Loss-Related Expense Provisions

Why the Expense Base Matters

A rate indication must account for expenses as well as expected losses. The expense provision depends on how the expense behaves and on the base used to express it. A commission that varies with premium is naturally measured against premium. A cost that varies with claim losses instead calls for a loss-based measure. Using the wrong denominator can produce a number that looks like a reasonable ratio but does not represent the expense relationship being estimated.

This tutorial focuses on expenses that are linked to losses, including claim adjustment costs. It explains how to estimate a loss-related ratio and translate it into a dollar provision. The amounts in the examples are hypothetical. No discounting, mortality assumption, or explicit probability distribution is needed: the examples use selected aggregate loss and expense amounts on a consistent basis.

Definition: A loss-related expense is an expense whose amount is expected to vary with the level or handling of losses. A common measure is the loss-related expense ratio, $r_L$, defined as selected loss-related expenses divided by the corresponding selected losses. It is a dimensionless ratio, not a dollar amount.

Loss adjustment expenses are a prominent example. Allocated loss adjustment expense (ALAE) is associated with particular claims, while unallocated loss adjustment expense (ULAE) is not assigned to individual claims in the same way. Both may be considered in estimating claim-related costs, but their observed relationship to losses can differ. Other costs may also be linked to claim volume or loss activity. Whether a cost belongs in a loss-related provision depends on its behavior and the ratemaking basis being used, not merely on the name of the account.

Measuring a Loss-Related Provision

Let $C_L$ denote the selected amount of loss-related expense and $L$ the corresponding selected losses. The empirical ratio is $r_L=C_L/L$, provided $L>0$. The numerator and denominator should refer to compatible experience: for example, the same coverage, period, and valuation basis. If the loss measure has been adjusted to a prospective basis, the expense measure should be selected consistently for that purpose as well.

r_L=\frac{C_L}{L}

If a rate indication has an expected loss amount $L_{\text{ind}}$ and the selected ratio is considered applicable, the indicated loss-related expense provision is $r_L L_{\text{ind}}$. For an exposure base $N$, if expected losses per exposure are $l=L_{\text{ind}}/N$, then the expense provision per exposure is $r_L l$. Multiplying that per-exposure amount by $N$ recovers the total provision.

\text{Expense provision}=r_L L_{\text{ind}}, \qquad \text{expense per exposure}=r_L\frac{L_{\text{ind}}}{N}

The selected ratio is an estimate of an expense relationship, not a claim that every dollar of expense is mechanically proportional to every dollar of loss. A proportional approximation is most defensible when the historical relationship is reasonably stable and the experience is relevant to the prospective period. Fixed claim-handling costs, changes in litigation practices, changes in claim mix, or expense caps can make a single ratio less representative.

Derivation: Why the Ratio Scales the Loss Amount

The conversion from a selected ratio to a provision follows directly from the proportional model. Suppose, as an approximation, loss-related expense $C_L$ is $r_L$ times losses $L$. For a prospective loss amount $L_{\text{ind}}$, applying the same relationship gives the expense provision $r_L L_{\text{ind}}$. Per exposure, divide the total by $N$. Since $l=L_{\text{ind}}/N$, the result is $r_L l$.

C_L=r_L L \quad\Longrightarrow\quad C_{L,\text{ind}}=r_L L_{\text{ind}} \quad\Longrightarrow\quad \frac{C_{L,\text{ind}}}{N} =r_L\frac{L_{\text{ind}}}{N} =r_L l

This argument uses a proportional relationship as the modeling assumption. It does not establish that expenses must be proportional in every portfolio. If expense amounts are mainly fixed per claim, for instance, a simple ratio to loss dollars may change when average claim severity changes. In that situation, the ratio may need refinement or another measure of the expense relationship.

Worked Example: Applying a Selected Ratio Per Exposure

A hypothetical personal property program has selected prospective losses of $1,500,000 for 12,000 exposures. Related claim expenses selected on the same basis are $225,000. Estimate the loss-related ratio, the expected loss per exposure, and the loss-related provision per exposure and in total.

Method and basis. Use the loss-related expense ratio, with expenses and losses from the same selected basis. No discounting or claim distribution is specified or required.

Step 1: Calculate the ratio.

r_L=\frac{\$225{,}000}{\$1{,}500{,}000}=0.15

Step 2: Calculate expected losses per exposure.

l=\frac{\$1{,}500{,}000}{12{,}000}=\$125

Step 3: Calculate the expense provision per exposure.

r_L l=0.15(\$125)=\$18.75

Step 4: Check the total. Multiplying the per-exposure provision by the exposure count gives $12{,}000(\$18.75)=\$225{,}000$. This matches the selected historical expense amount because the selected prospective loss amount equals the loss amount used to calculate the ratio.

Common mistake avoided: The ratio is $0.15$, not a dollar provision per exposure. The dollar provision per exposure is $18.75, found by applying the ratio to the $125 expected loss per exposure.

Combining Experience Without Misweighting It

When experience from multiple periods is combined, the aggregate ratio is total selected expense divided by total selected loss. It is therefore a loss-weighted average of the period ratios. A simple arithmetic average gives each period equal weight regardless of its loss volume, and generally does not reproduce the aggregate relationship.

r_L=\frac{\sum_j C_{L,j}}{\sum_j L_j} =\sum_j\left(\frac{L_j}{\sum_k L_k}\right) \left(\frac{C_{L,j}}{L_j}\right)

The equality follows because multiplying each period ratio $C_{L,j}/L_j$ by its loss weight $L_j/\sum_k L_k$ gives $C_{L,j}/\sum_k L_k$. Summing those terms produces total expense divided by total loss. The formula assumes each period has a positive loss amount and that the selected expenses and losses are comparable across periods.

Worked Example: Combining Two Periods Correctly

A hypothetical liability portfolio has the following experience: Period A has $500,000 of losses and $100,000 of loss-related expenses; Period B has $1,500,000 of losses and $180,000 of loss-related expenses. Find each period’s ratio and the combined ratio. Explain why the arithmetic average is not the appropriate combined result.

Method and basis. Calculate the aggregate ratio from total expenses divided by total losses. No discounting or claim distribution is specified or required.

Step 1: Calculate the period ratios.

r_{L,A}=\frac{\$100{,}000}{\$500{,}000}=0.20, \qquad r_{L,B}=\frac{\$180{,}000}{\$1{,}500{,}000}=0.12

Step 2: Aggregate the amounts. Total expenses are $\$100{,}000+\$180{,}000=\$280{,}000$. Total losses are $\$500{,}000+\$1{,}500{,}000=\$2{,}000{,}000$.

Step 3: Divide aggregate expense by aggregate loss.

r_L=\frac{\$280{,}000}{\$2{,}000{,}000}=0.14

As a check, Period A represents $500{,}000/2{,}000{,}000=0.25$ of the losses and Period B represents $1{,}500{,}000/2{,}000{,}000=0.75$. The loss-weighted ratio is $0.25(0.20)+0.75(0.12)=0.05+0.09=0.14$.

Common mistake avoided: The simple average, $(0.20+0.12)/2=0.16$, is not the aggregate ratio because it gives the smaller-loss period the same weight as the larger-loss period.

Keep Premium-Related Loads Separate

A premium-related expense is measured relative to premium; a loss-related expense is measured relative to losses. These are different relationships and should not be interchanged. For example, a commission that is defined as a proportion of premium belongs in a premium-based calculation. Dividing a claim expense by premium instead of losses produces an expense-to-premium ratio, which answers a different question.

In a broader rate calculation, the expense provision can be expressed on a loss basis and then combined with other rate components using a consistent formulation. The calculation here identifies only the loss-related provision; it does not convert premium-related expenses into loss loads or prescribe how expenses are allocated among rate components.

Expense relationship	Typical measurement base	Question answered
Loss-related	Losses	How much expense is associated with the selected loss amount?
Premium-related	Premium	How much expense is associated with the selected premium amount?
Fixed per exposure or policy	Exposure or policy count	How much expense is assigned to each unit of business?

Worked Example: Avoiding the Wrong Denominator

A hypothetical commercial equipment insurer selects $162,000 of claim-related expense and $1,350,000 of corresponding losses. Earned premium for the same experience is $3,000,000. A prospective indication uses 2,500 exposures and expected losses of $72 per exposure. Calculate the loss-related ratio and prospective provision, and explain what the premium denominator would measure.

Method and basis. Use expense divided by corresponding losses to estimate the loss-related ratio, then apply it to prospective losses. No discounting or claim distribution is specified or required.

Step 1: Calculate the loss-related ratio.

r_L=\frac{\$162{,}000}{\$1{,}350{,}000}=0.12

Step 2: Calculate prospective losses.

L_{\text{ind}}=2{,}500(\$72)=\$180{,}000

Step 3: Apply the selected ratio.

C_{L,\text{ind}}=0.12(\$180{,}000)=\$21{,}600

The provision per exposure is $\$21{,}600/2{,}500=\$8.64$. As a check, $0.12(\$72)=\$8.64$ per exposure, which gives the same total when multiplied by 2,500.

Dividing the historical expense by premium gives $\$162{,}000/\$3{,}000{,}000=0.054$. That is an expense-to-premium ratio, not the loss-related ratio requested. Applying it as a loss-related ratio would use the wrong denominator and misstate the intended expense relationship.

Common mistake avoided: Do not substitute premium merely because it is available in the data. Select the denominator that matches the expense relationship being estimated.

When a Simple Ratio Needs Care

A loss-related ratio is useful only to the extent that the selected experience represents the prospective business. Before applying it, check that:

The expense numerator includes only expenses intended to be treated as loss-related.
The losses and expenses refer to compatible periods, coverage, and valuation bases.
The experience has been selected consistently for prospective use; an observed historical relationship may not remain stable if conditions change.
The expense does not have a large fixed or threshold component that would make a proportional relationship to losses misleading.
The loss denominator is positive. If selected losses are zero, the ratio is undefined and cannot be calculated by division.

A common exam trap is to label a ratio by its numerator and overlook its denominator. Always write the fraction before calculating: loss-related expense divided by loss for $r_L$; premium-related expense divided by premium for a premium-based ratio. Then check the units. A ratio has no dollar units, while a provision per exposure has dollars per exposure.

Key takeaway: Select the expense category and denominator deliberately. A loss-related ratio is $C_L/L$; applying it to expected losses gives a loss-related dollar provision. Keep that calculation distinct from premium-based expense loads and check whether a proportional relationship is appropriate.

Check Your Understanding

For each hypothetical problem, identify the appropriate base and calculate the requested loss-related measure.

A small specialty portfolio has $48,000 of selected loss-related expenses and $400,000 of corresponding losses. Find $r_L$. (Answer: 0.12)
A hypothetical book has 1,800 exposures and prospective losses of $90 per exposure. The selected loss-related ratio is 0.08. Find the provision per exposure and in total. (Answer: $7.20 per exposure; $12,960 total)
Period A has $300,000 losses and $45,000 expenses; Period B has $900,000 losses and $153,000 expenses. Find the combined ratio using aggregate amounts. (Answer: 0.165)
A hypothetical insurer has $81,000 of loss-related expenses, $900,000 of corresponding losses, and $1,500,000 of premium. Find the loss-related ratio and the expense-to-premium ratio, respectively. (Answer: 0.09 and 0.054)
A prospective indication has 4,000 exposures and expected losses of $125 per exposure. The applicable selected loss-related ratio is 0.14. Find the total loss-related provision and confirm it using the per-exposure amount. (Answer: $70,000 total; $17.50 per exposure)

What You'll Learn