Why Severity and Coverage Must Be Kept Separate
A claim payment can change even when the underlying loss has not. For example, an insurer might pay less on an otherwise identical claim because the deductible increased or the policy limit decreased. Conversely, ground-up claim severity can rise while a coverage change partly or entirely offsets the increase in the amount paid. To interpret a change correctly, identify what describes the loss and what describes the contract.
Ground-up severity is the amount of the covered loss before applying the policy’s deductible, limit, or coinsurance. A trend factor describes a change in the price level of losses between periods. Deductibles, limits, and coinsurance are contract terms that determine how much of a ground-up loss is paid by the insurer. These are different inputs to a claim calculation and should not be treated as interchangeable.
The formula for \(Y^L\) assumes that the deductible is applied first and that \(u\) is a maximum insurer payment after the deductible. Other policy wordings can define limits differently, so the contract’s definition must be checked before using the formula. In particular, an observed insurer payment is not automatically a ground-up loss or a measure of ground-up severity.
A Framework for Separating the Inputs
A useful comparison names three things separately: the loss amount and its price basis; the coverage terms; and the resulting payment. For a historical loss, a trend factor may be calculated from the relevant period indices. The deductible and limit remain contract parameters unless the scenario explicitly changes them. When comparing two periods, first ask whether the comparison is intended to isolate a change in severity, a change in coverage, or the combined effect.
| Quantity | What it describes | Examples |
|---|---|---|
| Ground-up severity | The loss before coverage modification | Claim amount \(X\), mean ground-up loss |
| Severity trend | A change in loss price basis between periods | Trend factor \(k\) applied to historical loss amounts |
| Coverage terms | The contract’s sharing of the loss | Deductible \(d\), payment limit \(u\), coinsurance percentage |
| Insurer payment | The amount paid under the specified contract | \(Y^L\) after deductible and limit |
One way to isolate a severity effect is to compare loss amounts on a common price basis while keeping the coverage design fixed. To isolate a coverage effect, compare different contract terms for the same ground-up loss amounts. If both severity and coverage have changed, report both changes rather than attributing the entire payment difference to either one. This separation is a bookkeeping principle: it does not imply that claim data alone can always identify every underlying quantity.
Worked Comparisons
The examples use hypothetical claims. Unless stated otherwise, there is no interest or mortality assumption because neither is relevant to these loss comparisons. Amounts are treated as exact; any displayed rounding is stated explicitly.
Worked Example: One Trended Loss Under Two Contracts
A hypothetical ground-up claim of $20,000 is observed in an earlier price period. A stated trend factor of 1.20 converts it to the comparison period. Compare the insurer payment under (a) a $2,000 deductible and a $20,000 payment limit, and (b) a $5,000 deductible and a $15,000 payment limit. Assume in both contracts that the limit is the maximum insurer payment after the deductible.
Method. First identify the trended ground-up loss. Then calculate each payment using that loss and the relevant contract terms. The distribution is a single deterministic hypothetical loss; no probability model is needed.
The loss on the comparison-period price basis is
For contract (a), the amount above the deductible is \(24{,}000-2{,}000=22{,}000\). Applying the payment limit gives
For contract (b), the amount above the deductible is \(24{,}000-5{,}000=19{,}000\). Applying that contract’s limit gives
Check. Both contracts are applied to the same $24,000 ground-up loss. The payments differ by $5,000, while the ground-up severity is identical. The difference is due to the coverage terms, not a difference in loss severity. The calculation also shows why the amounts above the deductibles must be computed before comparing them with the limits: they are $22,000 and $19,000, respectively.
Common mistake avoided: It would be incorrect to describe the $15,000 payment under contract (b) as a $15,000 ground-up loss. It is the insurer’s payment after both the deductible and the limit have been applied.
Worked Example: Separating a Sample Trend From a Contract Change
A hypothetical portfolio has three historical ground-up losses of $10,000, $15,000, and $20,000. The losses are equally weighted, and a common trend factor of 1.20 is specified for comparison with a later period. The historical contract has a $2,000 deductible and a $20,000 payment limit. A later contract has a $5,000 deductible and a $15,000 payment limit. Assume both limits are maximum insurer payments after the deductible. Calculate the historical mean severity, the trended mean severity, and mean insurer payments under each contract when the trended losses are used.
Method. Calculate the ground-up sample mean before any coverage modification. Trend each loss by the stated factor to obtain the later-period comparison amounts. Then calculate payments separately under the two contracts. The severity model here is an equally weighted empirical distribution on the three stated losses; no parametric distribution is assumed.
The historical mean ground-up severity is
The three trended losses are \(10{,}000(1.20)=12{,}000\), \(15{,}000(1.20)=18{,}000\), and \(20{,}000(1.20)=24{,}000\). Their mean is
Under the historical contract terms, the payments on the trended losses are:
Thus the mean payment under those terms is \((10{,}000+16{,}000+20{,}000)/3=46{,}000/3=\$15{,}333.33\), rounded to the nearest cent.
Under the later contract terms, the payments on the same trended losses are:
The mean payment under the later terms is \((7{,}000+13{,}000+15{,}000)/3=35{,}000/3=\$11{,}666.67\), rounded to the nearest cent.
Check. The ground-up mean rises from $15,000 to $18,000 on the stated price basis. For the same trended loss sample, the mean payment is lower under the later contract terms: $11,666.67 rather than $15,333.33. The mean payments therefore reflect both the loss amounts and the contract used. They should not be substituted for the mean ground-up severity.
Common mistake avoided: Comparing only the two payment means and calling their difference a severity trend would mix the trend with a change in deductible and limit. To describe the severity change, compare the ground-up means on the stated price basis.
Worked Example: Recovering Severity From an Uncapped Payment
A hypothetical historical claim produced an insurer payment of $8,000 under a $2,000 deductible and a $20,000 payment limit. A comparable later claim produced an $8,000 payment under a $5,000 deductible and a $15,000 payment limit. In both cases, assume the limit is applied after the deductible. A trend factor of 1.20 is specified for the historical claim. Determine the ground-up losses implied by the payments and compare them on the later-period price basis.
Method. First check whether each payment is below the limit. If so, the payment is an interior payment, and under the stated contract formula it equals the ground-up loss less the deductible, provided the loss exceeds the deductible.
For the historical claim, $8,000 is below the $20,000 limit. Its implied ground-up loss is \(8{,}000+2{,}000=\$10{,}000\). The claim exceeds the deductible, and the implied payment is below the limit, as required. Trending this historical severity gives \(10{,}000(1.20)=\$12{,}000\).
For the later claim, $8,000 is below the $15,000 limit. Its implied ground-up loss is \(8{,}000+5{,}000=\$13{,}000\). This also exceeds the deductible, and its payment remains below the limit.
Check. The two recorded payments are identical, but their implied ground-up losses are not: $10,000 historically and $13,000 later. On the later-period price basis, the trended historical amount is $12,000, compared with the later amount of $13,000. These arithmetic comparisons do not, by themselves, establish a portfolio-wide trend; the example concerns only the stated claims and assumptions.
Common mistake avoided: Treating both $8,000 payments as identical ground-up losses would ignore the different deductibles. Also, one cannot recover a ground-up loss by adding the deductible if the payment is capped at the limit; the payment may then correspond to many possible loss amounts.
Why a Payment at the Limit Is Different
For an interior payment under the stated deductible-and-limit formula, the ground-up loss can be recovered as payment plus deductible. But a payment at the limit does not identify the exact loss. If \(Y^L=u\), all that the formula establishes is that the amount above the deductible is at least \(u\). Therefore, \(X\ge d+u\). A limit-censored payment can support a lower bound on severity, but not an exact value, without further information.
This distinction matters when historical claim records contain insurer payments rather than ground-up losses. A change in the distribution of payments might reflect a change in severity, a change in the contract, a change in which claims are capped, or a combination. To draw a severity conclusion, the analyst needs the claim basis and the coverage terms, and must account for payments whose ground-up amounts cannot be uniquely recovered.
Common Exam Traps
- Calling a payment a loss. State whether an amount is ground-up severity or an insurer payment. The same dollar value can represent either, depending on context.
- Attributing every payment change to trend. A deductible or limit change can alter payments even when the underlying loss amounts stay fixed.
- Ignoring the price basis. A historical severity and a later-period severity are not directly comparable as nominal amounts when a trend factor has been specified.
- Inverting a capped payment as if it were interior. A payment equal to the limit generally gives a lower bound on \(X\), not an exact ground-up amount.
- Changing more than one input without saying so. If both the severity basis and contract terms differ, identify both before interpreting the resulting payment comparison.
Check Your Understanding
For each problem, distinguish the loss quantity from the coverage terms before calculating. All amounts are hypothetical.
- A ground-up loss of $16,000 is converted to a later price basis using a trend factor of 1.10. The contract has a $1,600 deductible and a $30,000 payment limit applied after the deductible. Find the later-basis loss and insurer payment. (Answer: ground-up loss $17,600; payment $16,000.)
- A $24,000 ground-up loss is subject to either a $2,000 deductible and $20,000 payment limit, or a $5,000 deductible and $15,000 payment limit. Find both payments and state whether the ground-up severity changed. (Answer: $20,000 under the first contract and $15,000 under the second; ground-up severity is $24,000 under both.)
- An equally weighted empirical sample of ground-up losses is $8,000, $12,000, and $20,000. A common trend factor is 1.25. Find the historical and trended sample means, before coverage modification. (Answer: historical mean $13,333.33; trended mean $16,666.67, each rounded to the nearest cent.)
- A claim payment is $7,000 under a $3,000 deductible and a $18,000 payment limit. The payment is below the limit. Recover the ground-up loss and then express it on a later price basis using a trend factor of 1.15. (Answer: ground-up loss $10,000; later-basis amount $11,500.)
- A claim payment equals a $10,000 limit under a contract with a $4,000 deductible, with the limit applied after the deductible. What can be concluded about the ground-up loss? If the claim is historical and a trend factor of 1.20 applies, what lower bound follows on its later-basis amount? (Answer: ground-up loss is at least $14,000; later-basis amount is at least $16,800, but neither value is identified exactly.)