Occurrence Definitions and Hours Clauses

Why Occurrence Wording Matters

A catastrophe can generate many individual claims, but a catastrophe excess-of-loss treaty usually responds to an occurrence or event total, not to each claim in isolation. The contract wording determines which claims belong to that total. A different grouping can change whether the attachment is reached, how much of a layer is used, and how many occurrence limits are available.

An hours clause is one way a contract sets a time boundary around losses arising from a catastrophe. It does not, by itself, decide every question about what caused a loss or whether that loss is covered. The event definition and the hours clause must be read together: the definition identifies the relevant event or cause, while the time wording may restrict the losses that can be grouped for that event.

Definition: An occurrence is the event or defined group of related losses to which the contract applies its per-occurrence terms. An hours clause specifies a time period, commonly a stated number of consecutive hours, within which losses attributed to a defined catastrophe may be grouped as one occurrence. The exact wording controls; there is no universal rule that every loss within a time window must be grouped or that every loss with a common cause must be grouped.

For a time window of $h$ hours beginning at time $s$, use the interval $[s,s+h)$: a loss at exactly $s$ is included, while a loss at exactly $s+h$ is not. This convention makes boundary decisions explicit. A contract may instead define its endpoints or the start of the period differently, so use the wording supplied in the problem.

Group Losses Before Applying the Layer

For each occurrence, first add the amounts of eligible claims. Only then apply the per-occurrence excess-of-loss terms. Let $Z_j$ be the aggregate loss assigned to occurrence $j$, $A$ the attachment amount, and $L$ the layer limit. This tutorial assumes the limit is stated as a dollar limit, in addition to the attachment.

R_j=\min\bigl((Z_j-A)^+,L\bigr),\qquad (x)^+=\max(x,0).

Thus, a layer attaching at $500,000 with a $1,000,000 limit pays at most $1,000,000 for one occurrence. Equivalently, its exhaustion point is $1,500,000: losses above that point do not increase the recovery for that occurrence. If the contract instead states the layer as “$1,000,000 excess of $500,000,” the standard meaning is a $1,000,000 limit attaching at $500,000, not a $1,000,000 exhaustion point.

The reason the layer formula works is that the reinsurer responds only to loss above the attachment, so the amount before the limit is $(Z_j-A)^+$. If that amount exceeds the limit, the contract caps it at $L$. Taking the smaller of those two amounts gives the payment in every case: zero below attachment, the excess above attachment within the layer, and the full limit once the layer is exhausted.

R_j= \begin{cases} 0, & Z_j\le A,\\ Z_j-A, & A<Z_j<A+L,\\ L, & Z_j\ge A+L. \end{cases}

When there are several occurrences and no additional annual cap or other contract constraint applies, total recovery is the sum of the occurrence recoveries. A per-occurrence limit applies separately to each occurrence; it is not automatically one limit for the entire catastrophe year.

A Reliable Method for Applying an Hours Clause

Read the event definition.
Identify the catastrophe or causal event the wording recognizes. Do not assume that closeness in time alone makes unrelated events one occurrence.

Set the permitted time window.
Record its duration, start-time rule, and endpoint convention. Treat hours as elapsed time, not as calendar dates, unless the contract says otherwise.

Assign claims to occurrences.
Use both the event definition and the time rule. Follow any wording on election, overlapping windows, and losses outside the selected period.

Aggregate and calculate.
Add the eligible claims in each occurrence, then apply the attachment and limit to each occurrence total.

A clause may allow an insurer or cedent to elect when the hours period begins, and wordings can differ on how many periods may be selected for a continuing catastrophe. Do not assume an election right, or assume that every out-of-window claim becomes a second occurrence. Where a question supplies the election or out-of-window rule, apply it as written. Where it does not, state the assumption needed for the calculation.

Worked Example: Grouping Claims in a 72-Hour Window

A hypothetical catastrophe treaty defines one occurrence as losses arising from the same storm within a single 72-consecutive-hour period. For this example, the period begins at the first reported loss, and any later loss outside it is a separate occurrence. Claims are $700,000 at hour 0, $900,000 at hour 30, and $600,000 at hour 70. The per-occurrence layer attaches at $1,000,000 and has a $1,000,000 limit. What is the recovery? These are deterministic claim amounts; no interest, mortality, or probability distribution is assumed.

Method. Apply the specified time window first, aggregate the claims in each occurrence, and then apply the layer formula.

The window is $[0,72)$ hours. All three claim times fall within it because $0\le 0<72$, $0\le 30<72$, and $0\le 70<72$. The occurrence total is

Z_1=\$700{,}000+\$900{,}000+\$600{,}000=\$2{,}200{,}000.

The amount above the $1,000,000 attachment is $2,200,000 − $1,000,000 = $1,200,000. Applying the $1,000,000 limit gives

R_1=\min(\$1{,}200{,}000,\$1{,}000{,}000)=\$1{,}000{,}000.

Check. The recovery does not exceed the $1,000,000 per-occurrence limit, and the occurrence total exceeds the $2,000,000 exhaustion point. The layer is therefore exhausted, as calculated.

Common mistake avoided: Applying the attachment separately to the three claims would produce no recovery, because each claim is below $1,000,000. That calculation ignores the occurrence aggregation required by the stated wording.

Alternative Windows Can Change Recoveries

A longer hours period can combine more losses into one occurrence, but that does not guarantee a larger total recovery. Grouping may help a group of claims reach the attachment; it may also combine losses into one occurrence subject to one limit instead of allowing separate occurrence limits. The direction of the effect depends on the claim amounts, the attachment, the limit, and the contract’s rules about selecting periods.

Worked Example: Comparing 48-Hour and 72-Hour Periods

A hypothetical treaty uses the same storm definition under either of two alternative wordings. In both cases the period starts at hour 0, and a claim outside the period is a separate occurrence. Claims are $600,000 at hour 0, $700,000 at hour 35, and $900,000 at hour 65. The layer attaches at $500,000 and has a $700,000 limit per occurrence. Compare total recovery with a 48-hour period and a 72-hour period. No interest, mortality, or stochastic loss model is used.

48-hour wording. The window is $[0,48)$, so the claims at hours 0 and 35 form the first occurrence:

Z_1=\$600{,}000+\$700{,}000=\$1{,}300{,}000.

Its loss above the attachment is $1,300,000 − $500,000 = $800,000, capped at $700,000. The hour-65 claim is a separate occurrence of $900,000, which produces $900,000 − $500,000 = $400,000 of recovery. Therefore,

R_{\text{48}}=\$700{,}000+\$400{,}000=\$1{,}100{,}000.

72-hour wording. All three claims are in $[0,72)$, so the single occurrence total is $600,000 + $700,000 + $900,000 = $2,200,000. Its recovery is

R_{\text{72}}=\min(\$2{,}200{,}000-\$500{,}000,\$700{,}000)=\$700{,}000.

Check and interpretation. The 48-hour wording produces $1,100,000, which is $400,000 more than the 72-hour wording. Under the shorter window, the last claim generates a second recovery above the attachment. Under the longer window, all claims share one per-occurrence limit, which is exhausted.

Common mistake avoided: It is incorrect to compare only the aggregate losses or to apply one $700,000 limit to the whole treaty year. The calculation must follow the occurrences created by each wording, with the stated limit applied to each one.

Event Identity Is Not Just a Clock Test

A time window is not a substitute for the contract’s causal definition. Two unrelated storms occurring within the same 72 hours may remain separate occurrences if the wording defines an occurrence by each storm. Conversely, a contract may define a particular catastrophe and its related losses together, subject to the stated hours limit. The claim dates alone cannot settle that interpretation.

Worked Example: Two Defined Storms During One Time Period

A hypothetical treaty defines each distinct storm as a separate occurrence, even if two storms happen within one 72-hour period. The first storm produces claims of $800,000 and $700,000; the second produces a claim of $1,200,000. The storms occur 20 hours apart. The per-occurrence layer attaches at $500,000 and has a $1,000,000 limit. Calculate recovery. All amounts are fixed hypothetical losses; no interest, mortality, or probability distribution is assumed.

Method. Apply the explicit event definition before considering the clock. Since the claims arise from two distinct storms, calculate two occurrence totals, not one total for the 72-hour span.

For the first storm, $Z_1=\$800{,}000+\$700{,}000=\$1{,}500{,}000$. Its loss above attachment is $1,500,000 − $500,000 = $1,000,000, so

R_1=\min(\$1{,}000{,}000,\$1{,}000{,}000)=\$1{,}000{,}000.

For the second storm, $Z_2=\$1{,}200{,}000$. Its loss above attachment is $1,200,000 − $500,000 = $700,000, so

R_2=\min(\$700{,}000,\$1{,}000{,}000)=\$700{,}000.

Total recovery is $1,000,000 + $700,000 = $1,700,000. Check: neither occurrence recovery exceeds the $1,000,000 limit, and each occurrence independently exceeds the $500,000 attachment.

Common mistake avoided: Pooling the storms because they occurred within 72 hours would produce one $2,700,000 total and only a $1,000,000 recovery under the layer. That result contradicts the stipulated definition of each distinct storm as a separate occurrence.

Exam Traps and Contract Checks

Confusing elapsed hours with calendar days. A 48-hour period beginning at 6 p.m. on Monday ends at 6 p.m. on Wednesday, not at midnight Tuesday.
Applying the layer before grouping. Determine the occurrence total first; then apply its attachment and limit.
Misreading layer terminology. “$1,000,000 excess of $500,000” means a $1,000,000 limit attaching at $500,000. A separate statement that a layer has a $1,200,000 limit should be treated as a $1,200,000 limit.
Assuming one time window settles event identity. The cause and event definition still matter, even if losses fall within the stated hours.
Assuming out-of-window treatment. A claim outside the selected period may be a separate occurrence, outside the grouped occurrence, or treated under another contract rule. Follow the wording rather than silently choosing a result.
Applying one limit across all occurrences. A per-occurrence limit is applied separately to each occurrence, subject to any additional treaty-wide terms stated in the problem.

Key takeaway: Interpret the event definition and hours clause together. Establish the permitted window and claim grouping, aggregate claims by occurrence, and only then calculate each layer recovery. The wording—not the clock alone—determines the result.

Check Your Understanding

For each problem, use the stated event and window rules before applying the layer. All monetary amounts are hypothetical.

A clause groups losses from one event in the 72-hour period $[0,72)$; losses outside it are separate occurrences. Claims are $300,000 at hour 0, $500,000 at hour 24, and $400,000 at hour 60. A layer attaches at $750,000 and has a $500,000 limit per occurrence. What is total recovery? (Answer: $450,000.)
A clause fixes the window at $[5,53)$ hours and treats a claim outside it as a separate occurrence. Claims are $800,000 at hour 5, $600,000 at hour 45, and $700,000 at hour 80. A layer attaches at $500,000 and has a $600,000 limit per occurrence. Calculate total recovery. (Answer: $800,000.)
A treaty explicitly defines two distinct storms as separate occurrences, although they happen within the same 72-hour period. Storm A causes $1,100,000 of aggregate loss and Storm B causes $900,000. A layer attaches at $500,000 and has a $500,000 limit per occurrence. Find total recovery. (Answer: $900,000.)
A clause groups claims in $[0,72)$, with any later claim treated as a separate occurrence. Claims are $400,000 at hour 0, $900,000 at hour 50, and $800,000 at hour 100. The layer attaches at $500,000 and has a $800,000 limit per occurrence. Calculate total recovery. (Answer: $1,100,000.)
A treaty fixes the hours period to begin at the first loss. Three claims are $600,000 at hour 0, $700,000 at hour 35, and $900,000 at hour 65. Any claim outside the period is a separate occurrence. The layer attaches at $500,000 and has a $700,000 per-occurrence limit. Compare total recovery under a 48-hour period and a 72-hour period. (Answer: 48-hour recovery = $1,100,000; 72-hour recovery = $700,000; the 48-hour wording pays $400,000 more.)

What You'll Learn