Incremental Layer Costs from Increased Limits Factors

Why Incremental Layer Costs Matter

An increased limits factor summarizes how expected payment changes as a per-claim limit increases. But a rating or pricing question may ask for the cost of just the additional coverage above an attachment point, rather than the cost of all coverage from the ground up. The key is to compare expected payments at the two endpoints of the layer.

For example, the expected payment under a $500,000 limit includes losses paid below $250,000 as well as payments in the $250,000-to-$500,000 layer. Subtracting the expected payment under a $250,000 limit removes the lower portion, leaving the expected cost of the incremental layer.

Definition: Let $X$ be a nonnegative ground-up loss amount for one claim, and let $u$ be a per-claim policy limit with no deductible. The expected capped loss at limit $u$ is $C(u)=E[\min(X,u)]$. For attachment $a$ and upper limit $b$, where $0\leq a\leq b$, the payment in the layer is $Y_{a,b}=\min((X-a)^+,b-a)$, where $z^+=\max(z,0)$. Its expected cost per claim is $C(b)-C(a)$.

From Capped Losses to Layer Cost

The capped loss $C(u)$ is the expected amount paid under a limit of $u$, when the insurer pays from the ground up and pays no more than $u$ for a claim. The layer from $a$ to $b$ instead pays nothing until the loss exceeds $a$, then pays the excess above $a$, capped at the layer width $b-a$. Thus, $a$ is the attachment point and $b$ is the upper limit; the layer width is not the same as the upper limit.

The identity behind the calculation can be checked for each possible loss amount. If $X\leq a$, both capped losses are $X$, so their difference is zero. If $a<X\leq b$, the difference is $X-a$. If $X>b$, the difference is $b-a$. In all three cases, the difference is exactly the layer payment.

\min(X,b)-\min(X,a)=\min((X-a)^+,b-a)

Taking expectations on both sides gives the layer-cost formula:

E[Y_{a,b}]=E[\min(X,b)]-E[\min(X,a)]=C(b)-C(a)

This is a difference of expected amounts, not a ratio. It applies whether the loss distribution is continuous, discrete, or a mixture, provided the same ground-up loss basis and payment definition are used at both endpoints. It does not directly apply when the two capped costs use different deductibles, claim definitions, or other payment terms.

For a nonnegative loss, the same expected layer cost can also be expressed using the survival function $S_X(x)=P(X>x)$:

E[Y_{a,b}]=\int_a^b S_X(x)\,dx

This follows because a layer pays one dollar for each dollar of layer width that the loss reaches. At each level $x$ between $a$ and $b$, the probability that the loss reaches that level is $P(X>x)$. Integrating those probabilities across the layer gives the expected payment. This form is useful when the survival function is known; otherwise, subtracting capped expected losses is often more direct.

Using Increased Limits Factors

An increased limits factor (ILF) expresses capped expected loss at a selected limit relative to capped expected loss at a designated basic limit $L_0$. Under a consistent ground-up, per-claim basis, define

ILF(u)=\frac{C(u)}{C(L_0)}

If the expected capped loss at the basic limit is $C(L_0)$, then $C(u)=ILF(u)C(L_0)$. It follows that the expected cost per claim for a layer from $a$ to $b$ is

E[Y_{a,b}]=\bigl(ILF(b)-ILF(a)\bigr)C(L_0)

Both factors must be measured relative to the same basic limit and on the same basis. If the attachment or upper limit is not listed, this formula by itself does not provide the missing factor. Determining a factor between listed limits is a separate interpolation question.

The result so far is a cost per claim. If the expected claim frequency is $\lambda$ claims per exposure unit, the corresponding expected layer cost per exposure is $\lambda E[Y_{a,b}]$. This multiplication assumes the frequency measure and the per-claim severity measure refer to the same claim population and exposure basis.

Worked Example: Layer Cost from Two ILFs

A hypothetical liability program has an expected capped loss of $500 per claim at its $100,000 basic limit. Its ILF is 1.30 at $250,000 and 1.55 at $500,000. The expected claim frequency is 0.08 claims per exposure. Find the expected cost of the $100,000-to-$250,000 layer per claim and per exposure.

Method and assumptions. Use the difference between capped expected losses, calculated from ILFs with the same $100,000 basic limit. The factors are supplied, so no claim severity distribution or distribution parameters are assumed. Interest and mortality assumptions are not applicable.

First calculate the capped expected losses at the layer endpoints:

C(100{,}000)=500,\qquad C(250{,}000)=1.30(500)=650

Subtract the lower-limit cost from the upper-limit cost:

E[Y_{100{,}000,250{,}000}]=650-500=150

The expected layer cost is $150 per claim. Convert to cost per exposure using the given frequency:

$$ 0.08(150)=12 $$

The result is $150.00 per claim and $12.00 per exposure. As a check, the layer is only the increase from the $100,000 cap to the $250,000 cap; it does not include the $500 already expected under the basic limit.

Common mistake avoided: Treating $1.30(500)=650$ as the layer cost would include the expected payment below $100,000. Subtracting the capped cost at the attachment leaves only the incremental layer.

Worked Example: Layer Cost from an Exponential Severity Model

For a hypothetical property portfolio, suppose ground-up severity $X$ for a claim follows an exponential distribution with mean $\theta=\$50{,}000$. Find the expected per-claim payment in the layer attaching at $50,000 and extending to a $150,000 upper limit. The expected claim frequency is 0.04 claims per exposure.

Method and assumptions. Use the capped-loss difference. Here $X$ is exponentially distributed with mean $\theta=\$50{,}000$, so $S_X(x)=e^{-x/50{,}000}$ for $x\geq0$. No deductible applies. Interest and mortality assumptions are not applicable.

For an exponential severity with this mean, the capped expected loss at limit $u$ is

C(u)=\int_0^u e^{-x/50{,}000}\,dx =50{,}000\left(1-e^{-u/50{,}000}\right)

At the $150,000 upper limit:

C(150{,}000)=50{,}000(1-e^{-3}) =50{,}000(1-0.0497871) =\$47{,}510.65

At the $50,000 attachment:

C(50{,}000)=50{,}000(1-e^{-1}) =50{,}000(1-0.3678794) =\$31{,}606.03

Subtract the capped expected losses:

E[Y_{50{,}000,150{,}000}] =47{,}510.65-31{,}606.03 =\$15{,}904.62

Then convert the per-claim cost to a per-exposure cost:

0.04(15{,}904.62)=\$636.1848 \approx \$636.18

Rounded to cents, the expected cost is $15,904.62 per claim and $636.18 per exposure. The calculation is consistent with the survival-integral form: the layer cost is $\int_{50{,}000}^{150{,}000} e^{-x/50{,}000}\,dx$.

Common mistake avoided: The layer width is $100,000, but that does not mean the insurer pays $100,000 on every claim. Payment depends on how often severity reaches each level within the layer.

Worked Example: Splitting a Layer at an Intermediate Limit

A hypothetical commercial policy has a $100,000 basic limit, with expected capped loss of $720 per claim at that limit. The ILFs are 1.27 at $250,000, 1.49 at $500,000, and 1.63 at $1,000,000. The expected claim frequency is 0.06 claims per exposure. Find the expected cost per claim and per exposure for the layer from $250,000 to $1,000,000, and verify it by splitting the layer at $500,000.

Method and assumptions. Use differences of ILFs, all relative to the same $100,000 basic limit, and multiply by the basic-limit expected capped loss. No severity distribution is assumed beyond the supplied factors. Interest and mortality assumptions are not applicable.

For the full layer, the factor difference is $1.63-1.27=0.36$. Thus:

E[Y_{250{,}000,1{,}000{,}000}] =0.36(720)=\$259.20\text{ per claim}

For the lower sublayer, from $250,000 to $500,000:

(1.49-1.27)(720)=0.22(720)=\$158.40\text{ per claim}

For the upper sublayer, from $500,000 to $1,000,000:

(1.63-1.49)(720)=0.14(720)=\$100.80\text{ per claim}

The sublayer costs sum to $158.40+100.80=\$259.20$, matching the full-layer result. The expected cost per exposure is:

0.06(259.20)=\$15.552\approx \$15.55

The result is $259.20 per claim and $15.55 per exposure, with the final per-exposure amount rounded to cents. The sublayer check confirms that the intermediate limit partitions the layer without changing its total expected cost.

Common mistake avoided: Adding the ILFs at the endpoints would combine costs below the attachment with the layer. The correct operation is to subtract the factors first, then multiply by the common basic-limit cost.

Exam Traps and Practical Checks

Incremental layer calculations are straightforward once the payment basis is fixed, but small interpretation errors can produce large differences. Before calculating, identify the ground-up loss basis, the attachment, the upper limit, and whether the requested answer is per claim or per exposure.

Do not confuse the attachment with the layer width. A layer attaching at $100,000 and extending to $250,000 has width $150,000, but its upper limit is $250,000.
Do not report the upper-limit capped loss as the layer cost. The amount $C(b)$ includes all covered payment from the ground up. Subtract $C(a)$.
Do not subtract incompatible factors. The factors must use the same basic limit and consistent loss and claim definitions.
Keep units straight. An ILF difference times a basic-limit expected loss is a per-claim amount. Multiply by frequency only when a per-exposure cost is requested.
Check the direction and scale. For $a\leq b$, $C(b)-C(a)$ cannot be negative. A layer cost should not exceed the layer width per claim, because $Y_{a,b}\leq b-a$.

For a sequence of adjacent layers, the costs telescope: the cost from $a$ to $c$ equals the cost from $a$ to $b$ plus the cost from $b$ to $c$, whenever $a\leq b\leq c$. This is another useful check on a table of incremental costs.

Check Your Understanding

For each hypothetical situation, calculate the requested layer cost. Unless a severity distribution is specified, use only the supplied expected capped losses or ILFs. Round dollar answers to cents.

A program has a $600 expected capped loss per claim at a $100,000 basic limit. The ILF at $300,000 is 1.25. Expected frequency is 0.12 claims per exposure. Find the expected cost per exposure for the $100,000-to-$300,000 layer. (Answer: $18.00 per exposure)
Ground-up severity is exponential with mean $\theta=\$40{,}000$. Expected claim frequency is 0.03 per exposure. Find the expected cost per exposure for the layer attaching at $20,000 and extending to $80,000. Use $C(u)=\theta(1-e^{-u/\theta})$. (Answer: $565.43 per exposure)
A $200,000 basic limit has expected capped loss of $250 per claim. The ILFs are 1.28 at $400,000 and 1.52 at $1,000,000. Expected frequency is 0.06 claims per exposure. Find the expected cost per exposure for the $400,000-to-$1,000,000 layer. (Answer: $3.60 per exposure)
A hypothetical policy’s expected capped losses are $1,120 per claim at $50,000 and $1,780 per claim at $200,000. The layer attaches at $50,000, has a $150,000 layer limit, and expected frequency is 0.04 per exposure. Find the cost per claim and per exposure. (Answer: $660.00 per claim; $26.40 per exposure)
A program has an expected capped loss of $900 per claim at its $50,000 basic limit. ILFs at $150,000, $300,000, and $600,000 are 1.22, 1.39, and 1.51. Expected frequency is 0.025 per exposure. Find the $150,000-to-$600,000 layer cost per claim and per exposure, and verify by splitting at $300,000. (Answer: $261.00 per claim; $6.53 per exposure; sublayer costs are $153.00 and $108.00 per claim)

What You'll Learn