Maintaining Coverage Consistency in Loss Cost Comparisons

Why Coverage Consistency Matters

A pure premium compares losses with exposures. That comparison is meaningful only when the losses represent the coverage being priced. If historical policies had a lower payment limit, a higher deductible, or narrower benefits than the proposed policy, their recorded losses do not directly represent the proposed coverage. The difference is not necessarily a change in underlying claim frequency or severity; it may simply be a change in what the insurer pays.

The earlier material on matching losses and exposures establishes the need to use a consistent basis for both. This tutorial focuses on a further part of that consistency: the historical losses must also be translated, where possible, to the benefits, limits, and deductibles of the coverage underlying the indicated rate. Such a translation is a coverage adjustment. It is distinct from loss trend, which projects claim costs through time, and from changing the exposure denominator.

Definition: A coverage-consistent loss comparison expresses historical losses on the same benefit terms as the target coverage for which the pure premium is being estimated. The target coverage basis specifies such terms as covered benefits, deductibles, and limits. A coverage-adjusted loss is an estimate of what the historical claim experience would have cost under that target basis, subject to the data and assumptions used.

For a claim with ground-up loss $X$, suppose the policy has a deductible $d$ and an insurer payment limit $u$. In this tutorial, $u$ is the maximum amount the insurer pays per claim, after applying the deductible. The insurer payment per loss is

Y^L(X;d,u)=\min\bigl((X-d)^+,u\bigr), \qquad (z)^+=\max(z,0).

If there is no binding payment limit, take $u$ to be sufficiently large. Policy forms sometimes define a limit differently—for example, as a maximum covered loss amount before the deductible. The definition in the policy must therefore be checked before calculating a payment. The formula above applies only to the stated payment-limit convention.

Translate Claims to the Target Coverage

When ground-up losses and relevant claim details are available, a direct approach is to calculate each claim payment twice: once under the historical terms and once under the target terms. Sum the target payments and divide by the appropriate exposure. For $n$ claims and exposure $E$, the target-basis pure premium is

\widehat{\mathrm{PP}}_{\mathrm{target}} = \frac{\sum_{j=1}^{n}Y^L(X_j;d_{\mathrm{target}},u_{\mathrm{target}})} {E},

with the payment function expanded or modified as necessary for the actual benefit provisions. The numerator should represent the selected loss measure used in the indication, such as ultimate losses, with the coverage adjustment applied consistently to that measure. This calculation assumes the claim information is sufficient to evaluate the target terms and that the selected exposures are appropriate for the same population and period.

Worked Example: Increasing a Payment Limit

A hypothetical portfolio has 800 exposure units and three selected ground-up losses of $18,000, $64,000, and $130,000. Historical coverage has no deductible and a $50,000 maximum insurer payment per claim. The proposed coverage has no deductible and a $90,000 maximum payment. Assume the ground-up losses are known even when the historical payment was capped. Calculate the historical and target-basis pure premiums. Method: recalculate each claim under each limit.

Under the historical limit, the three payments are:

\min(18{,}000,50{,}000)=18{,}000,\quad \min(64{,}000,50{,}000)=50{,}000,\quad \min(130{,}000,50{,}000)=50{,}000.

Historical losses total $18{,}000+50{,}000+50{,}000=\$118{,}000$. The historical-basis pure premium is $\$118{,}000/800=\$147.50$ per exposure unit.

Under the target limit, the payments are $\$18{,}000$, $\$64{,}000$, and $\$90{,}000$. Target-basis losses total $18{,}000+64{,}000+90{,}000=\$172{,}000$, so the target-basis pure premium is $\$172{,}000/800=\$215.00$ per exposure unit. The coverage adjustment increases the aggregate by $\$172{,}000-\$118{,}000=\$54{,}000$. The calculation avoids treating the old capped payments as if they included amounts above the historical limit.

Check: The target limit is higher, so no recalculated claim payment can be lower than its historical payment. The target total, $\$172{,}000$, exceeds $\$118{,}000$, as expected.

Deductibles and Benefit Changes

A deductible changes the amount the insurer pays, not the ground-up amount of the claim. If the deductible increases and all other terms are unchanged, the insurer payment for any claim cannot increase. This monotonicity is a useful check. Conversely, a broader benefit can add losses that were not included in the historical insurer payments. Those newly covered amounts need a defensible estimate; they cannot be recovered merely by rescaling losses that were recorded under narrower coverage.

A simple aggregate adjustment factor can be used when the same claims are represented under both coverage bases. If $L_{\mathrm{hist}}$ is the aggregate historical-basis loss and $L_{\mathrm{target}}$ is the aggregate loss recalculated for the target terms, define

C_{\mathrm{cov}}=\frac{L_{\mathrm{target}}}{L_{\mathrm{hist}}}.

Then $C_{\mathrm{cov}}L_{\mathrm{hist}}=L_{\mathrm{target}}$. The identity is useful, but it does not establish that the factor is appropriate for other claims or periods. Applying one aggregate factor assumes the claims being adjusted are sufficiently representative of the claims to which the factor will be applied. It also requires $L_{\mathrm{hist}}>0$. When claim sizes, benefit types, or coverage changes vary materially, claim-level adjustment or a more segmented estimate may be necessary.

Worked Example: Raising the Deductible

A hypothetical book has 250 exposure units. Its selected ground-up losses are $3,000, $9,000, and $24,000. The historical deductible is $1,000 per claim, and the proposed deductible is $4,000. Neither contract has a binding payment limit. Calculate both pure premiums. Method: apply each deductible to each ground-up loss before summing.

Under the historical deductible, the insurer payments are:

(3{,}000-1{,}000)^+=2{,}000,\quad (9{,}000-1{,}000)^+=8{,}000,\quad (24{,}000-1{,}000)^+=23{,}000.

Historical losses total $2{,}000+8{,}000+23{,}000=\$33{,}000$. The historical pure premium is $\$33{,}000/250=\$132.00$ per exposure unit.

Under the proposed deductible, the payments are $(3{,}000-4{,}000)^+=0$, $(9{,}000-4{,}000)^+=\$5{,}000$, and $(24{,}000-4{,}000)^+=\$20{,}000$. Target-basis losses total $\$25{,}000$, giving a pure premium of $\$25{,}000/250=\$100.00$ per exposure unit. The adjustment reduces losses by $\$8{,}000$. This method avoids subtracting the deductible only from claims that exceed it by an assumed average amount; the deductible is applied to each claim.

Check: The proposed deductible is larger, and each target payment is less than or equal to its historical payment. Accordingly, the target total and pure premium are lower.

Coverage extensions require special care. If a historical policy excluded a type of loss, the insurer may have no historical payment for it. A multiplier applied to the old paid losses cannot create information about the excluded claims. The extension must be estimated from suitable supplementary claim records, exposure-level models, or other justified information. The estimate should be identified separately from the observed historical losses so that its basis and uncertainty are visible.

Worked Example: Adding a Newly Covered Benefit

A hypothetical policy is being broadened to cover a particular equipment failure. For 600 exposure units, selected historical losses for benefits already covered total $19,000. A supplementary incident file, assumed complete for this illustration, identifies two equipment-failure losses of $4,000 and $9,000 that were excluded by the historical policy but would be covered by the proposed policy. Assume no deductible or limit applies to these amounts. Calculate the old and target-basis pure premiums. Method: add the separately estimated newly covered losses to the existing covered losses.

The historical-basis loss total is $\$19{,}000$, so the historical pure premium is $\$19{,}000/600=\$31.666\ldots$, or $\$31.67$ per exposure unit rounded to the nearest cent.

The newly covered losses total $\$4{,}000+\$9{,}000=\$13{,}000$. Thus target-basis losses are $\$19{,}000+\$13{,}000=\$32{,}000$, and the target-basis pure premium is $\$32{,}000/600=\$53.333\ldots$, or $\$53.33$ per exposure unit rounded to the nearest cent. The calculation avoids applying an unsupported factor to the old loss total when the relevant extension losses were absent from that total.

Check: The target loss total includes both the previously covered losses and the newly covered losses: $\$19{,}000+\$13{,}000=\$32{,}000$. The target pure premium is therefore greater than the historical pure premium.

Why the Adjustment Works—and When It Does Not

The pure premium is a loss total per exposure unit. Holding the exposure basis fixed, replacing historical payments with target-coverage payments changes only the numerator. If the historical payments for the same claims sum to $L_{\mathrm{hist}}$, and their target payments sum to $L_{\mathrm{target}}$, then

\frac{L_{\mathrm{hist}}}{E} \cdot \frac{L_{\mathrm{target}}}{L_{\mathrm{hist}}} = \frac{L_{\mathrm{target}}}{E}, \qquad L_{\mathrm{hist}}>0.

This justifies the aggregate factor as an algebraic conversion for those losses. It does not justify applying that factor to a different claim population without evidence that the factor transfers. In particular, a single factor may be misleading if the coverage change affects only certain claim types, if large claims are disproportionately limited, or if the historical records omit losses that the target policy would cover.

A deductible and a limit are nonlinear features: the payment changes differently depending on the size of each loss. Therefore, multiplying every recorded payment by a factor is generally not equivalent to recalculating each claim. For example, changing a deductible by $1,000 does not reduce every payment by $1,000; losses below the deductible can have zero payment, while losses well above it can be affected differently once a limit also applies.

Specify the target terms.
State which benefits are covered, how the deductible applies, and exactly what the limit caps.

Check what the data contain.
Determine whether ground-up losses and the information needed to apply target terms are available. Historical payments alone may conceal losses above limits or omit excluded claims.

Recalculate or estimate target losses.
Use claim-level payment calculations when possible. Estimate newly covered losses separately when they are absent from historical insurer payments.

Divide by the matching exposure.
Use the exposure basis appropriate to the same population and period represented by the target-basis losses.

Coverage adjustment and loss trend should not be confused. A trend projects claim costs from one time period to another; a coverage adjustment changes which portion of a claim the insurer pays. Their order can matter when a trend is applied to ground-up severity and the target contract has a fixed deductible or limit. For instance, a $10,000 ground-up loss with a $3,000 deductible produces a $7,000 payment. If ground-up severity is projected upward by 20% while the deductible remains fixed, the projected loss is $12,000 and the payment is $9,000. Multiplying the historical payment by 1.20 instead gives $8,400. These calculations differ because the deductible is applied to the projected ground-up loss in the first method. The appropriate sequence depends on what the trend represents and whether contract terms are fixed or themselves changing.

Common Exam Traps

Using historical capped payments as ground-up losses. A payment at the old limit does not reveal the claim amount above that limit. Do not calculate a new limit payment from capped data as if the uncapped amount were known.
Applying a new deductible as a flat reduction to every claim. Recalculate each payment using the deductible rule. A claim below the deductible pays zero, not a negative amount.
Scaling narrow-coverage losses to represent a new benefit. A factor cannot identify excluded losses that were never in the historical loss total. Use a separate, supportable estimate for the added coverage.
Changing the numerator but not checking the denominator. A coverage adjustment does not automatically change the exposure basis. The exposure must still represent the risks and period relevant to the adjusted losses.

Key takeaway: Compare pure premiums only after expressing losses on the target coverage basis. Recalculate claim payments when the data permit it; identify and estimate missing benefits separately when they do not. A coverage factor is a conversion for a defined set of losses, not a substitute for understanding what the policy covers.

Check Your Understanding

For each problem, state the target-basis loss total or pure premium as requested. Assume all loss amounts are ground-up unless the problem says otherwise.

A hypothetical portfolio has 500 exposure units and losses of $15,000, $55,000, and $80,000. The historical payment limit is $40,000 per claim, with no deductible; the target limit is $70,000. What is the target-basis pure premium? (Answer: $280.00 per exposure unit)
A book has 300 exposure units and ground-up losses of $1,000, $8,000, and $20,000. The deductible increases from $2,000 to $6,000, with no binding limit. What is the target-basis pure premium? (Answer: $53.33 per exposure unit)
Historical covered losses total $48,000 for 400 exposure units. A separate, suitable estimate of losses newly covered by a proposed benefit is $12,000 for the same exposure base. What target-basis pure premium results? (Answer: $150.00 per exposure unit)
A ground-up loss is $10,000. The target deductible is fixed at $3,000, and a 1.20 severity trend applies to ground-up loss. What is the trended insurer payment? (Answer: $9,000; trending the old $7,000 payment instead would give the incorrect $8,400)
For 100 exposure units, ground-up losses are $500, $4,000, and $10,000. Historical terms are a $1,000 deductible and a $5,000 payment limit; target terms are a $2,000 deductible and an $8,000 payment limit. What is the target-basis pure premium? (Answer: $100.00 per exposure unit)

What You'll Learn