same insured item or other exposure basis and the same coverage being assigned.
Use coverage-effective start and end dates, including valid cancellation, renewal, and replacement dates, rather than relying only on record creation or processing dates.
Split the exposure interval wherever the term providing coverage changes. Treat a boundary as belonging to the new term under the half-open convention.
Apply the stated exposure basis and day-count convention, then record the amount against the policy-term identifier for the coverage in force.
Compare the sum assigned across terms with the total covered interval. Resolve gaps or overlaps from authoritative coverage records before aggregation.
Worked Example: Allocating a Partial-Term Exposure
A hypothetical one-vehicle policy provides coverage from January 1, 2025, through January 1, 2026. A constant exposure basis applies from April 1, 2025, to October 1, 2025. Using half-open dates and a 365-day annual exposure unit, calculate the exposure and assign it to the policy term.
Method. Intersect the vehicle’s covered interval with the policy term. The entire interval lies inside the term, so no split is needed. The term identifier is the policy’s 2025–2026 term, not merely the policyholder’s name or an earlier term’s identifier.
Count covered days. The interval is \([{\rm April\ 1},{\rm October\ 1})\). Its days are \(30+31+30+31+31+30=183\): April 1–30, May, June, July, August, and September. October 1 itself is excluded.
Calculate. The assigned exposure is \(183/365=0.5013698\ldots\), or 0.5014 annual exposure units when rounded to four decimal places. The whole amount is assigned to the 2025–2026 term.
Check. The interval is shorter than a full 365-day year, and \(0.5014\times365\) is approximately 183 days. A common mistake is to include October 1 and count 184 days; that would conflict with the stated half-open interval.
Renewals and Replacement Policies
A renewal starts a new policy term, even if the coverage continues seamlessly. The exposure before the renewal belongs to the expiring term; exposure from the renewal effective date belongs to the renewed term. The continuity of coverage does not make the term identifiers interchangeable.
A replacement can similarly split exposure before the original expiration date. The replacement’s effective date matters: if the old term ends when the replacement starts, allocate the adjacent segments separately. If the records overlap or leave a gap, do not resolve the issue by automatically summing or discarding days. First determine whether the overlap represents two genuinely in-force contracts, a duplicate record, or an incorrect effective date.
Worked Example: Separating Exposure at a Replacement
A hypothetical one-vehicle policy begins January 1, 2023, and is effective through January 1, 2024, represented as \([{\rm Jan\ 1,\ 2023},{\rm Jan\ 1,\ 2024})\). A replacement policy starts July 1, 2023, and coverage under the first policy ends at that same instant. Calculate exposure by term using 365 days per annual unit.
Method. The two terms are adjacent: the first covers \([{\rm Jan\ 1},{\rm Jul\ 1})\), and the replacement covers \([{\rm Jul\ 1},{\rm Jan\ 1,\ 2024})\). July 1 is assigned only to the replacement term.
Count the first segment. January through June 2023 contains \(31+28+31+30+31+30=181\) days. The first term receives \(181/365=0.4958904\ldots\), or 0.4959 units to four decimal places.
Count the replacement segment. July through December contains \(31+31+30+31+30+31=184\) days. The replacement receives \(184/365=0.5041096\ldots\), or 0.5041 units to four decimal places.
Check. The two segments contain \(181+184=365\) days, and their unrounded exposures sum to \(365/365=1\) annual unit. A common mistake is to assign the full 2023–2024 term to each policy record, doubling the exposure; the correct assignment follows the effective coverage boundary.
When Records Overlap
An overlap in stored dates is a warning to investigate, not proof that the insured had overlapping coverage. For example, an old policy may appear to end after a replacement begins because one record retained a scheduled expiration date even though a cancellation endorsement changed the actual end date. Conversely, two separate policies can genuinely cover the same risk during the same period. The appropriate assignment depends on which coverage was actually in force and on the exposure unit being measured.
If a correction establishes that the old term ended before the replacement began, revise the effective intervals before calculating exposure. If two distinct policies genuinely provide coverage concurrently, keep their policy-term assignments separate; do not force the records into a single term merely to avoid an overlap. If the target is one physical item’s time exposed to any coverage at all, overlapping days should not be counted twice in that combined measure. If the target is exposure by contract, the same time may appear under both contracts. State which measure is being calculated.
Worked Example: Resolving an Apparent Overlap
A hypothetical equipment schedule has an old policy record with dates \([{\rm March\ 1},{\rm May\ 1})\) and a replacement record with dates \([{\rm April\ 25},{\rm June\ 1})\). An authoritative cancellation endorsement confirms that the old policy actually ended on April 25, when the replacement took effect. For one item, calculate total exposure by term using 365 days per annual unit.
Method. Use the effective dates confirmed by the endorsement, not the stale end date on the old record. The corrected intervals are \([{\rm March\ 1},{\rm April\ 25})\) for the old term and \([{\rm April\ 25},{\rm June\ 1})\) for the replacement term.
Count the old-term segment. March 1 through March 31 contributes 31 days, and April 1 through April 24 contributes 24 days. Thus the old term receives 55 days, or \(55/365=0.1506849\ldots\), which is 0.1507 units to four decimal places.
Count the replacement segment. April 25–30 contributes 6 days, and May 1–31 contributes 31 days. The replacement receives 37 days, or \(37/365=0.1013698\ldots\), which is 0.1014 units to four decimal places.
Check. The corrected total is \(55+37=92\) days, equal to March 1 through June 1 under the half-open convention. The raw records appear to overlap from April 25 through April 30, but the endorsed termination date removes that overlap. Counting both raw intervals would incorrectly count those six days twice.
Common Mistakes and Practical Checks
- Using the wrong date. A data-entry date or scheduled expiration date may not be the effective coverage date. Use the date that reflects coverage in force.
- Assigning all time to the first policy record. A renewal or replacement creates a new term; split the exposure at its effective boundary.
- Counting the boundary twice. With half-open intervals, the old term excludes its end date and the new term includes its start date.
- Adding overlapping records without review. Determine whether the overlap is a record error, a true concurrent contract, or a consequence of the chosen exposure measure.
- Assuming every annual term has exactly 365 elapsed days. Use the specified day-count convention and handle leap days consistently. Do not silently change the denominator in the middle of a calculation.
- Failing to state the exposure basis. Time alone is not enough when the number of insured units or insured amount changes. Segment at those changes and calculate with the appropriate basis.
A useful final check is to compare the assigned segments with the underlying covered timeline. For a single continuously covered item with adjacent terms, assigned days should equal covered days. If they do not, look for a gap, duplicated dates, an omitted cancellation, or inconsistent date boundaries. For multiple contracts or exposure bases, reconcile each measure separately rather than assuming that one total must fit every view.
Check Your Understanding
For each problem, use half-open date intervals and a 365-day annual exposure unit unless the question states otherwise. Round final exposures to four decimal places.
- A hypothetical policy term is \([{\rm Feb\ 1,\ 2025},{\rm Feb\ 1,\ 2026})\). One insured unit is exposed from May 1 to August 1, 2025. How many days and annual exposure units belong to this term? (Answer: 92 days; 0.2521 units.)
- A one-item policy term covers \([{\rm Jan\ 1,\ 2022},{\rm Jan\ 1,\ 2023})\). A renewal starts January 1, 2023, with no gap. A covered interval runs from December 1, 2022, to February 1, 2023. Assign days and exposure to each term. (Answer: 31 days and 0.0849 units to the 2022 term; 31 days and 0.0849 units to the renewal.)
- A hypothetical vehicle’s old term ends September 15, 2024, and a replacement starts on September 15. The covered interval is \([{\rm September\ 1},{\rm October\ 1})\). How many days and units are assigned to each term? (Answer: 14 days and 0.0384 units to the old term; 16 days and 0.0438 units to the replacement.)
- Two records for one insured item both show coverage on November 10–12. A correction confirms the old policy ended November 10 and the replacement began November 10. For a combined measure of days the item was covered by any policy, how many days should the overlap contribute? Explain. (Answer: The overlap contributes no extra days; count the covered interval once, because the corrected terms are adjacent and the item is measured once.)
- A corrected interval for one item is \([{\rm March\ 1},{\rm April\ 1})\) under one term, followed immediately by \([{\rm April\ 1},{\rm May\ 1})\) under a renewal. Calculate total annual exposure across both terms and state why April 1 is not counted twice. (Answer: 61 days total, or 0.1671 units; April 1 belongs to the renewal only under the half-open convention.)