Deriving the Premium-Based Expected Loss Formula

Introduction

An expected loss ratio is a proportion, while expected ultimate losses are a dollar amount. The premium-based expected loss formula connects the two: multiply the earned premium for a defined book of business by the expected loss ratio applicable to that same book. The arithmetic is simple; the important work is making sure the premium and ratio refer to a consistent basis.

This formula is used in expected claims methods to develop an a priori estimate of ultimate losses. It does not require observed losses to be complete or mature. Instead, it translates a premium base and an expected loss ratio into a dollar estimate. The method is aligned with the CAS Exam 5 (Basic Ratemaking and Estimated Claim Liabilities) syllabus.

Definition: Let $P$ be earned premium for a specified portfolio and period, and let $U$ be the corresponding ultimate losses on the loss basis being estimated. The expected loss ratio $r$ is the expected ultimate loss amount per dollar of that earned premium. When the premium amount is fixed and the expected loss ratio applies to the same portfolio, $r=E[U]/P$, so the premium-based expected loss estimate is $E[U]=P r$.

Deriving the Formula

The derivation begins with the meaning of the expected loss ratio, not with a separate claim-count or claim-severity model. Suppose the earned premium $P$ is known for the cohort being estimated. The expected loss ratio $r$ is defined as the expected ultimate loss amount divided by that premium:

r=\frac{E[U]}{P}.

Multiplying both sides by $P$ gives:

$$ E[U]=P r. $$

The multiplication is valid because $P$ is the fixed premium base in this definition. If the expected loss ratio is 0.65, it means an expected 65 cents of loss for each dollar of the applicable earned premium. Thus, for $1,000,000 of such premium, the implied expected losses are $650,000.

The result can also be understood conditionally when premium varies across possible outcomes. If the assumption is $E[U\mid P]=rP$, then taking expectations on both sides gives $E[U]=rE[P]$, provided $r$ is constant under that assumption. In an estimate for a specified cohort, the usual calculation uses the cohort’s known or estimated earned premium as its premium base. One must not silently treat a random premium and a fixed premium as interchangeable.

Assumptions and scope: The premium and loss ratio must refer to the same portfolio, coverage, period, and loss basis. The ratio must be appropriate for the premium being multiplied. The formula produces an expected ultimate loss estimate; it does not assert that the eventual realized losses will equal that amount. The formula itself specifies no claim-size distribution, interest rate, or mortality assumption.

Units and Basis Checks

The units provide a quick check on the calculation. Earned premium is measured in dollars, and a loss ratio is a dimensionless proportion. Their product is therefore a dollar amount:

\text{dollars of earned premium} \times \frac{\text{expected dollars of loss}}{\text{dollars of earned premium}} = \text{expected dollars of loss}.

A loss ratio of 68% must be entered as 0.68, not 68. If premium is reported in thousands of dollars, the result is also in thousands until converted to dollars. The calculation cannot repair a mismatch in basis: multiplying a ratio for one coverage or loss definition by premium for another does not become valid merely because the units appear to cancel.

When separate segments have different expected loss ratios, calculate each segment’s expected loss amount using its own premium and ratio, then add those amounts. Equivalently, the overall loss ratio is the premium-weighted average of the segment ratios. It is generally not the simple arithmetic average unless the segment premiums are equal.

E[U_{\text{total}}]=\sum_{j=1}^{k} P_j r_j, \qquad r_{\text{overall}}= \frac{\sum_{j=1}^{k} P_j r_j}{\sum_{j=1}^{k}P_j}.

Worked Applications

The examples treat each supplied expected loss ratio as an input. They do not select or justify that ratio; the calculation is about translating it into expected ultimate loss dollars. In each example, the premium and ratio are stated to be on the same basis.

Worked Example: Converting a Cohort Ratio to Expected Losses

A hypothetical specialty insurer has $1,250,000 of earned premium for a defined cohort. An expected loss ratio of 64% applies to that cohort and to the ultimate loss definition being estimated. Calculate expected ultimate losses and verify the result in two ways.

Model specification. Use the premium-based expected loss method with fixed earned premium $P=\$1,250,000$ and expected loss ratio $r=0.64$. No interest rate or mortality assumption applies, and no claim-size distribution or parameters are specified or needed.

Step 1: Convert the percentage to a proportion. The ratio is $64\%=64/100=0.64$.

Step 2: Multiply by earned premium.

E[U]=P r =\$1,250,000(0.64) =\$800,000.

Step 3: Check by interpreting the ratio. The premium contains 125 blocks of $10,000. At 64% of each block, expected losses per block are $6,400, and $125(\$6,400)=\$800,000$, agreeing with the formula.

Method and trap avoided. This is the premium-based expected loss method. Treating 64% as the number 64 would give $80,000,000, one hundred times too large. The percentage must be converted to a decimal proportion before multiplication.

Worked Example: Combining Two Segments Correctly

A hypothetical insurer estimates losses for two coverage segments. Segment A has $840,000 of earned premium and an expected loss ratio of 58%. Segment B has $360,000 of earned premium and an expected loss ratio of 72%. Find the expected ultimate losses for each segment, the total, and the overall expected loss ratio.

Model specification. Apply the premium-based formula separately to each segment, using fixed earned premium and the corresponding segment ratio. Both ratios use the same stated loss basis as their premiums. No interest rate, mortality assumption, or claim-size distribution is specified or needed.

Step 1: Estimate Segment A losses.

E[U_A]=\$840,000(0.58)=\$487,200.

Step 2: Estimate Segment B losses.

E[U_B]=\$360,000(0.72)=\$259,200.

Step 3: Add expected losses and premium. Total expected losses are $\$487,200+\$259,200=\$746,400$. Total earned premium is $\$840,000+\$360,000=\$1,200,000$.

Step 4: Calculate the implied overall ratio.

r_{\text{overall}}= \frac{\$746,400}{\$1,200,000} =0.622=62.2\%.

Check. Applying 62.2% to total premium gives $\$1,200,000(0.622)=\$746,400$, which matches the sum of the segment estimates.

Method and trap avoided. The segment estimates are aggregated by premium-weighted results. A simple average of the ratios, $(58\%+72\%)/2=65\%$, would imply $\$1,200,000(0.65)=\$780,000$, which is $33,600 too high. That average gives equal weight to the segments despite their unequal premiums.

Worked Example: Working in Thousands and Checking the Total

A hypothetical portfolio is summarized in thousands of dollars. Its earned premium is $475 thousand for a group with an expected loss ratio of 66%, and $325 thousand for a second group with an expected loss ratio of 54%. Calculate total expected ultimate losses in dollars and the overall expected loss ratio.

Model specification. Use $P_1=475$ thousand dollars and $P_2=325$ thousand dollars, with fixed premium amounts and ratios applying to their respective groups on the same loss basis. No interest rate or mortality assumption applies; no claim-size distribution is assumed.

Step 1: Calculate expected losses in thousands.

E[U_1]=475(0.66)=313.5\text{ thousand dollars}, $$ $$ E[U_2]=325(0.54)=175.5\text{ thousand dollars}.

Step 2: Sum the group estimates. Total expected losses are $313.5+175.5=489.0$ thousand dollars, or $\$489,000$. Total earned premium is $475+325=800$ thousand dollars, or $\$800,000$.

Step 3: Find the overall ratio and verify.

r_{\text{overall}}= \frac{489}{800} =0.61125=61.125\%, \qquad 800(0.61125)=489\text{ thousand dollars}.

Check. Both the direct sum and the aggregate-ratio calculation give $489,000. The displayed ratio is exact to the shown precision; rounding it before multiplying is unnecessary.

Method and trap avoided. This is a segment-by-segment premium-based estimate followed by aggregation. The intermediate amounts remain in thousands of dollars. Reading 489 thousand as $489, or multiplying a premium stated in thousands as though it were dollars, would create a factor-of-1,000 error.

Interpretation and Common Exam Traps

The formula is a translation rule, not a guarantee of the realized outcome. If the selected ratio is 0.60, the estimate says expected losses are 60% of the applicable premium base. It does not mean every policy has losses equal to 60% of its premium, nor does it mean the observed ultimate loss ratio must equal 60%.

The most important check is conceptual: does the expected loss ratio describe losses per dollar of this earned premium? If not, $P r$ may be arithmetically correct and actuarially inappropriate. A calculation also needs a clearly defined loss basis. For example, the ratio and resulting estimate must treat loss adjustment expense consistently; if the provided ratio includes it, the output includes it, and if it excludes it, the output excludes it.

Common errors include using written premium when the specified basis is earned premium, entering a percentage without converting it to a decimal, averaging segment ratios without premium weights, and presenting a result in the wrong monetary unit. For every calculation, label the premium basis, state the ratio as a proportion, show the multiplication, and check that the result has units of dollars.

Define the cohort and loss basis.
Specify which business, period, and ultimate loss definition are represented.

Confirm the premium and ratio match.
Use earned premium and an expected loss ratio applicable to that same basis.

Calculate and check.
Multiply premium by the decimal ratio, then verify the units and, for multiple segments, reconcile to the total.

Key takeaway: For a specified cohort with earned premium $P$ and applicable expected loss ratio $r$, the expected ultimate loss estimate is $P r$. The formula’s reliability depends on a consistent premium and loss basis, not on the multiplication itself.

Check Your Understanding

For each problem, identify the applicable premium base, convert percentages to proportions, and check the units.

A hypothetical book has $720,000 of earned premium and an applicable expected loss ratio of 67.5%. What is its expected ultimate loss estimate? (Answer: $486,000.)
A cohort has $1,600,000 of earned premium. Its expected ultimate losses are estimated using a 62.5% expected loss ratio. Calculate the estimate and state the ratio as a decimal. (Answer: 0.625 and $1,000,000.)
Two hypothetical segments have earned premiums of $400,000 and $600,000, with expected loss ratios of 55% and 70%, respectively. Find the total expected losses and implied overall ratio. (Answer: $640,000 and 64%.)
A portfolio is stated in thousands: Group A has 350 thousand of earned premium at a 60% expected loss ratio; Group B has 450 thousand at 68%. Find total expected losses in dollars and the overall ratio. (Answer: $516,000 and 64.5%.)
A hypothetical program has $900,000 of earned premium divided between two groups. Group 1 has $500,000 at an expected loss ratio of 56%; Group 2 has the remaining premium at an expected loss ratio of 71%. Calculate each group’s expected losses, the total, and the overall ratio. (Answer: $280,000 and $284,000 by group; $564,000 total; 62.6667% overall, approximately.)

What You'll Learn