Handling Immature Experience in Component Estimates

Why Immature Experience Needs Component-Level Attention

A frequency-severity estimate separates expected claim cost into the number of claims per exposure unit and the average cost per claim. That separation is useful only when both components describe a consistent population and valuation basis. Recent experience may have unreported claims, reported claims with unsettled amounts, or both. As a result, reported frequency can understate ultimate frequency, while average incurred severity can understate or overstate ultimate severity.

The key question is not whether a period is simply “immature.” It is which component is immature, and what development is appropriate for that component? A count development factor addresses claims not yet reported; a severity development factor addresses the change from the current average loss per claim to the selected ultimate average. One factor does not automatically replace the other.

Definition: For a specified exposure base, reported frequency is the number of claims reported by the valuation date divided by exposures. Ultimate frequency is the estimated ultimate number of claims for the experience period divided by exposures. Incurred loss at a valuation date is paid loss plus case reserves for reported claims. Ultimate severity is estimated ultimate loss divided by estimated ultimate claim count, using a consistent claim population and loss basis.

Reported counts can develop because claims have not yet reached the insurer or administrator. Incurred losses can develop because reported claims are still being adjusted, paid, reopened, or otherwise changed before ultimate settlement. Case reserves are estimates, not payments; including them in incurred loss does not make the claim’s ultimate amount certain.

Development factors are estimates based on suitable experience or selections. In the examples below, factors are supplied or derived from supplied factors; they are hypothetical assumptions, not universal constants. Applying a factor assumes that the selected development pattern is relevant to the experience being estimated. Material changes in reporting speed, claim handling, coverage, or settlement practices can undermine that assumption.

Keep the Components on a Consistent Basis

Let $E$ denote exposure units, $N$ the ultimate claim count, and $L$ the ultimate loss amount. Ultimate frequency is $N/E$, and ultimate severity is $L/N$. Their product gives the pure premium, or expected loss per exposure unit, when the quantities refer to the same period, claims, and loss basis.

\text{Ultimate pure premium} = \frac{N}{E}\times\frac{L}{N} = \frac{L}{E}.

This identity also explains a frequent practical issue: count development and loss development are related, but they are not interchangeable. If aggregate losses and counts are developed from the same population and basis, their development factors imply a severity factor.

Derivation: Suppose the reported count is $N_r$, reported incurred loss is $L_r$, the count development factor is $C=N/N_r$, and the aggregate loss development factor is $A=L/L_r$. Reported average severity is $L_r/N_r$, while ultimate average severity is $L/N$. Therefore, $$ \frac{L/N}{L_r/N_r} = \frac{L/L_r}{N/N_r} = \frac{A}{C}. $$ Thus the implied severity development factor is $A/C$, provided the count and loss factors apply to the same claims, period, valuation basis, and scope. This condition matters: dividing unrelated factors does not produce a valid severity factor.

If a selected severity factor is already supplied directly, apply it to the corresponding current severity. If instead an aggregate loss factor and a count factor are supplied on a consistent basis, their ratio gives the implied severity factor. Applying the aggregate loss factor to severity and then also developing frequency would count claim-count development twice.

Worked Examples

Worked Example: Developing Both Components

A hypothetical equipment insurer has 4,000 exposure units for a recent period. At the valuation date, it has 180 reported claims and $360,000 of incurred losses on those claims. Assume the selected count development factor is 1.15 and the selected incurred-severity development factor is 1.06. Estimate ultimate frequency, ultimate severity, and pure premium per exposure unit.

Method and assumptions. Apply the count factor to reported claims and the severity factor to reported incurred severity. Assume both selected factors apply to this period, exposure base, and claim population. The severity factor is already a severity factor, not an aggregate loss factor.

First develop the claim count:

N=180(1.15)=207\text{ claims}.

Ultimate frequency is:

\frac{207}{4{,}000}=0.05175\text{ claims per exposure unit}.

The current incurred severity is $360,000 divided by 180 reported claims:

\frac{\$360{,}000}{180}=\$2{,}000\text{ per claim}.

Develop severity to ultimate:

\$2{,}000(1.06)=\$2{,}120\text{ per claim}.

The pure premium is the product of the developed components:

0.05175(\$2{,}120)=\$109.71\text{ per exposure unit}.

Check. Estimated ultimate losses are $207(\$2{,}120)=\$438{,}840$. Dividing by 4,000 exposures gives $\$438{,}840/4{,}000=\$109.71$, matching the component calculation. The common mistake avoided is applying the count factor to the count and then applying it again to severity; the stipulated 1.06 factor already develops severity.

Worked Example: Inferring Severity Development From Aggregate Development

A hypothetical commercial property portfolio has 2,500 exposure units, 100 reported claims, and $200,000 in reported incurred losses. Assume a count development factor of 1.20 and an aggregate incurred-loss development factor of 1.32. Both factors apply to the same claims and valuation basis. Estimate the implied severity factor and ultimate pure premium.

Method and assumptions. Because the supplied loss factor develops aggregate losses, derive the severity factor by dividing it by the count factor. This uses the relationship derived above and assumes the two factors are comparable.

The implied severity factor is:

\frac{1.32}{1.20}=1.10.

Develop the claim count and calculate ultimate frequency:

N=100(1.20)=120,\qquad \frac{120}{2{,}500}=0.048\text{ claims per exposure unit}.

Current incurred severity is:

\frac{\$200{,}000}{100}=\$2{,}000\text{ per reported claim}.

Develop that severity and combine the components:

\$2{,}000(1.10)=\$2{,}200\text{ per claim}, \qquad 0.048(\$2{,}200)=\$105.60\text{ per exposure unit}.

Check. Ultimate aggregate losses are $\$200{,}000(1.32)=\$264{,}000$. The component calculation implies $120(\$2{,}200)=\$264{,}000$, and $\$264{,}000/2{,}500=\$105.60$. A common mistake is to apply 1.32 directly to the $2,000 average severity and then also develop count by 1.20. That would use an aggregate loss factor as if it were a severity factor and double-count count development.

Worked Example: Mature Counts but Unsettled Losses

A hypothetical specialty insurer has 2,000 exposure units and 80 reported claims for a period. Assume reporting is mature, so the count development factor is 1.00. Incurred losses on the claims total $200,000. The selected aggregate incurred-loss development factor is 1.12 and applies to this same period and claim population. Estimate ultimate frequency, severity, and pure premium.

Method and assumptions. Since the count factor is 1.00, claim counts are treated as ultimate for this estimate. Loss amounts are not yet mature, so derive severity development from the aggregate loss factor and count factor.

The implied severity factor is $1.12/1.00=1.12$. Ultimate frequency is:

\frac{80(1.00)}{2{,}000}=0.04\text{ claims per exposure unit}.

Current incurred severity is $\$200{,}000/80=\$2{,}500$ per claim. Therefore:

\text{Ultimate severity}=\$2{,}500(1.12)=\$2{,}800, \qquad \text{Pure premium}=0.04(\$2{,}800)=\$112.00.

Check. The developed aggregate losses are $\$200{,}000(1.12)=\$224{,}000$. The component estimate gives $80(\$2{,}800)=\$224{,}000$, or $\$224{,}000/2{,}000=\$112.00$ per exposure unit. The common mistake avoided is assuming that mature reporting also means mature claim amounts. Count maturity and settlement maturity are separate questions.

A Practical Assessment Sequence

Before projecting from immature experience, establish what each observed quantity represents and what remains incomplete. The following sequence keeps the component estimates aligned.

Define the basis.
Specify the experience period, valuation date, exposure measure, claim definition, and loss measure. Confirm that the counts and losses describe the same population.

Assess count maturity.
Determine whether unreported claims remain material. If so, develop reported counts to ultimate using a selected count basis appropriate to the experience.

Assess loss maturity separately.
Determine whether incurred amounts on reported claims are expected to change. If they are, select or derive severity development on a compatible basis.

Combine only after development.
Calculate ultimate frequency and ultimate severity, then multiply. Check that their product agrees with ultimate losses per exposure when the underlying bases are consistent.

The assessment need not result in two factors greater than 1.00. Counts can be mature while amounts remain unsettled, as in the third example. Conversely, a severity estimate might be credible on a suitable mature subset while reporting is incomplete for the full period. A factor can also be below 1.00 if the selected development pattern indicates a decline from the current observed measure. What matters is that the factor reflects the component and basis it is meant to develop.

Changes in claim reporting or reserving practices deserve particular attention. Faster reporting can make a recent period look more complete than it would under an older pattern; changes in case-reserve practices can move incurred severity without changing ultimate claim costs. If such changes make historical factors unrepresentative, mechanical application of those factors may not produce a sound estimate. The factor selection and its limitations must be considered before the calculation.

Common Exam Trap: Applying the Wrong Factor

Suppose a problem supplies a count development factor and an aggregate loss development factor. The aggregate loss factor develops total losses, which reflect both the number of claims and their average cost. It is not automatically the factor for average severity. On a consistent basis, derive the implied severity factor as aggregate loss development divided by count development.

For example, if count development is 1.20 and aggregate loss development is 1.32, the implied severity factor is $1.32/1.20=1.10$, not 1.32. Applying 1.32 to severity while also applying 1.20 to frequency would yield an aggregate development of $1.20(1.32)=1.584$, inconsistent with the supplied aggregate factor of 1.32.

Key takeaway: Develop claim counts for reporting emergence and average claim amounts for loss emergence. Use factors tied to their stated basis, and verify that developed frequency times developed severity reproduces ultimate losses per exposure when the components refer to the same claims and period.

Check Your Understanding

For each hypothetical portfolio, identify the relevant component development and calculate the requested estimate. Assume each stated factor is appropriate for the stated experience basis.

A portfolio has 3,000 exposure units, 120 reported claims, and $240,000 of incurred losses. The count development factor is 1.25 and the severity development factor is 1.05. What are ultimate frequency, ultimate severity, and pure premium per exposure unit? (Answer: 0.0500 claims per exposure unit; $2,100.00 per claim; $105.00 per exposure unit.)
A portfolio has 4,000 exposure units, 200 reported claims, and $500,000 of incurred losses. The count factor is 1.10 and the aggregate loss factor is 1.21, both on the same basis. Find the implied severity factor and pure premium. (Answer: Severity factor 1.10; $151.25 per exposure unit.)
A portfolio has 5,000 exposure units and 150 claims. Assume counts are fully developed, incurred losses are $450,000, and the aggregate loss development factor is 1.20 on the same claim population. Estimate the pure premium. (Answer: $108.00 per exposure unit.)
A portfolio has 10,000 exposure units, 400 reported claims, and $800,000 of incurred losses. The count factor is 1.08 and the aggregate loss factor is 1.188. Find ultimate frequency, implied severity factor, ultimate severity, and pure premium. (Answer: 0.0432 claims per exposure unit; severity factor 1.10; $2,200.00 per claim; $95.04 per exposure unit.)
A recent period has a reported count factor of 1.15 and an aggregate loss factor of 1.15, selected on the same basis. What is the implied severity factor? Explain what it suggests about the relationship between count and aggregate loss development; do not infer that claims are settled. (Answer: 1.00; the selected aggregate loss development is fully explained by count development, but this does not establish that individual claim amounts are settled.)

What You'll Learn