Projecting the Exposure Basis to the Future Period

Why Project the Exposure Basis?

A pure premium is a loss amount per exposure unit. The relevant denominator is therefore not automatically the number of exposures in the historical data. It is the exposure volume that corresponds to the losses being divided and, for a prospective indication, to the future period and book being considered. If the projected losses represent a larger or smaller expected book than the historical book, the denominator must be selected on a consistent basis.

This is a volume adjustment, not a loss trend adjustment. The earlier material on pure premium indications establishes the loss-per-exposure calculation. Here, the focus is how to project the exposure volume so that the numerator and denominator describe the same prospective basis. A change in the number of exposure units can change total projected losses even when the expected loss cost per unit is unchanged.

Definition: The prospective exposure basis is the expected amount of the selected exposure unit during the future period to which the indicated rate will apply. A pure premium on that basis is projected losses divided by projected exposures, provided both quantities refer to the same period, coverage, population, and exposure unit.

An exposure unit might be a vehicle-year, policy-year, insured location-year, or a stated amount of payroll. The exposure basis must be stated explicitly: “projected exposure” is not enough unless the unit and period are clear. For example, 8,000 vehicle-years and 8,000 policies are not interchangeable if policies have different durations or different numbers of insured vehicles.

A projection commonly starts with historical exposure volume and applies a forecast of volume change. If expected annual growth is $g$ for $t$ years, with growth compounded annually, then the projected volume is:

E_{\text{pros}}=E_{\text{hist}}(1+g)^t

Here $E_{\text{hist}}$ is the exposure volume on the starting basis, and $E_{\text{pros}}$ is the projected volume on the future basis. This formula assumes a constant annual compound growth rate over the interval. If growth is specified separately for each year, apply each year’s factor in sequence. The formula does not apply as written when the stated growth is a total, noncompounded change over the whole interval.

Project the Volume, Then Match the Losses

The essential consistency test is whether the projected loss amount and the exposure denominator describe the same expected book. If the loss estimate is for 10,000 exposure units, dividing it by 12,000 projected units would understate the loss cost per exposure. Conversely, dividing a loss estimate for a 12,000-unit book by only 10,000 units would overstate it.

Exposure growth alone does not imply that the pure premium per exposure has changed. Under the assumption that expected loss cost per exposure remains constant, total expected losses grow in proportion to exposure volume. This gives a useful check on projections.

\frac{L_{\text{hist}}}{E_{\text{hist}}}=p \quad\Longrightarrow\quad L_{\text{pros}}=pE_{\text{pros}} \quad\Longrightarrow\quad \frac{L_{\text{pros}}}{E_{\text{pros}}}=p

The equality follows because the prospective loss amount is assumed to equal the same per-exposure loss cost $p$ multiplied by the prospective exposure count. Thus, increasing volume changes the aggregate loss dollars and the denominator by the same factor; it does not, by itself, change the loss cost per exposure. This is a volume-only illustration. If expected losses per exposure also change, the loss projection must reflect that separately.

In practice, an actuary may receive a prospective loss estimate directly, or may forecast exposure volume from policy counts, average insured units, and expected time in force. Whatever the approach, the numerator and denominator must be reconciled before calculating the pure premium.

Specify the exposure unit and period.
For example, identify whether the basis is earned vehicle-years during a calendar year or another defined unit.

Establish the starting volume.
Use the relevant historical or current exposure volume, not premium as a substitute for exposure.

Apply the volume forecast.
Use the stated growth assumption, policy-count forecast, or other exposure forecast method.

Check the loss basis.
Confirm that the projected losses describe the same book and future period as the projected exposures.

Calculate and interpret the pure premium.
Divide projected losses by projected exposures and report the result per stated exposure unit.

Worked Applications

Worked Example: Compound Growth in Vehicle-Years

A hypothetical personal auto program has 12,000 earned vehicle-years on its starting basis. For planning purposes, the exposure volume is expected to grow by 5% per year for two years, compounded annually. The projected loss estimate for the prospective two-year book is $1,984,500. Calculate the prospective exposure volume and pure premium. Assume the same coverage and exposure definition apply throughout; no interest, mortality, or probability distribution assumption is needed for this deterministic calculation.

Step 1: Project the exposures. The two annual growth factors are each $1.05$, so:

E_{\text{pros}}=12{,}000(1.05)^2 =12{,}000(1.1025) =13{,}230\text{ vehicle-years}

Step 2: Divide projected losses by projected exposures.

p_{\text{pros}}=\frac{\$1{,}984{,}500}{13{,}230} =\$150.00\text{ per vehicle-year}

Check. A $150 loss cost applied to 13,230 vehicle-years gives $13{,}230\times \$150=\$1{,}984{,}500$, matching the projected losses. A common mistake is to use the original 12,000 vehicle-years as the denominator, producing $\$1{,}984{,}500/12{,}000=\$165.375$. That calculation mismatches prospective losses with starting-period exposures. The method here avoids that denominator error.

Worked Example: Forecasting Exposure From Policy Counts

A hypothetical equipment insurer expects to write 4,500 policies during a future year. Based on the planned timing of new business and cancellations, each policy is expected to contribute an average of 0.82 earned policy-years during that year. The projected losses for that same book and year are $738,000. Calculate the projected exposure volume and the pure premium. Assume the average earned policy-year estimate already accounts for the policy durations; do not apply an additional duration adjustment.

Step 1: Convert expected policy count to exposure units. The expected earned exposure is:

E_{\text{pros}}=4{,}500\text{ policies}\times 0.82 =3{,}690\text{ earned policy-years}

The factor 0.82 is exposure per policy, not a loss adjustment. Multiplying by it converts the count of policies into earned policy-years.

Step 2: Calculate the pure premium.

p_{\text{pros}}=\frac{\$738{,}000}{3{,}690} =\$200.00\text{ per earned policy-year}

Check. $3{,}690\times \$200=\$738{,}000$. The result is in dollars per earned policy-year, not per policy written. Treating all 4,500 policies as full-year exposures would instead give $\$738{,}000/4{,}500=\$164.00$, which uses a different denominator and ignores the stated average duration. The method avoids treating a policy count as though it were automatically an earned exposure count.

Worked Example: Reconcile a Prospective Loss Estimate to a Volume Forecast

A hypothetical commercial property portfolio has 10,000 insured location-years on its starting basis. The prospective period is expected to have 11,500 location-years. The selected prospective loss cost is $150 per location-year, and the projected aggregate losses are stated to be consistent with that cost and the forecast volume. Calculate the aggregate loss projection and verify the pure premium. Assume the exposure unit, coverage, and loss basis are unchanged; no interest, mortality, or probability distribution assumption is needed.

Step 1: Calculate aggregate prospective losses.

L_{\text{pros}}=11{,}500\times \$150 =\$1{,}725{,}000

Step 2: Verify the pure premium on the prospective basis.

p_{\text{pros}}=\frac{\$1{,}725{,}000}{11{,}500} =\$150.00\text{ per location-year}

Check. The calculated pure premium equals the selected $150 per location-year. If the starting 10,000 location-years were mistakenly used, the result would be $\$1{,}725{,}000/10{,}000=\$172.50$. That is not evidence of a higher loss cost: it is the result of dividing losses for 11,500 location-years by only 10,000. This example illustrates the volume-only assumption; it does not assert that loss cost will remain constant when other conditions change.

Common Traps and Interpretation

The projected exposure denominator is a substantive selection, not just a convenient number in a spreadsheet. A reliable calculation makes the basis visible and asks whether the exposure count belongs to the same book and period as the projected losses.

Using historical exposure volume with prospective losses. This mismatches the numerator and denominator when the future book is expected to have a different volume.
Confusing policies with exposure units. A policy count may need to be converted using expected duration, insured units, or another stated exposure measure.
Applying growth for the wrong interval. Two years of 5% annual compounded growth gives a factor of $1.05^2$, not $1+2(0.05)$ unless the assumption explicitly specifies simple, noncompounded growth.
Applying volume growth twice. If projected losses already reflect the forecast volume, do not scale them by the same growth factor again without a reason.
Interpreting more total losses as a higher pure premium. A larger portfolio may have larger aggregate losses solely because it has more exposure units. Compare losses per consistent exposure unit to assess loss cost.

A practical audit is to multiply the calculated pure premium by the selected prospective exposures. The product should reproduce the projected loss amount, subject only to stated rounding. If it does not, check the units, time interval, growth factor, and whether the losses were already projected to the forecast book.

Key takeaway: Project the exposure volume that corresponds to the prospective losses and rate period. A change in portfolio volume affects total losses and total exposures together; only a consistent numerator and denominator yield an interpretable pure premium per exposure unit.

Check Your Understanding

For each problem, identify the prospective exposure denominator before calculating or interpreting the pure premium.

A hypothetical liability portfolio has 9,000 earned unit-years and is forecast to grow by 4% over the next year. Projected losses for that prospective book are $1,404,000. Find the prospective exposures and pure premium. (Answer: 9,360 unit-years; $150.00 per unit-year)
A hypothetical insurer expects 2,500 policies, each contributing an average of 1.10 earned policy-years during the prospective period. Projected losses are $412,500. Find the exposure volume and pure premium. (Answer: 2,750 earned policy-years; $150.00 per earned policy-year)
A hypothetical product starts with 7,500 exposure units. Volume is forecast to increase by 6% for one year and then by 2% for a second year, with annual compounding. Projected losses for the resulting prospective book are $1,008,000. Find the projected exposure volume and pure premium. (Answer: 8,109 exposure units; approximately $124.31 per exposure unit)
A hypothetical portfolio has 18,000 earned exposures. Expected volume grows by 8% in the first year and 5% in the second year, compounded annually. Projected losses for the prospective book are $3,061,800. Find the projected exposures and pure premium. (Answer: 20,412 exposures; $150.00 per exposure)
A hypothetical book is expected to grow from 10,000 to 11,500 exposure units. A prospective loss estimate of $1,725,000 is already for the larger book. Calculate the correctly matched pure premium and the result from mistakenly dividing by the starting volume. Explain the difference. (Answer: $150.00 per exposure correctly; $172.50 using the mismatched 10,000-exposure denominator, an overstatement caused by the mismatch)

What You'll Learn