Introduction
A policy record can contain several dates that look equally important: the date coverage begins, the date it ends, the date the policy was issued, and the date someone reviewed the record. These dates answer different questions. To determine whether a particular event occurred during the contract’s coverage period, begin with the policy’s effective date and expiration date, then apply the contract’s stated times and time zone.
This tutorial focuses on interpreting those dates and distinguishing them from transaction and valuation dates. It does not calculate written or earned exposure. All examples are deterministic comparisons of hypothetical contract dates; no interest rate, mortality assumption, or probability distribution is used.
The Coverage Interval and Its Boundaries
A useful standard representation treats the effective date-time as included and the expiration date-time as excluded. In mathematical notation, the coverage interval is \([E,X)\), where \(E\) is the effective date-time and \(X\) is the expiration date-time. This convention means that a time \(t\) is within the interval exactly when \(E \leq t < X\).
The half-open interval avoids counting the shared boundary twice when one policy period ends exactly when a succeeding period begins. At the exact expiration time, the expiring policy no longer covers. If another policy begins at that same time, its interval includes the boundary. The contract’s actual language controls; if it specifies a different convention, use that wording rather than assuming this representation.
Some policies specify a time such as 12:01 a.m. local time on both the effective and expiration dates. That detail matters. “Effective October 1” does not necessarily mean coverage starts at midnight on October 1, and “expires October 1” does not necessarily mean coverage continues through the whole day. Also check which local time zone applies. A database timestamp stored in a standard time zone may need conversion before it can be compared with the contract’s local time.
Why the Interval Convention Works
The coverage test follows directly from the meaning of a start and an end. A time before \(E\) occurs before coverage begins, so it is outside the period. A time at or after \(E\) has passed the start boundary. But a time at \(X\) has reached the point at which coverage ends, so the period includes only times strictly earlier than \(X\).
Thus the two requirements must hold together: the time must not precede the start, and it must precede the end. If consecutive periods use the same boundary \(B\), the first has \(t < B\), while the next has \(t \geq B\). The boundary belongs to the second interval, not both.
This is a convention for representing consecutive coverage intervals, not a substitute for reading the contract. If the policies use different times or time zones at a boundary, their actual date-times may not match, and a gap or overlap may result.
Contract Dates Versus Other Dates
| Date type | Question it answers | What it does not establish by itself |
|---|---|---|
| Effective date-time | When does the contract’s coverage begin? | When the policy was entered into an administrative system |
| Expiration date-time | When does the contract’s coverage end? | That coverage continues through the entire calendar date shown |
| Transaction date | When did an administrative event occur? | That the event changed the contract’s coverage boundaries |
| Valuation date | As of what date is a record or measurement being considered? | That the policy was in force on that date |
A transaction might be entered before a policy becomes effective, or a record might be valued after its expiration. Neither fact is inherently inconsistent. The relevant question is what the transaction did under the contract and what date the measurement is intended to represent. A record’s presence in a system, by itself, does not prove that coverage existed at every time shown in that record.
Worked Example: Testing Events at Both Policy Boundaries
A hypothetical commercial property policy states that coverage begins at 12:01 a.m. local time on April 8, 2026, and expires at 12:01 a.m. local time on April 8, 2027. Determine whether each event time is within the policy period: (a) April 8, 2026, at 12:00 a.m.; (b) April 8, 2026, at 12:01 a.m.; (c) April 8, 2027, at 12:00 a.m.; and (d) April 8, 2027, at 12:01 a.m.
Method. Treat the stated date-times as \(E\) and \(X\), and apply \(E \leq t < X\). All times below are in the policy’s stated local time.
(a) The event is one minute before \(E\). Therefore \(t < E\), so it is not within coverage.
(b) This event is exactly at \(E\). Since the start boundary is included, \(E \leq t\), and it is earlier than \(X\). It is within coverage.
(c) This event is one minute before \(X\). It is at or after \(E\) and strictly earlier than \(X\), so it is within coverage.
(d) This event is exactly at \(X\). The required condition \(t < X\) fails, so it is not within coverage.
Check. The covered interval runs from the April 8, 2026, start boundary up to—but not including—the April 8, 2027, expiration boundary. The common mistake is to treat each named calendar date as covered in full; the correction is to compare the complete date-time with both stated boundaries.
Transaction Dates and Valuation Dates
A transaction date describes an administrative event rather than automatically describing the coverage interval. For example, an insurer might issue a policy on one date with a later effective date. The issue date answers when the policy was processed or issued; the effective date answers when coverage begins under the contract. Unless the contract provides otherwise, the issue date does not move the effective boundary.
A valuation date is an “as of” date for a calculation, data extract, or review. It helps specify the timing of the information being considered. A valuation date that falls inside a policy period does not prove that a particular exposure or claim is covered: the relevant coverage terms and event facts still matter. Likewise, a policy may appear in records valued after its expiration without being in force on the valuation date.
Worked Example: Separating Issue, Effective, and Valuation Dates
A hypothetical equipment policy is issued on September 22, 2026. Its contract states that coverage begins at 12:01 a.m. local time on October 1, 2026, and expires at 12:01 a.m. local time on October 1, 2027. An analyst reviews the policy record as of December 31, 2026. A reported incident occurred on September 29, 2026, at 3:00 p.m. local time. Identify what each date means and whether the incident time is inside the stated coverage interval.
Method. First identify the contract boundaries; then compare the incident time with those boundaries. Keep the issue and valuation dates separate from that comparison.
The issue date is September 22, 2026. It records when the policy was issued. It is not the stated start of coverage.
The effective date-time is October 1, 2026, at 12:01 a.m. local time, and the expiration date-time is October 1, 2027, at 12:01 a.m. local time. These define the stated interval \([E,X)\).
The incident occurred on September 29, 2026, at 3:00 p.m., which is earlier than October 1, 2026, at 12:01 a.m. Therefore \(t < E\), and the incident is outside the stated coverage interval.
The valuation date is December 31, 2026. It identifies when the analyst is considering the record; it does not change the policy’s effective date or make the earlier incident fall within the interval.
Check. The incident precedes the contract’s effective boundary, even though the policy was issued before the incident and the later valuation date falls during the stated policy period. The common mistake is to infer coverage from the issue date or the valuation date; the correction is to test the incident against the contract’s effective and expiration date-times.
Date-Only Records and Local Time
A data field that contains only a date can hide a material detail. If a contract says “effective October 1” but the system stores only “October 1,” the time of day and possibly the time zone are missing from that field. Do not silently replace missing information with midnight unless the contract or a documented data convention supports that interpretation.
When comparing records, make the assumptions explicit. Check whether the dates are local or standardized, whether the contract states a time, and whether the recorded times use the same time zone. Daylight-saving changes can also affect the relationship between a local boundary and a stored timestamp. When those details are unavailable, flag the ambiguity rather than asserting a precise coverage result at the boundary.
Worked Example: Resolving a Date-Only Expiration Field
A hypothetical policy record contains an effective date of January 31, 2028, and an expiration date of January 31, 2029, but no times. A separate contract extract states that both boundaries occur at 12:01 a.m. local time. A loss is recorded as occurring on January 31, 2029, at 12:00 a.m. local time. Determine whether the date-only field alone settles the question, then use the contract extract to test the event time.
Method. Separate the information in the data field from the more precise contract wording, then apply the interval test using the precise date-times.
The date-only field by itself does not specify whether the expiration boundary is at the beginning or end of January 31. It is insufficient to decide the boundary-time question without an applicable convention or more complete contract information.
The contract extract supplies that information. The expiration time is January 31, 2029, at 12:01 a.m. The loss time, 12:00 a.m., is one minute earlier. It is after the effective boundary and before the expiration boundary, so it is within the stated interval.
Check. The result relies on the contract extract’s time, not on an assumption that the date-only expiration field means the policy ends at midnight. The common mistake is to treat an incomplete date field as a complete date-time; the correction is to obtain and apply the contract’s stated boundary.
Common Mistakes and Practical Checks
- Using the issue date as the start date. Read the stated effective date-time; an administrative issue date does not automatically define when coverage starts.
- Assuming the expiration date is covered through the whole day. Compare the event time with the actual expiration time. At the expiration boundary, the interval \([E,X)\) has ended.
- Treating a valuation date as proof of coverage. A valuation date describes when a record is being considered; it does not establish that the contract was in force at an event time.
- Ignoring time zone or boundary-time details. A comparison is reliable only when the event and policy boundaries are expressed on a consistent time basis.
- Filling in missing times without support. If a record supplies dates only, consult the contract or a documented convention before deciding a boundary case.
These distinctions support accurate interpretation of policy records before any exposure measure is selected or calculated. The essential habit is to identify the contract’s two boundaries first, express them as date-times on a consistent basis, and compare the event time with both boundaries. A transaction or valuation date can provide useful context, but it cannot replace that test.
Check Your Understanding
For each hypothetical scenario, identify the relevant boundary or date type and apply the stated interval convention. Assume dates and times use the policy’s local time unless indicated otherwise.
- A policy begins at 12:01 a.m. on June 12, 2027, and expires at 12:01 a.m. on June 12, 2028. An event occurs at 12:00 a.m. on June 12, 2028. Is it within the policy period? (Answer: Yes; it is one minute before expiration.)
- A policy is issued on February 3, 2027, but its effective date is February 10, 2027, at 12:01 a.m. A loss occurs February 8 at 9:00 p.m. Which date controls the interval start, and is the loss time within that interval? (Answer: The effective date controls; no, the loss is before the effective boundary.)
- A policy ends at 12:01 a.m. on November 5, 2028, and a succeeding policy begins at that same date and time. Under the half-open convention, which policy includes an event occurring exactly at the shared boundary? (Answer: The succeeding policy only.)
- A record valued as of May 31, 2027, lists a policy expiration date of May 1, 2027, with no time. What can be concluded from the valuation date alone about whether coverage existed on May 31? (Answer: Nothing definitive; the valuation date does not establish coverage, and the actual contract boundaries and any relevant changes must be checked.)
- A policy’s date-only effective field says March 4, 2029. The contract specifies 12:01 a.m. local time as the effective time. An event occurs at 12:00 a.m. that day. Is the event within the policy period, assuming the expiration is later? (Answer: No; it occurs one minute before the effective boundary.)