Pure Premium Indications With Changing Exposure Volume

Why Exposure Volume and Loss Cost Must Be Separated

Total losses can rise because an insurer covers more exposure, because expected losses per exposure have increased, or because both changed. A pure premium indication must separate these effects. Exposure volume helps determine the aggregate amount of loss to project; loss cost per exposure is the quantity that informs a loss-only rate per exposure.

For example, if a hypothetical book doubles its number of exposure units while expected loss per unit stays constant, its total expected losses double. That does not mean the indicated loss cost per unit has doubled. Conversely, total losses might stay level even as the loss cost rises, if exposure volume falls enough to offset the increase.

The calculations below treat losses as ultimate losses and exposures as measured on the same, consistent exposure base. They are deterministic illustrations: no interest discounting, mortality assumption, or claim probability distribution is used. No expense provision is included, so the indicated amount is a loss-only pure premium, not necessarily a complete charged rate.

Definition: Let $E$ denote exposure volume and $L$ denote the corresponding ultimate losses. The pure premium, or loss cost per exposure, is $L/E$, when $E>0$. Thus $L=E(L/E)$. Exposure volume measures how much insurance was provided; pure premium measures expected loss per unit of that insurance.

Separating the Two Sources of Loss Change

Compare two periods with exposures $E_1$ and $E_2$, losses $L_1$ and $L_2$, and pure premiums $C_1=L_1/E_1$ and $C_2=L_2/E_2$. Then total losses in each period equal exposure volume multiplied by loss cost. Their ratio therefore separates exactly into the exposure ratio and the loss-cost ratio:

\frac{L_2}{L_1} = \frac{E_2}{E_1}\times\frac{C_2}{C_1}

This identity explains why the percentage change in total losses is not, by itself, a rate indication. It combines changes in volume and in loss cost. To identify a loss-only rate per exposure, compare loss costs on a consistent basis, not unadjusted aggregate losses from periods with different exposure volumes.

It is also possible to express the dollar change in losses as a volume component and a loss-cost component. One useful convention is to change exposure first, holding the first-period loss cost fixed, and then change loss cost at the second-period exposure level:

$$ L_2-L_1 = (E_2-E_1)C_1+E_2(C_2-C_1) $$

To verify the decomposition, substitute $L_1=E_1C_1$ and $L_2=E_2C_2$. The right side becomes $E_2C_1-E_1C_1+E_2C_2-E_2C_1=E_2C_2-E_1C_1=L_2-L_1$. The first term is the change attributable to volume at the initial loss cost; the second is the change attributable to loss cost at the ending volume. This order assigns the interaction between volume and loss cost to the second term. Reversing the order would produce a different allocation between components, but the total change would still be $L_2-L_1$.

Worked Example: Decomposing an Increase in Total Losses

A hypothetical commercial equipment portfolio has 8,000 exposure units and $1,440,000 of ultimate losses in the first period. In the second period it has 10,000 exposure units and $2,000,000 of ultimate losses. Calculate each period’s loss cost and decompose the increase in aggregate losses using the convention that volume changes first. Assume the losses are deterministic ultimate amounts, with no interest adjustment or mortality or claim distribution model.

Step 1: Calculate the loss costs. For period 1, $C_1=\$1{,}440{,}000/8{,}000=\$180$ per exposure. For period 2, $C_2=\$2{,}000{,}000/10{,}000=\$200$ per exposure.

Step 2: Calculate the total loss increase. Losses increased by $\$2{,}000{,}000-\$1{,}440{,}000=\$560{,}000$.

Step 3: Calculate the volume component. Exposure increased by 2,000 units. At the initial loss cost of $180, the volume component is $2{,}000\times\$180=\$360{,}000$.

Step 4: Calculate the loss-cost component. Loss cost increased by $20 per exposure. At the ending volume of 10,000 exposures, the component is $10{,}000\times\$20=\$200{,}000$.

Check: $\$360{,}000+\$200{,}000=\$560{,}000$, the observed total increase. The loss-only cost per exposure rose from $180 to $200, or $200/180-1=11.11\%$, even though aggregate losses rose by $560{,}000/1{,}440{,}000=38.89\%$. The larger aggregate percentage includes exposure growth. The method avoids treating the entire 38.89% increase as a per-exposure rate change.

Combining Experience With Different Exposure Volumes

When experience from several periods is combined, the aggregate pure premium is total losses divided by total exposures. This is an exposure-weighted average of the period loss costs: periods with more exposure receive more weight. A simple arithmetic average of period loss costs is appropriate only when the periods have equal exposure volumes, or when equal weighting is specifically intended for another reason.

C_{\mathrm{combined}} = \frac{\sum_j L_j}{\sum_j E_j} = \frac{\sum_j E_j C_j}{\sum_j E_j}

If loss costs must first be trended to a common cost level, trend each period’s loss cost (or its losses) before combining the periods. Then divide the adjusted total losses by the corresponding total exposures. The denominator remains exposure, not premium. Before interpreting the result, confirm that the exposure units and coverage are consistent across the periods.

Worked Example: Combining Periods With Unequal Exposure

A hypothetical data-processing liability portfolio has 4,000 exposures and $720,000 of ultimate losses in an earlier period, and 6,000 exposures and $1,260,000 of ultimate losses in a later period. Assume a deterministic annual loss-cost trend of 4%, applied once to the earlier period to bring it to the later period’s cost level. Exposures are already on a consistent basis. There is no interest adjustment, mortality assumption, or claim distribution model. Find the exposure-weighted indicated loss cost and the projected losses at 12,500 future exposures.

Step 1: Calculate the unadjusted period loss costs. The earlier period has $\$720{,}000/4{,}000=\$180$ per exposure. The later period has $\$1{,}260{,}000/6{,}000=\$210$ per exposure.

Step 2: Trend the earlier loss cost. Applying the 4% trend once gives $\$180\times1.04=\$187.20$ per exposure. On the 4,000 earlier exposures, this corresponds to adjusted losses of $4{,}000\times\$187.20=\$748{,}800$. The later-period losses remain $1,260,000.

Step 3: Combine using exposure weighting. Adjusted losses total $\$748{,}800+\$1{,}260{,}000=\$2{,}008{,}800$. Exposures total $4{,}000+6{,}000=10{,}000$. The indicated loss cost is therefore $\$2{,}008{,}800/10{,}000=\$200.88$ per exposure.

Step 4: Project aggregate losses. At 12,500 projected exposures, expected losses are $12{,}500\times\$200.88=\$2{,}511{,}000$.

Check: Dividing $2,511,000 by 12,500 recovers $200.88 per exposure. The unweighted average of the adjusted period loss costs, $(\$187.20+\$210)/2=\$198.60$, is not the combined pure premium here because the periods have different exposure volumes. The exposure-weighted method avoids giving the smaller period the same weight as the larger one.

From Projected Loss Cost to a Pure Premium Indication

A practical calculation separates the selection of projected loss cost from the projection of exposure volume. First establish a loss cost per exposure using losses and exposures on consistent coverage and cost-level bases. Then multiply that selected loss cost by projected exposures to estimate aggregate losses. Dividing those projected losses by projected exposures returns the selected loss cost.

Establish the experience loss cost.
Divide applicable ultimate losses by their matching exposure volume; adjust the loss basis for any required trend or other stated projection.

Select the projected exposure volume.
Use the projected number of exposure units on the same base as the experience denominator.

Calculate projected aggregate losses.
Multiply selected loss cost per exposure by projected exposures.

State the loss-only indication.
Report the selected loss cost per exposure. If comparing it with a current loss-only rate, use the ratio of projected loss cost to that comparable current rate.

This procedure also clarifies when exposure growth matters. If projected loss cost is unchanged, a larger projected exposure volume increases projected aggregate losses but does not change the loss-only indication per exposure. If the task asks for total projected losses, volume is essential. If it asks for the per-exposure loss cost, volume is the denominator that normalizes the total.

A percentage indication relative to a current loss-only rate $R_0$ is $C_{\mathrm{projected}}/R_0-1$, provided both amounts use the same exposure unit, coverage, and cost level. This comparison does not by itself account for expenses, profit, or other rate components. In particular, it should not be interpreted as a complete premium-rate indication when those provisions have not been considered.

Worked Example: Projecting Losses Without Mistaking Volume for Rate Change

A hypothetical renewable-energy equipment program has 6,000 exposure units and $1,080,000 of ultimate losses in its experience period. Assume a deterministic annual loss-cost trend of 10% for two years, unchanged coverage, and projected volume of 9,000 exposures. A comparable current loss-only rate is $200 per exposure. There is no interest adjustment, mortality assumption, or claim distribution model. Calculate projected aggregate losses and the loss-only percentage indication.

Step 1: Calculate the historical loss cost. The experience loss cost is $\$1{,}080{,}000/6{,}000=\$180$ per exposure.

Step 2: Project the loss cost for two years. The two-year trend factor is $1.10^2=1.21$. Projected loss cost is $\$180\times1.21=\$217.80$ per exposure.

Step 3: Project total losses using projected volume. Expected losses are $9{,}000\times\$217.80=\$1{,}960{,}200$.

Step 4: Compare the loss cost with the current rate. The loss-only indication factor is $\$217.80/\$200=1.089$. Thus the indicated change is $1.089-1=0.089$, or 8.9%.

Check: Projected aggregate losses are 1.815 times historical losses: $\$1{,}960{,}200/\$1{,}080{,}000=1.815$. This equals the exposure ratio $9{,}000/6{,}000=1.5$ times the loss-cost ratio $\$217.80/\$180=1.21$. The aggregate increase is 81.5%, but the loss-only indication relative to the stated current rate is 8.9%. Comparing total losses directly would confuse projected exposure growth and trend with the per-exposure rate change.

Common Exam Trap: Comparing Aggregate Losses Across Different Volumes

Suppose aggregate losses rise 30% while exposure volume rises 20%. It is incorrect to conclude that loss cost rose 30%. The implied loss-cost factor is $1.30/1.20=1.0833$, so loss cost rose about 8.33%, not 30%, assuming the exposure and loss bases are consistent. Similarly, a decline in total losses does not necessarily indicate a lower pure premium if exposure volume declined by even more.

A related error is to multiply projected loss cost by projected exposures and then use projected losses alone as the indicated rate. The resulting amount is aggregate losses, not a rate per exposure. Divide by projected exposures to recover the loss cost. Also avoid averaging period pure premiums equally when exposures differ; use total adjusted losses over total corresponding exposures.

Key takeaway: Separate the numerator’s scale from its per-unit level. Exposure volume scales aggregate losses; pure premium normalizes losses by exposure. A sound indication uses consistent losses and exposures, exposure-weighted experience where appropriate, and a clearly stated projected exposure base.

Check Your Understanding

Work each problem before checking the answer shown at the end of the question. Treat all loss amounts as deterministic ultimate losses; no interest, mortality, or claim distribution assumption is required.

A hypothetical repair plan has 5,000 exposures and $850,000 of ultimate losses. At a projected 7,000 exposures with no change in loss cost, what are the indicated loss cost per exposure and projected aggregate losses? (Answer: $170 per exposure; $1,190,000 aggregate losses.)
A hypothetical portfolio has 3,000 exposures and $510,000 of losses in one period, and 5,000 exposures and $950,000 of losses in another. With no trend adjustment, calculate the combined exposure-weighted loss cost. (Answer: $182.50 per exposure.)
A hypothetical experience set has 4,000 exposures and $960,000 of ultimate losses. Apply a deterministic annual loss-cost trend of 3% for two years, then project 6,000 exposures. What are the projected loss cost and total losses? (Answer: $254.62 per exposure and $1,527,696 in projected losses, with loss cost rounded to cents.)
Exposure volume increases by 20%, while loss cost per exposure decreases by 5%. By what percentage do aggregate losses change? (Answer: Aggregate losses increase by 14%, since $1.20\times0.95=1.14$.)
A hypothetical book’s aggregate losses increase from $2,400,000 to $2,880,000 while exposure increases from 12,000 to 16,000 units. Calculate the change in loss cost per exposure and explain why the aggregate loss comparison alone is misleading. (Answer: Loss cost decreases from $200 to $180 per exposure, a 10% decrease; aggregate losses rose 20% because exposure volume grew 33.33%.)

What You'll Learn