Tutorials › CAS Exam 5 - Basic Ratemaking and Estimated Claim Liabilities › Using Ultimate Losses as an Indication Input

CAS Exam 5 · Indication Framework and Data · Tutorial 212 of 989

Using Ultimate Losses as an Indication Input

Distinguish reported experience from selected ultimate losses and place the ultimate estimate on the correct basis as an indication input.

Beginner 11 min read

What You'll Learn

  • Define reported, incurred, and ultimate losses in an indication context
  • Explain why immature historical losses may understate their ultimate amount
  • Apply supplied development estimates to reported or incurred losses
  • Select an ultimate loss estimate when multiple estimates are provided
  • Check that the loss input and premium base refer to the same experience
  • Distinguish development to ultimate from projection to the prospective period

Why Ultimate Losses Matter in an Indication

Historical claims do not all become fully known at the same time. A recent accident period may have claims that have not yet been reported, reported claims that have not been fully paid, or open claims whose final costs are not yet settled. Comparing that incomplete loss amount with premium can make the experience appear more favorable than it ultimately will be.

For a rate level indication, the loss input should represent the estimated cost of the claims arising from the selected experience, including amounts not yet observed in the data. That estimate is called the ultimate loss. The central task here is to identify the selected ultimate estimate and use it as the historical loss input—not to project it to a future period. Projection and trend adjustments are separate steps.

Definition: Reported losses are losses recorded as reported by the valuation date, under the data definition being used. Incurred losses commonly include paid losses plus case outstanding reserves for reported claims. Ultimate losses are the selected estimates of the total losses attributable to the experience period after all covered claims have emerged and their costs have been fully developed. An ultimate estimate can include future payments on known claims, claims not yet reported, and revisions to current estimates.

“Ultimate” describes the estimated final amount attributable to the experience period; it does not mean that every claim has actually been settled. A selected ultimate estimate is an actuarial estimate at a stated evaluation date. The experience period, coverage, loss definition, and valuation basis should still match the premium and exposure basis used in the indication.

Development Is Not Prospective Trend

Development adjusts for the fact that losses from a past period may be incomplete at the valuation date. Trend, by contrast, adjusts the cost level from the historical experience period toward a prospective period. These adjustments answer different questions and should not be substituted for one another.

Key distinction: Development to ultimate estimates the final cost of claims arising from the historical experience. Trend projects the cost level between periods. A loss can require development even if no trend adjustment is made; a fully developed ultimate loss can still require trend before it represents a prospective loss cost.

For this tutorial, a development factor is defined as ultimate losses divided by the observed losses at a specified evaluation age. If \(L_{\text{obs}}\) is the observed amount and \(f\) is the applicable selected factor, then:

$$ L_{\text{ult}}=L_{\text{obs}}f $$

This multiplication is valid only when the factor is defined on that same observed basis and applies to the relevant period, coverage, and loss measure. For instance, a factor selected for incurred losses at a particular age should not automatically be applied to paid losses at a different age. If a selected ultimate estimate is supplied directly, use that estimate rather than applying another factor to it.

Why the Ultimate Estimate Belongs in the Loss Input

An indication compares expected claim costs with the premium basis. If the numerator includes only currently observed losses, it can omit costs that belong to the historical experience but have not yet emerged. Using an ultimate estimate brings the numerator closer to the total claim cost attributable to that experience.

The rationale can be stated as a simple decomposition. At the valuation date, the ultimate amount consists of the observed amount plus the estimated remaining amount:

$$ L_{\text{ult}}=L_{\text{obs}}+L_{\text{remaining}} $$

A development factor expresses this same relationship as a ratio, \(f=L_{\text{ult}}/L_{\text{obs}}\), when \(L_{\text{obs}}>0\). Rearranging gives \(L_{\text{ult}}=L_{\text{obs}}f\). Thus, if the observed losses are only part of the eventual cost, using \(L_{\text{obs}}\) alone leaves out \(L_{\text{remaining}}\). The factor is an estimate of the total-to-observed relationship, not an additional amount to add on top of the observed amount.

In an indication framework, the selected ultimate loss estimate is paired with a consistent premium base. For example, an ultimate loss ratio is the selected ultimate losses divided by the appropriate premium for the same experience and rate-level basis. If loss adjustment expenses are part of the selected loss measure, they must be handled consistently with the premium and expense basis. The earlier material on matching loss adjustment expenses to the loss basis addresses that alignment.

Worked Example: Developing an Incurred Loss Estimate to Ultimate

A hypothetical small-equipment program has incurred losses of $720,000 for an experience period. At the valuation date, the selected incurred-to-ultimate factor for this period is 1.20. Calculate the ultimate loss estimate and explain which amount is the indication input. No prospective trend adjustment is requested.

Step 1: Identify the observed basis. The $720,000 is incurred losses at the stated valuation date, not ultimate losses. The supplied factor is specifically an incurred-to-ultimate factor, so its basis matches.

Step 2: Apply the factor.

$$ L_{\text{ult}}=\$720{,}000(1.20)=\$864{,}000 $$

The implied estimated remaining development is \(\$864{,}000-\$720{,}000=\$144{,}000\). This is already included in the ultimate estimate; it is not an amount to add a second time.

Step 3: Identify the input. The selected ultimate loss input is $864,000. The $720,000 is the observed incurred amount, and the $144,000 is the implied remaining development.

Check. The ultimate estimate exceeds the incurred amount by 20% of $720,000, or $144,000. Also, $720,000 plus $144,000 equals $864,000. The common mistake avoided is treating the factor as a percentage to add: adding 1.20 times the observed amount to the observed amount would count the observed losses twice.

Selecting an Ultimate Estimate

Sometimes a problem supplies one applicable development factor; sometimes it supplies several estimates or states that an actuary has selected an ultimate amount. The indication uses the selected estimate, not every estimate added together. The selection may reflect the relative reliability of the available information, but a stated selection should not be replaced by an unrequested method.

Before placing the selected number into an indication, check four things:

1
Confirm the experience period.
Make sure the loss estimate belongs to the period whose premium and exposures are being used.
2
Confirm the valuation and basis.
Distinguish paid, reported, or incurred losses from a selected ultimate estimate, and check that any supplied factor applies to that exact observed basis.
3
Use the selected ultimate amount once.
If the selected estimate is already ultimate, do not multiply it by another factor or add development again.
4
Keep other adjustments distinct.
Development estimates historical claim maturity. On-leveling and prospective trend adjustments address different changes in the indication framework.

Worked Example: Choosing the Stated Ultimate Selection

A hypothetical recreational-vehicle insurer reviews one experience period. The observed incurred losses are $500,000. An applicable factor produces an actuarial estimate of $625,000, while a separate review produces an estimate of $610,000. The problem states that the selected ultimate loss estimate is $620,000. The matched on-level premium for the same experience is $1,000,000. Find the loss input and the ultimate loss ratio for this step of the indication.

Step 1: Separate estimates from the selection. The factor-based estimate is $625,000, and the separate review estimate is $610,000. Neither is the stated selection. The selected ultimate amount supplied in the problem is $620,000.

Step 2: Use the selected ultimate amount as the numerator. The indication loss input is $620,000. Do not add the $625,000 or $610,000 estimates to it, and do not apply the factor again.

Step 3: Divide by the matched premium.

$$ \text{Ultimate loss ratio} =\frac{\$620{,}000}{\$1{,}000{,}000} =0.62=62\% $$

Check. The selected ultimate losses are 62 cents per dollar of the stated matched premium. The result is a loss ratio for this step, not a complete rate indication; no conclusion about other provisions or prospective adjustments is being made. The common mistake avoided is selecting the largest or smallest estimate automatically instead of following the explicit selection.

Combining Periods Without Losing the Basis

When an indication uses multiple experience periods, each period’s selected ultimate losses should be identified before aggregation. If the periods have different observed amounts or development estimates, apply the appropriate estimate to each period and then sum the selected ultimate amounts. Do not apply one factor to a combined total unless the problem establishes that the factor is appropriate for that combined basis.

This order matters because the periods may be at different maturities. A single factor applied to a combined observed amount can misstate the result if one period has more remaining development than another. Likewise, the premium in the denominator must correspond to the experience represented by the ultimate-loss numerator.

Worked Example: Aggregating Ultimate Losses Across Two Periods

A hypothetical specialty property program uses two consecutive experience periods. Period A has $400,000 of observed incurred losses and an applicable incurred-to-ultimate factor of 1.10. Period B has $300,000 of observed incurred losses and an applicable factor of 1.30. The corresponding matched on-level premiums are $500,000 and $400,000. Calculate the total selected ultimate losses and the aggregate ultimate loss ratio.

Step 1: Develop Period A.

$$ L_{\text{ult,A}}=\$400{,}000(1.10)=\$440{,}000 $$

Step 2: Develop Period B.

$$ L_{\text{ult,B}}=\$300{,}000(1.30)=\$390{,}000 $$

Step 3: Aggregate numerator and denominator on matching period bases.

$$ L_{\text{ult,total}}=\$440{,}000+\$390{,}000=\$830{,}000 $$
$$ P_{\text{total}}=\$500{,}000+\$400{,}000=\$900{,}000 $$

Step 4: Calculate the aggregate ratio.

$$ \frac{\$830{,}000}{\$900{,}000}=0.9222\ldots=92.22\% $$

Check. Period A’s estimated remaining development is $40,000, and Period B’s is $90,000. Their sum, $130,000, added to total observed losses of $700,000, gives $830,000. The ratio uses the combined premium of $900,000. The common mistake avoided is applying one period’s factor to the combined $700,000 when the two periods have different supplied factors.

Common Traps and Practical Checks

The ultimate-loss input is easy to mishandle when the question places reported, incurred, and ultimate amounts together. Keep the data labels attached to each number throughout the calculation. A factor is meaningful only with its specified starting basis and valuation age; an ultimate selection is already the destination amount.

  • Using observed losses as if they were ultimate. If the experience is immature and an applicable selected estimate is provided, use the ultimate amount for the historical claim-cost input.
  • Adding a factor to losses instead of multiplying. A factor of 1.20 means ultimate losses are 1.20 times the observed amount, not observed losses plus 1.20 times the observed amount.
  • Developing an ultimate selection again. Once the problem states the selected ultimate losses, do not apply a second development adjustment.
  • Mixing paid and incurred bases. A factor selected for one observed measure is not automatically appropriate for another.
  • Confusing development with trend. Development estimates the final cost of historical claims; trend moves a cost level between periods.
  • Pairing mismatched totals. The loss and premium totals must represent the same experience and a consistent basis before a ratio is interpreted.
Key takeaway: Identify the observed loss basis, determine whether a development estimate is needed, and use the selected ultimate losses once as the historical loss input. Keep that input distinct from adjustments that address rate level or the prospective period.

Check Your Understanding

For each problem, identify the relevant ultimate loss input and show the key calculation.

  1. A hypothetical farm-equipment portfolio has $260,000 of incurred losses and a stated incurred-to-ultimate factor of 1.15. What is the ultimate estimate? (Answer: $299,000.)
  2. A hypothetical event-cancellation portfolio has $480,000 of observed losses. A factor-based estimate is $528,000, but the stated selected ultimate losses are $520,000. Which amount is the indication input, and what is the implied remaining development relative to observed losses? (Answer: $520,000; implied remaining development is $40,000.)
  3. A hypothetical marine portfolio has two experience periods. Period 1 has $200,000 observed incurred losses with factor 1.10 and $250,000 matched premium. Period 2 has $300,000 observed incurred losses with factor 1.20 and $350,000 matched premium. Find total ultimate losses and the aggregate ultimate loss ratio. (Answer: $580,000; $580,000 ÷ $600,000 = 96.67%.)
  4. A hypothetical warranty program has $700,000 of incurred losses. A factor of 1.25 is explicitly defined for paid losses, while no paid amount or incurred-to-ultimate factor is supplied. Can the factor be applied to the incurred amount to determine ultimate losses? (Answer: No; the factor’s paid-loss basis does not match the $700,000 incurred-loss basis, so the ultimate amount cannot be calculated from the supplied information alone.)
  5. A hypothetical cyber policy group has selected ultimate losses of $840,000 and matched on-level premium of $1,200,000. The ultimate estimate already includes all selected development. What is the ultimate loss ratio, and should another development factor be applied? (Answer: 70%; no further development factor should be applied.)
Clinical Biostats