Severity Trend Effects on Deductible and Limit Factors

Why Severity Trend Changes Coverage Factors

A deductible or policy limit is usually stated as a dollar amount. If claim severity rises while those dollar amounts stay fixed, the coverage represented by the deductible and limit changes relative to the loss distribution. A fixed deductible takes a smaller share of a larger typical loss; a fixed policy limit caps a smaller share of it. Consequently, a factor derived from expected payments can change even when the policy wording does not.

The calculation must apply trend to the severity distribution and then apply the unchanged dollar terms of coverage. Multiplying an old expected payment by a trend factor would instead be appropriate only if the deductible and limit were also multiplied by that factor. The distinction is important in revising factors for a new cost level.

Definition: Let $X$ be a nonnegative ground-up claim severity, and let $k=1+t$ be a proportional severity trend multiplier, where $t$ is the trend rate over the period. The trended severity is $X'=kX$. For a fixed deductible $d$ and a policy limit $u>d$, insurer payment per claim is $$ Y_{d,u}(X')=\min\big((X'-d)_+,u-d\big), $$ where $(z)_+=\max(z,0)$. With no policy limit, payment is $(X'-d)_+$. The combined factor relative to ground-up expected severity is $F_{d,u}(X')=E[Y_{d,u}(X')]/E[X']$. A factor is meaningful only with its reference basis stated.

Here $u$ is a limit on the total covered loss, measured from the ground up. Thus, when $u>d$, the maximum insurer payment is $u-d$. If a contract instead specifies a cap directly on insurer payment, that cap is the layer width and the corresponding upper loss threshold is $d$ plus the cap. This tutorial assumes a single claim, no expenses, and no discounting; interest and mortality assumptions are not applicable.

Expected Payment Under Proportional Trend

Write $S_X(x)=P(X>x)$ for the survival function of the original severity. For a fixed deductible and limit, the expected payment after trend is

E[Y_{d,u}(X')]=\int_d^u S_{X'}(y)\,dy =k\int_{d/k}^{u/k}S_X(z)\,dz,\qquad u>d.

For a deductible with no limit, replace $u$ by infinity. These expressions require a nonnegative severity and a finite expected payment over the coverage layer; the unlimited expected payment also requires $E[X]$ to be finite.

The first integral follows because a payment in the layer from $d$ to $u$ can be represented as the sum of small payment increments: the increment at loss level $y$ is received exactly when severity exceeds $y$. The expected increment is therefore $P(X'>y)\,dy=S_{X'}(y)\,dy$. For the second equality, $X'=kX$ implies $S_{X'}(y)=P(kX>y)=S_X(y/k)$. Substituting $z=y/k$, so $dy=k\,dz$, gives the scaled thresholds and multiplier shown. This derivation makes the key point explicit: fixed dollar thresholds become $d/k$ and $u/k$ when expressed in original-severity units.

Key result: With a fixed deductible and limit, calculate the revised payment from the trended distribution at the original dollar thresholds. Equivalently, use the original distribution with thresholds divided by $k$, and multiply the resulting expected layer payment by $k$. If both the severity and all coverage thresholds are multiplied by $k$, expected payment and ground-up mean both scale by $k$, so their ratio is unchanged.

Worked Applications

Worked Example: Exponential Severity With a Fixed Deductible

A hypothetical portfolio has single-claim severity $X$ exponentially distributed with mean $\$500$, so its rate parameter is $1/500$ per dollar. Severity is trended upward by 20%. The deductible remains fixed at $\$250$, and there is no policy limit. Find the revised expected payment above the deductible and the factor relative to trended ground-up mean. There is no discounting or mortality assumption.

Method. An exponential severity with mean $\theta$ has $S_X(x)=e^{-x/\theta}$, and its mean excess payment above deductible $d$ is $\theta e^{-d/\theta}$. Proportional trend preserves the exponential form and changes the mean to $k\theta$.

First calculate the trend multiplier and the trended mean:

k=1+0.20=1.20,\qquad E[X']=1.20(500)=\$600.

The post-trend exponential rate is $1/600$ per dollar. Apply the unchanged $\$250$ deductible:

E[(X'-250)_+]=600e^{-250/600} =600e^{-0.4166667} \approx \$395.54.

The factor uses the same trended distribution in its numerator and denominator:

F_{250}(X')=\frac{395.5444\ldots}{600} =e^{-250/600}\approx 0.6592.

Thus the expected payment is $\$395.54$, and the deductible factor is $0.6592$ to four decimal places. As a check, the payment cannot exceed the $\$600$ mean ground-up severity, and the factor is below 1. A common mistake is to retain the old deductible factor $e^{-250/500}$ as if it still applied: the deductible has not grown with severity, so the relevant ratio is $250/600$, not $250/500$.

Worked Example: Discrete Severity and a Fixed Deductible

A hypothetical severity distribution has $P(X=\$400)=0.4$ and $P(X=\$1{,}600)=0.6$. Trend severity by 10%, then apply a fixed $\$500$ deductible and no limit. Find expected payment and its factor relative to trended mean. There is no discounting or mortality assumption.

Method. Apply the trend multiplier to each possible claim amount, retain its probability, and then apply the deductible to each trended amount. The distribution is discrete; its probabilities do not change under this deterministic proportional trend.

k=1.10,\qquad X'= \begin{cases} \$440 & \text{with probability }0.4,\\ \$1{,}760 & \text{with probability }0.6. \end{cases}

The trended ground-up mean is

E[X']=0.4(440)+0.6(1{,}760)=176+1{,}056=\$1{,}232.

At $\$440$, the claim does not exceed the $\$500$ deductible, so payment is $\$0$. At $\$1{,}760$, payment is $\$1{,}760-\$500=\$1{,}260$. Therefore,

E[(X'-500)_+]=0.4(0)+0.6(1{,}260)=\$756, \qquad F_{500}(X')=\frac{756}{1{,}232}=0.613636\ldots\approx 0.6136.

The answer is $\$756$ and a factor of $0.6136$ to four decimal places. The result checks against the total mean: expected payment is positive and below $\$1{,}232$. The common incorrect calculation is to find the original expected payment, $0.4(0)+0.6(1{,}600-500)=\$660$, and multiply it by $1.10$, obtaining $\$726$. That scales the old payment layer as though its deductible had also risen to $\$550$. With the deductible fixed at $\$500$, the trended large claim pays $\$1{,}260$, not $1.10(1{,}100)=\$1{,}210$.

Worked Example: Exponential Severity With a Fixed Deductible and Policy Limit

A hypothetical claim severity is exponential with mean $\$1{,}000$, or rate $1/1{,}000$ per dollar. Severity is trended upward by 10%. The contract has a fixed $\$200$ deductible and a fixed $\$1{,}000$ ground-up policy limit. Find expected insurer payment and the combined factor relative to trended ground-up mean. There is no discounting or mortality assumption.

Method. The policy limit is measured from ground up, so the insurer pays at most $1{,}000-200=\$800$. The payment layer runs from loss amount $\$200$ through $\$1{,}000$. After trend, severity remains exponential, with mean $1.10(1{,}000)=\$1{,}100$.

E[Y_{200,1000}(X')] =\int_{200}^{1000}e^{-y/1100}\,dy =1100\left(e^{-200/1100}-e^{-1000/1100}\right).

Evaluate the two survival terms:

e^{-200/1100}=e^{-0.1818182}\approx0.8337529,\qquad e^{-1000/1100}=e^{-0.9090909}\approx0.4028903.

Thus the expected payment is

1{,}100(0.8337529-0.4028903) =1{,}100(0.4308626) \approx \$473.95.

The combined factor is

F_{200,1000}(X')=\frac{473.9489\ldots}{1{,}100} =0.4308626\ldots\approx0.4309.

The expected payment is below the maximum per-claim payment of $\$800$, and the factor is below 1, as expected. The calculation uses the policy limit as a ground-up threshold of $\$1{,}000$; integrating to $\$800$ would confuse the maximum insurer payment with the ground-up loss threshold. If both coverage thresholds were trended to $\$220$ and $\$1{,}100$, respectively, the coverage factor would equal its pre-trend value. Here the thresholds stay fixed, so that invariance does not apply.

Applying Trend Consistently

For a factor calculation, first specify exactly what the numerator represents and what is used as the denominator. In the examples above, the denominator is ground-up expected severity under the same trended distribution. This gives a factor relative to ground-up loss cost. A comparison between two coverage plans instead uses the expected payment under one plan divided by the expected payment under the other; both expectations must be evaluated under the same severity distribution and trend assumption.

A reliable calculation sequence is:

State the trend basis.
Set $k=1+t$, identify the original severity distribution, and determine its trended distribution or mean.

Hold fixed terms fixed.
Use the deductible $d$ and ground-up policy limit $u$ in the contract. Do not scale these dollar amounts unless the problem explicitly says the coverage terms change.

Calculate expected payment.
Apply the deductible and limit to each trended claim, or integrate the trended survival function over the covered layer.

Form the stated factor.
Divide by the specified reference expected payment, calculated on a consistent basis. Round only the final answer.

A factor is not a property of the deductible or limit alone: it depends on the severity distribution and on the factor's reference coverage. A factor calculated from an untrended distribution should not be carried forward automatically when severity changes. Recalculate it using the revised distribution whenever the dollar thresholds remain fixed.

Common exam trap: $kE[(X-d)_+]$ is generally not the expected payment after trend when $d$ is fixed. It equals the post-trend payment only when the deductible is also scaled to $kd$. For a fixed deductible, the correct expression is $E[(kX-d)_+]$, and with a policy limit the fixed limit must also be applied to $kX$.

Check Your Understanding

For each problem, use the factor basis stated and keep fixed dollar terms unchanged unless the problem explicitly changes them.

A hypothetical exponential claim severity has mean $\$800$ and rate $1/800$ per dollar. Severity is trended by 25%; a fixed $\$400$ deductible applies with no limit. Find expected payment and the factor relative to trended mean. There is no discounting or mortality. (Answer: $\$670.32$; factor $0.6703$.)
A hypothetical severity is $\$500$ with probability 0.3 and $\$1{,}500$ with probability 0.7. Trend it by 20%, then apply a fixed $\$700$ deductible and no limit. Find the expected payment and its factor relative to trended mean. There is no discounting or mortality. (Answer: $\$770$; factor $0.5347$.)
A hypothetical exponential severity has mean $\$600$ and rate $1/600$ per dollar. Trend it by 10%; apply a fixed $\$300$ deductible and a fixed $\$800$ ground-up policy limit. Find expected payment and the factor relative to trended mean. There is no discounting or mortality. (Answer: $660(e^{-300/660}-e^{-800/660})\approx\$222.53$; factor $0.3372$.)
A hypothetical severity is $\$200$, $\$800$, or $\$1{,}400$, with probabilities 0.2, 0.5, and 0.3, respectively. Trend it by 15%; apply a fixed $\$400$ deductible and fixed $\$1{,}100$ ground-up policy limit. Find expected payment and the factor relative to trended mean. There is no discounting or mortality. (Answer: $\$470$; factor $0.4752$.)

What You'll Learn