Tutorials › CAS Exam 5 - Basic Ratemaking and Estimated Claim Liabilities › Inflation Effects in Reserve Emergence

CAS Exam 5 · Reserve Reporting and Interpretation · Tutorial 970 of 989

Inflation Effects in Reserve Emergence

Learn to identify, quantify, and interpret inflation-driven changes in claim payments, incurred amounts, and selected unpaid liabilities.

Advanced 10 min read

What You'll Learn

  • Distinguish claim severity inflation from changes in claim counts and portfolio mix
  • Project future claim amounts using annual and cumulative severity trends
  • Interpret why paid and incurred emergence can reflect inflation at different times
  • Update a prior reserve selection without applying inflation twice
  • Recognize social inflation as a potential source of changing liability severity

Why Inflation Matters in Reserve Emergence

A rise in reported claim costs does not automatically mean that more claims occurred or that an earlier reserve was inadequate. The amount paid to settle a claim can change as medical costs, repair costs, litigation practices, or settlement values change. When this change in severity is not separated from claim counts, claim mix, and reserving practice, reserve emergence can be misread.

For reserve analysis, distinguish the amount of a claim from when that amount is paid or recorded. Paid losses are amounts already paid; incurred losses are paid losses plus case reserves for reported claims. A change in incurred losses can therefore reflect both cash payments and revisions to estimates of unpaid claims.

Definition: Claim severity inflation is an increase over time in the cost per claim, measured on a sufficiently consistent basis. If expected severity increases at a constant annual rate \(g\), a base amount \(S_0\) grows for \(t\) years to \(S_0(1+g)^t\). Social inflation describes liability-cost growth associated with changes in the legal, social, or settlement environment, such as shifts in litigation patterns or award levels. It may contribute to severity inflation, but it is not a separate mechanical adjustment that can be assumed to apply uniformly to every claim.

The formula uses a multiplicative annual rate: each year’s increase applies to the amount after the preceding year’s increase. In practice, the relevant rate may differ by coverage, claim type, geography, and period. A rate derived from average paid severity is informative only to the extent that the observations are comparable in claim maturity, limits, deductibles, and mix.

Projecting Inflation by Payment or Settlement Time

An unpaid liability may include claims expected to settle at different future dates. If a selected base amount represents severity at a defined starting cost level, project each portion to its expected payment date using the applicable annual rates. For a claim or group expected to be paid in year \(t\), with annual rates \(g_1,\ldots,g_t\), the cumulative factor is the product of the annual factors.

$$ F_t=\prod_{k=1}^{t}(1+g_k), \qquad \text{projected amount at }t=\text{base amount}\times F_t $$

This adjustment is appropriate only when the base amount has not already incorporated the same future inflation. If observed payments or current case estimates already reflect the relevant cost level, adding the full trend again would double count it. The starting cost level and the timing of each amount must therefore be explicit.

Justification: Under the assumption that expected severity in each successive year is multiplied by \(1+g_k\), the first year’s amount is \(S_0(1+g_1)\). Applying the second year’s increase gives \(S_0(1+g_1)(1+g_2)\). Repeating this step through year \(t\) gives \(S_0\prod_{k=1}^{t}(1+g_k)\). If the rate is constant at \(g\), this becomes \(S_0(1+g)^t\). For a fixed number \(N\) of comparable claims, expected aggregate severity is \(N\) times expected severity per claim, so the same factor applies to the aggregate under those assumptions.

Worked Example: Comparing Paid Severity Across Calendar Years

A hypothetical repair coverage has 100 paid claims in each of two calendar years. The first-year average paid severity is $12,000. Assume the claim mix and coverage terms are unchanged, and that the relevant annual severity inflation rate is 5%. Estimate the second-year average severity and total paid losses. Then compare the observed total with a no-inflation baseline.

Method. Apply the annual severity factor once because the second calendar year is one year after the base observation.

Second-year average severity:

$$ \$12{,}000(1.05)=\$12{,}600 $$

Second-year total paid losses:

$$ 100(\$12{,}600)=\$1{,}260{,}000 $$

At the first-year severity, 100 claims would produce \(100(\$12{,}000)=\$1{,}200{,}000\). The difference is \(\$1{,}260{,}000-\$1{,}200{,}000=\$60{,}000\), or 5% of the baseline total. The claim count did not change in this example; the increase is attributable to the assumed severity trend, conditional on the stated comparability assumption.

Check. The projected average is 5% above $12,000, and the projected total is 100 times that average. The method avoids interpreting higher paid totals as higher frequency when the assumed count is unchanged.

Worked Example: Updating a Prior Unpaid Claim Selection

A hypothetical liability portfolio has 40 open claims. A prior selection assigns each claim a base unpaid amount of $25,000, stated at the current valuation cost level, with no future severity trend included. Ten claims are expected to settle in one year and 30 in two years. Assume a 4% annual severity trend for both future years. Calculate the inflation-adjusted unpaid selection.

Method. Apply one year of trend to the first group and two compounded years to the second group.

For the 10 claims expected to settle in one year, the projected amount per claim is:

$$ \$25{,}000(1.04)=\$26{,}000 $$

Their projected total is \(10(\$26{,}000)=\$260{,}000\).

For the 30 claims expected to settle in two years, the cumulative factor is \(1.04^2=1.0816\). The projected amount per claim is:

$$ \$25{,}000(1.0816)=\$27{,}040 $$

Their projected total is \(30(\$27{,}040)=\$811{,}200\). The adjusted selection is therefore \(\$260{,}000+\$811{,}200=\$1{,}071{,}200\). Without future trend, the base selection would be \(40(\$25{,}000)=\$1{,}000{,}000\), so the indicated increase is \(\$71{,}200\).

Check. The result lies between applying one year of trend to all claims, which would give $1,040,000, and applying two years to all claims, which would give $1,081,600. The method avoids treating every open claim as if it had the same settlement date.

Paid and Incurred Emergence

Inflation can affect paid and incurred results on different schedules. Paid severity reflects the cost of claims as amounts are actually paid. Incurred severity also depends on when case reserves are revised and how well those reserves anticipate future settlement costs. If case estimates are promptly updated for expected inflation, incurred amounts may reflect an increase before the corresponding cash payments occur. If estimates are not updated, the change may appear more strongly in later paid emergence and in subsequent case-reserve revisions.

The following simplified example holds claim count and claim status fixed except for settlement. It illustrates the accounting identity for incurred losses; it does not assume that every real case reserve responds immediately or uniformly to inflation.

Worked Example: Inflation Reflected in Paid and Incurred Amounts

At the start of a year, a hypothetical group has 10 reported claims, each with a $100,000 case reserve and no payments to date. During the year, the expected settlement cost per claim increases by 6%. Four claims settle during the year for the updated estimate, and case reserves on the other six are also updated to the same amount. Assume no other payments, claim-count changes, or reserve revisions. Calculate year-end paid losses, case reserves, and incurred losses, then reconcile the incurred change.

Method. Increase each claim estimate by 6% before calculating the settled payments and remaining case reserves.

Updated estimate per claim:

$$ \$100{,}000(1.06)=\$106{,}000 $$

Four claims settle, so year-end paid losses are \(4(\$106{,}000)=\$424{,}000\). Six claims remain open, so year-end case reserves are \(6(\$106{,}000)=\$636{,}000\). Incurred losses are:

$$ \$424{,}000+\$636{,}000=\$1{,}060{,}000 $$

Beginning incurred losses were \(10(\$100{,}000)=\$1{,}000{,}000\). The incurred increase is \(\$1{,}060{,}000-\$1{,}000{,}000=\$60{,}000\), exactly 6% of the beginning amount. The settled claims account for $424,000 of paid losses, while the remaining liability is $636,000 in case reserves.

Check. The $1,060,000 year-end incurred total equals the updated estimate for all 10 claims, whether already paid or still reserved. It would be incorrect to add the $424,000 paid to the original $1,000,000 reserve without removing the reserves for the four settled claims.

Interpreting Prior Selections and Reserve Changes

A prior reserve selection is made from information and assumptions available at its valuation date. When subsequent experience reflects higher severities, the analyst should ask whether the earlier selection anticipated the cost level that actually emerged. A reserve increase is not automatically a pure inflation adjustment: the change may also reflect claim count, claim mix, reporting or settlement timing, policy limits, case-reserving practice, or new information about individual claims.

A useful interpretation starts by stating the basis of comparison. For example, compare claims at similar maturity and coverage terms, and identify whether amounts are paid, incurred, or selected unpaid liabilities. Then determine which cost level the prior selection represented and which future inflation, if any, it already included. Only after that should a trend adjustment be considered.

Social inflation can complicate this assessment. Changes in litigation frequency, the size or distribution of awards, settlement behavior, and legal practices may increase liability severity even when general price indices do not show a comparable change. A historical average that spans a change in the legal environment may not be a stable guide to future claim costs. Conversely, a rise in a few large claims does not by itself establish a portfolio-wide social inflation rate. The analyst should consider claim-level information and consistency of the observed data before applying a broad factor.

1
Define the comparison.
Specify the coverage, claim population, valuation dates, and whether the measure is paid, incurred, or unpaid.
2
Separate exposure and mix.
Check whether changes in claim counts, limits, deductibles, or the types of claims explain some of the observed difference.
3
Establish the cost-level basis.
Identify the severity level already reflected in the earlier selection and the timing of any future payments.
4
Apply only the missing trend.
Project amounts not already stated at the required future cost level; document the rate and period used.

Common Mistakes and Exam Traps

  • Adding annual rates instead of compounding. For two years at 5%, the incorrect factor \(1+0.05+0.05=1.10\) gives $1,100,000 on a $1,000,000 base. The correct factor is \(1.05^2=1.1025\), giving $1,102,500. The second year’s 5% applies to the already increased amount.
  • Applying trend twice. If a selected case estimate already reflects the expected future settlement cost, multiplying it by the full future trend again overstates the liability.
  • Calling every rise in total losses inflation. Aggregate losses can increase because there are more claims or a different mix of claims, even if severity is unchanged.
  • Assuming paid and incurred changes occur together. Payments depend on settlement timing; incurred amounts also depend on case-reserve updates. The two measures can show inflation at different stages.
Key takeaway: Interpret reserve emergence in relation to the cost level already embedded in the estimate, the expected timing of unpaid claims, and the consistency of the comparison. Inflation is a severity assumption—not a substitute for analyzing claim counts, mix, or reserving practice.

Check Your Understanding

For each problem, state the relevant inflation assumption and keep the cost-level basis in view.

  1. A hypothetical coverage has average severity of $18,000 at the start of a two-year period. The annual severity rates are 3% in year one and 7% in year two. What is the projected severity at the end of year two? (Answer: $19,837.80)
  2. A portfolio has 24 comparable open claims, each with a $15,000 base unpaid estimate. The estimate excludes future inflation, and all claims are expected to settle in three years. At a constant annual rate of 4%, what is the projected total? (Answer: $404,951.04)
  3. Eight thousand dollars is the base unpaid estimate per claim. Five claims are expected to settle after one year and seven after two years. Apply 5% inflation in year one and 4% in year two. What is the projected unpaid total? (Answer: $103,152)
  4. Beginning incurred losses are $900,000. During the year, three claims settle for $110,000 each, and the seven remaining claims each have a $90,000 case reserve. What are ending incurred losses and the change from beginning incurred losses? (Answer: Ending incurred losses are $960,000; the increase is $60,000)
  5. A prior selection is $500,000 for open claims and is stated at the current cost level, with no future trend included. All claims are expected to settle in two years, and the assumed annual severity trend is 5%. What is the inflation-adjusted selection? (Answer: $551,250)
Clinical Biostats