Allocating Expected Losses with D-Ratios

Why Allocate Expected Losses?

Experience rating plans may distinguish between primary expected losses and excess expected losses. The distinction reflects how losses are assigned to the portions of the plan below and above a specified split point. This tutorial concerns the allocation of an already determined total expected loss amount; it does not derive the D-ratio from a severity distribution.

A D-ratio converts a total expected loss amount into its primary component. The remaining amount is the excess component. The arithmetic is simple, but the inputs must be on a consistent basis: the D-ratio must correspond to the applicable split point and expected-loss basis, and it must be applied to expected losses rather than confused with a claim count or an observed loss amount.

Definition: For a specified split point and loss basis, the D-ratio, denoted $D$, is the ratio of primary expected losses to total expected losses. Thus, $D$ is the proportion of total expected losses allocated to the primary component, and $1-D$ is the proportion allocated to the excess component. A valid D-ratio satisfies $0 \leq D \leq 1$.

If $EL_T$ is total expected losses, the component amounts are:

EL_P = D\,EL_T

EL_X = (1-D)\,EL_T = EL_T-EL_P

Here $EL_P$ denotes primary expected losses and $EL_X$ denotes excess expected losses. The subscript $X$ labels the excess component; it is not a claim random variable in these equations. If expected losses are calculated from an exposure amount $E$ and an expected loss rate $r$ on a per-exposure-unit basis, first calculate $EL_T=Er$, then allocate that amount.

Why the Allocation Works

The D-ratio is a ratio of expected amounts, not a ratio of claim counts. To see why multiplying by it gives the primary expected loss, consider a nonnegative loss $X$ and a split point $s$, using the same loss basis for both components. The primary part of one loss is $\min(X,s)$, and the excess part is $(X-s)^+$, where $z^+=\max(z,0)$. For every possible value of $X$,

X=\min(X,s)+(X-s)^+

Taking expectations and using additivity of expectation gives $E[X]=E[\min(X,s)]+E[(X-s)^+]$. By definition, $D=E[\min(X,s)]/E[X]$, provided $E[X]$ is finite and positive. Therefore $E[\min(X,s)]=D E[X]$, while the excess expectation is $E[X]-D E[X]=(1-D)E[X]$. This justifies the allocation formulas.

The identity uses components that partition the same loss amount. If the total expected loss and D-ratio use different coverage, trend, development, or split-point bases, the ratio may not represent the primary share of that total. In that case, the arithmetic still produces numbers, but the allocation may not be actuarially consistent.

Worked Example: Allocating One Total Expected Loss Amount

A hypothetical account has total expected losses of $240,000 for a specified coverage and experience-plan split point. The applicable D-ratio is 0.35. Calculate primary and excess expected losses and check the result.

Method. Apply the D-ratio to total expected losses for the primary amount, and allocate the remaining fraction to excess.

Step 1: Calculate primary expected losses.

EL_P=0.35(240{,}000)=84{,}000

Step 2: Calculate excess expected losses. The excess share is $1-0.35=0.65$, so

EL_X=0.65(240{,}000)=156{,}000

Step 3: Check the allocation.

$$ 84{,}000+156{,}000=240{,}000 $$

The components reconcile exactly to total expected losses. Common mistake avoided: applying 0.35 to a claim count or treating it as the excess share would answer a different question. Here it is specifically the primary share of expected loss.

Allocating by Classification or Coverage

When an account contains several classifications or coverages, calculate the expected loss amount and allocation for each component group when the expected loss rates or D-ratios differ. Add the resulting primary amounts and excess amounts separately. This preserves the relationship between each group’s expected losses and its applicable D-ratio.

EL_{P,\mathrm{total}}=\sum_j D_j EL_{T,j} \qquad\text{and}\qquad EL_{X,\mathrm{total}}=\sum_j (1-D_j)EL_{T,j}

Here $j$ indexes the groups, $EL_{T,j}$ is group $j$'s total expected loss, and $D_j$ is its applicable D-ratio. If a single aggregate D-ratio is needed, it must reproduce the total primary expected losses. It is therefore a total-expected-loss-weighted average, not generally the simple average of the group D-ratios:

D_{\mathrm{aggregate}} = \frac{\sum_j D_j EL_{T,j}}{\sum_j EL_{T,j}}

Worked Example: Allocating Expected Losses Across Two Classes

A hypothetical account has two classifications. Class A has 120 exposure units and an expected loss rate of $180 per unit, with a D-ratio of 0.40. Class B has 80 exposure units and an expected loss rate of $300 per unit, with a D-ratio of 0.25. Calculate the total, primary, and excess expected losses.

Method. First calculate total expected losses separately by class as exposure times expected loss rate. Then apply each class’s D-ratio to its own expected losses.

Step 1: Class A total expected losses.

EL_{T,A}=120(180)=21{,}600

Its primary amount is $0.40(21{,}600)=8{,}640$. Its excess share is $1-0.40=0.60$, giving excess expected losses of $0.60(21{,}600)=12{,}960$.

Step 2: Class B total expected losses.

EL_{T,B}=80(300)=24{,}000

Its primary amount is $0.25(24{,}000)=6{,}000$. Its excess share is $1-0.25=0.75$, giving excess expected losses of $0.75(24{,}000)=18{,}000$.

Step 3: Add the corresponding components.

EL_{T,\mathrm{total}}=21{,}600+24{,}000=45{,}600

EL_{P,\mathrm{total}}=8{,}640+6{,}000=14{,}640

EL_{X,\mathrm{total}}=12{,}960+18{,}000=30{,}960

Step 4: Check. The component sum is $14{,}640+30{,}960=45{,}600$, which equals total expected losses. Common mistake avoided: using one class’s D-ratio for both classes ignores the given class-specific allocation factors.

Aggregate D-Ratios and Weighting

An aggregate D-ratio is useful when a single ratio must represent several groups together. Since each group contributes a different amount of total expected loss, each D-ratio must be weighted by that amount. A group with $120,000 of expected losses has twice the influence on the aggregate primary share as a group with $60,000, all else equal.

Worked Example: Finding the Aggregate D-Ratio

A hypothetical portfolio has two coverages. Coverage A has total expected losses of $50,000 and a D-ratio of 0.60. Coverage B has total expected losses of $150,000 and a D-ratio of 0.20. Find the combined primary and excess expected losses and the aggregate D-ratio.

Method. Allocate each coverage separately, add the components, and divide combined primary expected losses by combined total expected losses to find the aggregate D-ratio.

Step 1: Allocate Coverage A. Primary expected losses are $0.60(50{,}000)=30{,}000$. Excess expected losses are $0.40(50{,}000)=20{,}000$.

Step 2: Allocate Coverage B. Primary expected losses are $0.20(150{,}000)=30{,}000$. Excess expected losses are $0.80(150{,}000)=120{,}000$.

Step 3: Combine the amounts.

EL_{T,\mathrm{total}}=50{,}000+150{,}000=200{,}000

EL_{P,\mathrm{total}}=30{,}000+30{,}000=60{,}000

EL_{X,\mathrm{total}}=20{,}000+120{,}000=140{,}000

Step 4: Calculate the aggregate D-ratio.

D_{\mathrm{aggregate}}=\frac{60{,}000}{200{,}000}=0.30

The check is $0.30(200{,}000)=60{,}000$, the combined primary amount calculated by adding the coverage allocations. Common mistake avoided: the unweighted average of 0.60 and 0.20 is 0.40, which would imply $80,000 of primary expected losses. That is not the correct combined allocation because the coverages do not have equal total expected losses.

Checks and Common Errors

The reconciliation is an essential control: primary plus excess expected losses must equal total expected losses. A mismatch usually signals an arithmetic error, an incorrectly calculated complementary share, or an inconsistent basis. Use the following sequence when performing an allocation:

Confirm the total expected loss base.
Identify the total expected loss amount to be split and the relevant group, coverage, or account.

Match the D-ratio.
Check that the ratio applies to the same split point and loss basis as the expected losses being allocated.

Calculate both components.
Use $D$ for primary and $1-D$ for excess, or calculate excess as total less primary.

Reconcile.
Verify that primary plus excess equals total, and aggregate components only after applying the appropriate group ratios.

A D-ratio does not mean that the same percentage of claims fall below the split point. It describes a share of expected loss dollars. Nor should an analyst take a simple average of different D-ratios unless the groups have equal total expected losses (or the particular values happen to make the simple average correct). If every group has the same D-ratio, applying it to the combined total is equivalent to allocating each group and then adding.

These calculations support consistent primary and excess expected-loss inputs in experience-plan work. They are aligned with the CAS Exam 8 (Advanced Ratemaking) syllabus. The most reliable presentation states the total expected losses, the applicable D-ratio, both component calculations, and the reconciliation.

Key takeaway: Multiply total expected losses by the applicable D-ratio to obtain primary expected losses; multiply by its complement to obtain excess expected losses. For multiple groups, allocate group by group and then sum. Always confirm that the two components add back to the original total.

Check Your Understanding

For each problem, calculate the requested allocation and use the reconciliation principle to check it.

A hypothetical account has $72,000 of total expected losses and a D-ratio of 0.30. What are its primary and excess expected losses? (Answer: primary $21,600; excess $50,400.)
A hypothetical class has 40 exposure units at an expected loss rate of $250 per unit and a D-ratio of 0.45. Calculate total, primary, and excess expected losses. (Answer: total $10,000; primary $4,500; excess $5,500.)
A hypothetical coverage has $125,000 of total expected losses and a D-ratio of 0.28. How much is allocated to each component? (Answer: primary $35,000; excess $90,000.)
A hypothetical account has Coverage A with $80,000 of total expected losses and a D-ratio of 0.60, and Coverage B with $120,000 and a D-ratio of 0.20. Find total primary and excess expected losses and the aggregate D-ratio. (Answer: total $200,000; primary $72,000; excess $128,000; aggregate D-ratio 0.36.)
A hypothetical portfolio has $78,000 of total expected losses. Its primary expected losses are $31,200. Find the D-ratio and excess expected losses, and verify the total. (Answer: D-ratio 0.40; excess $46,800; $31,200 + $46,800 = $78,000.)

What You'll Learn