Splitting Exposure Across Coverages

Why Exposure Must Be Split by Coverage

A single policy can provide several coverages, but those coverages need not protect the same items, apply for the same amount of time, or use the same exposure basis. A fleet policy, for example, might provide liability coverage for every vehicle while collision coverage applies only to selected vehicles. Counting the policy once does not reveal either coverage’s exposure.

For ratemaking and experience analysis, the relevant exposure is generally the amount of a particular coverage’s exposure base during the period being measured. The policy is the container for the coverages; it is not automatically the correct unit of exposure for every coverage. A policy-level total can be useful for counting policies, but it cannot replace separate coverage totals when those totals differ.

Definition: A policy-level exposure is an exposure measure attached to a policy as a whole, such as one policy-year for a policy in force for one year. A coverage-specific exposure is the exposure measured for one named coverage, using that coverage’s specified exposure basis and the time for which the coverage applies. Examples of bases include vehicle-years, location-years, payroll, sales, or insured-value-years.

The choice of basis is not arbitrary: use the basis defined for the coverage being analyzed. When the basis is a count of units and the number of covered units remains constant, exposure is the number of units multiplied by the fraction of a year covered. When an amount such as insured value changes, calculate exposure separately for each interval with a constant amount, then add the interval results. This calculation assumes exposure accumulates in proportion to time within each interval.

E_{p,c}=\sum_{k} q_{p,c,k}\,t_{p,c,k}

Here $E_{p,c}$ is the exposure for policy $p$ and coverage $c$; $q_{p,c,k}$ is the amount of the coverage’s exposure base during interval $k$; and $t_{p,c,k}$ is the fraction of a year in that interval. The formula applies when the chosen base is accumulated linearly over time and each $q$ is constant within its interval. For a vehicle-year basis, $q$ is the number of covered vehicles. For an insured-value-year basis, $q$ is the insured amount expressed in the chosen units. If the basis is not time-proportional, this multiplication is not appropriate; use the basis’s prescribed measurement instead.

Record exposure by coverage before aggregating. Even when two coverages happen to use the same unit, their separate totals answer different questions. When their units differ, adding the numbers does not produce a meaningful combined exposure.

Calculating Coverage-Specific Exposure

For each policy, first list its coverages and the basis applicable to each. Then identify the covered units or amounts and the fraction of the measurement period for which each applies. Split an interval whenever the covered units, insured amount, or coverage status changes. Calculate each coverage separately; only combine values that have the same defined unit and are intended to be reported together.

Name the coverage and its basis.
For example, identify whether the calculation uses vehicle-years or insured-value-years. Do not assume the whole policy has one universal basis.

Identify covered units and intervals.
Use the applicable coverage record to determine which vehicles, locations, or insured amounts are covered, and for how much of the measurement period.

Calculate and sum within that coverage.
Multiply each interval’s exposure base by its time fraction when the basis is time-proportional, then add the interval results for that coverage.

Keep the coverage totals identifiable.
Report each total with its coverage name and unit. A policy count or a sum across unlike bases does not provide the missing coverage detail.

Worked Example: One Fleet Policy With Different Vehicle Selections

A hypothetical one-year fleet policy covers two vehicles for liability for the entire year. Collision coverage applies to Vehicle A for the entire year, while comprehensive coverage applies to Vehicle A for the first half-year only. Use vehicle-years for all three coverage bases. Find each coverage’s exposure and compare the results with the policy-level exposure.

Method. Calculate the covered vehicle count times the covered fraction of a year separately for each coverage. No interest or mortality assumptions are needed because this is a deterministic exposure calculation.

Liability. Two vehicles are covered for one full year:

E_{\text{liability}}=2\times 1=2.00\text{ vehicle-years}.

Collision. One vehicle is covered for one full year:

E_{\text{collision}}=1\times 1=1.00\text{ vehicle-year}.

Comprehensive. One vehicle is covered for one-half year:

E_{\text{comprehensive}}=1\times 0.5=0.50\text{ vehicle-year}.

The policy itself is in force for one year, so its policy-level exposure is $1.00$ policy-year. The three coverage totals are $2.00$, $1.00$, and $0.50$ vehicle-years, respectively. Their arithmetic sum is $3.50$ coverage vehicle-years, but that sum counts exposure across coverage records, not unique policy-years. It does not replace any of the three individual coverage totals.

Check. Liability exposure exceeds one vehicle-year because two vehicles have liability coverage all year. Comprehensive exposure is half a vehicle-year because only one vehicle has that coverage for half the year. The calculation avoids the mistake of assigning the policy’s single policy-year to every coverage.

Changing Amounts and Different Bases

The same separation is needed when a coverage’s insured amount changes. If the amount is constant within each interval, calculate the amount-times-time contribution for each interval. Do not use just the amount at inception or the amount at the end of the year for the entire period. The result should retain the coverage name and the unit used to express the insured amount.

Worked Example: Building and Contents Exposure on a Package Policy

A hypothetical commercial property policy is measured over one year. Building coverage has an insured value of $500,000 for the first 0.4 year and $700,000 for the remaining 0.6 year. Contents coverage has an insured value of $180,000 for the full year. Calculate each coverage’s exposure in $100,000 insured-value-years.

Method. Apply the piecewise exposure formula, converting each insured value into units of $100,000 before multiplying by its time fraction.

For building coverage, $500,000 is 5 units of $100,000, and $700,000 is 7 such units. Thus:

E_{\text{building}}=(5\times 0.4)+(7\times 0.6)=2.0+4.2=6.2

The building exposure is $6.2$ $100,000 insured-value-years. Equivalently, before scaling, the amount-time calculation is $$500{,}000\times0.4)+($700{,}000\times0.6)=$620{,}000$ insured-value-years.

For contents coverage, $180,000 is 1.8 units of $100,000 and applies for the whole year:

E_{\text{contents}}=1.8\times 1=1.8

The contents exposure is $1.8$ $100,000 insured-value-years, or $180,000 insured-value-years before scaling. The building and contents figures use the same stated unit, so their arithmetic sum is $8.0$ $100,000 insured-value-years. However, that combined number still does not tell an analyst how much exposure belongs to building versus contents. Preserve the separate totals when analyzing the coverages separately.

Check. The building value averaged over the year is $620,000 by the amount-time calculation, not $700,000, the ending amount. The contents amount does not change. Using the ending building value for all twelve months would overstate building exposure.

Why a Policy Total Cannot Recover Coverage Totals

A policy-level total compresses information. If the total is one policy-year, it tells the analyst that a policy was in force for one year under that particular policy-level measure; it does not say which coverages were selected, how many covered units each had, or how their amounts changed. Different coverage arrangements can therefore produce the same policy-level total but different coverage-specific exposures.

The reason is visible from the basic calculation. Suppose a policy has two coverages, $A$ and $B$. The policy-level measure may be one policy-year regardless of their individual exposure amounts. The coverage record instead supplies a pair $(E_{p,A},E_{p,B})$. Knowing only one policy-level number does not determine both components. Even a sum of coverage exposures cannot generally be reversed to recover the separate components.

E_{p,A}+E_{p,B}=T \quad\not\Rightarrow\quad \text{known values of }E_{p,A}\text{ and }E_{p,B}

For example, a total of 3 vehicle-years across two coverages could arise from 2 liability vehicle-years and 1 collision vehicle-year, or from 1 liability vehicle-year and 2 collision vehicle-years. The same aggregate does not identify which coverage has which amount. And if the coverages use unlike bases, such as vehicle-years and insured-value-years, their numerical values should not be added at all.

Worked Example: Aggregating a Small Fleet Without Losing Coverage Detail

A hypothetical one-year fleet policy covers three vehicles. Liability applies to all three vehicles for the whole year. Collision applies to two vehicles for the whole year and to a third vehicle for the final 0.5 year. Comprehensive applies to one vehicle for 0.25 year. Calculate coverage-specific vehicle-years, the policy-level exposure, and the arithmetic sum of the coverage exposures.

Method. Add vehicle-year contributions within each coverage, then report the policy-level and coverage-level measures separately.

Liability covers three vehicles for one year:

E_{\text{liability}}=3\times1=3.00\text{ vehicle-years}.

Collision covers two vehicles for one year and one vehicle for half a year:

E_{\text{collision}}=(2\times1)+(1\times0.5)=2.00+0.50=2.50\text{ vehicle-years}.

Comprehensive covers one vehicle for one-quarter year:

E_{\text{comprehensive}}=1\times0.25=0.25\text{ vehicle-year}.

The policy-level exposure is $1.00$ policy-year. The arithmetic sum across the three coverage vehicle-year totals is $3.00+2.50+0.25=5.75$ coverage vehicle-years. That sum is not the policy-level exposure and does not imply 5.75 distinct vehicles or policy-years; the same vehicle can contribute exposure under multiple coverages. The coverage results needed for analysis remain liability $3.00$, collision $2.50$, and comprehensive $0.25$ vehicle-years.

Check. Liability must have the largest total because all three vehicles are covered throughout the year. Comprehensive is only a quarter vehicle-year. Treating the single policy-year as the exposure for all coverages would erase these differences.

Common Mistakes and Practical Checks

The central discipline is to keep the unit of analysis visible. Before accepting a total, ask: Which coverage does it describe? What is its exposure basis? What period does it cover? A number without those labels may be difficult to interpret and may not be usable for coverage-level analysis.

Using one policy-year for every coverage. A policy can have multiple covered units for one coverage and only a subset for another. Calculate each coverage from its own covered units.
Counting a covered item once without regard to time. A vehicle covered for half a year contributes 0.5 vehicle-year, not one full vehicle-year, under a time-proportional vehicle-year basis.
Using only the beginning or ending amount. When an insured amount changes, split the period and calculate each constant-amount segment. Using a single snapshot can misstate the amount-time exposure.
Combining unlike units. A vehicle-year and a dollar of insured-value-year are different measures. Adding them has no meaningful exposure interpretation.
Assuming a combined total preserves the detail. Even when coverage units match, retaining the coverage breakdown matters: the aggregate alone cannot reveal its components.

Key takeaway: Calculate exposure for each coverage using that coverage’s own basis, covered units or amounts, and applicable time. Policy-level exposure and coverage-level exposure answer different questions; preserve coverage labels and units rather than substituting one total for another.

Check Your Understanding

For each problem, calculate and label the coverage-specific exposure before comparing it with any policy-level measure.

A one-year hypothetical vehicle policy has liability on three vehicles for the full year and collision on two vehicles for the full year. Using vehicle-years, find both coverage exposures and the policy-level exposure. (Answer: liability $3.00$ vehicle-years, collision $2.00$ vehicle-years, and $1.00$ policy-year.)
A hypothetical equipment policy has $300,000 of equipment coverage for 0.25 year and $500,000 for the remaining 0.75 year. Express exposure in $100,000 insured-value-years. (Answer: $0.75+3.75=4.50$ $100,000 insured-value-years.)
A one-year fleet policy has four vehicles. Liability covers all four for the full year; collision covers three for the full year and one for 0.5 year. Find each vehicle-year total and their sum. (Answer: liability $4.00$, collision $3.50$, and $7.50$ coverage vehicle-years in total.)
On a hypothetical one-year property policy, building coverage of $400,000 applies for the full year. Contents coverage is $200,000 for 0.5 year and $100,000 for 0.5 year. In $100,000 insured-value-years, find each coverage exposure. (Answer: building $4.00$ and contents $1.50$ $100,000 insured-value-years.)
A policy-level record reports one policy-year. Can that value determine the separate vehicle-year exposures for liability and collision? State why, and give an example of two different coverage exposure pairs with the same policy-level total. (Answer: No; for example, $(2,1)$ and $(1,2)$ vehicle-years both can occur on a one-policy-year record, so the policy total does not identify the coverage components.)

What You'll Learn