Diagnosing Reserve Adequacy Distortions

Why Reserve Adequacy Can Distort Development

Incurred losses combine amounts already paid with estimates of what remains to be paid. A change in case-reserving practice can therefore change an incurred triangle even when the underlying loss emergence has not changed. For example, more conservative case estimates may raise incurred losses at a given age; later payments may then reduce case outstanding without any corresponding change in the ultimate cost of the claims.

The diagnostic question is not simply whether incurred losses increased. It is whether the change is concentrated in case outstanding, or whether paid losses also show a consistent change after differences in settlement patterns have been considered. A comparison must also hold development age constant: comparing one accident period at 12 months with another at 24 months confounds accident-period differences with ordinary development.

Definition: For accident period $i$ and development age $j$, let $P_{i,j}$ be cumulative paid loss, $C_{i,j}$ be case outstanding loss, and $I_{i,j}$ be cumulative incurred loss. By definition, $I_{i,j}=P_{i,j}+C_{i,j}$. A reserve-adequacy distortion is a difference in the level or pattern of case estimates that changes incurred development relative to a reasonably comparable paid-loss pattern.

The comparison is strongest when periods have a common valuation basis, similar coverage and claim populations, and the same development age. Where settlement patterns have changed, use paid losses standardized for settlement rates, as in the earlier material on standardizing paid losses. Raw paid loss can otherwise change because claims are being settled at a different pace, not because ultimate loss emergence has changed.

Read the Three Patterns Together

At each development age, inspect paid, case outstanding, and incurred losses together. Useful descriptive measures include the case outstanding-to-paid ratio $C_{i,j}/P_{i,j}$ and the incurred-to-paid ratio $I_{i,j}/P_{i,j}$, provided paid loss is positive and large enough for the ratios to be stable. Because incurred equals paid plus case outstanding, these ratios satisfy:

\frac{I_{i,j}}{P_{i,j}}=1+\frac{C_{i,j}}{P_{i,j}}.

A higher ratio at a common age means more case outstanding relative to paid loss. It is evidence consistent with a difference in reserve adequacy, but it is not conclusive by itself: claim mix, claim severity, coverage, or other changes can also affect the relationship. Conversely, if standardized paid and incurred losses both rise in a similar proportion and the case-to-paid relationship remains stable, the pattern is more consistent with a change in loss emergence than with a reserve-adequacy shift alone.

Diagnostic principle: A reserve-adequacy concern is stronger when, at the same development age, incurred losses or case outstanding move materially while standardized paid losses do not move in a corresponding way. A genuine change in loss emergence is more plausible when paid and incurred measures move together and the case-reserve relationship remains broadly stable. Neither pattern proves its cause; the conclusion should be stated as a diagnosis supported by the available data.

This distinction matters because an incurred development factor can be distorted by changing case adequacy. If case estimates are strengthened in a later accident period, incurred loss may appear to develop more slowly or to start at a higher level, depending on the timing of the change. A mechanical comparison of incurred factors can then attribute a reserve-practice change to the loss process itself.

A Short Derivation of the Ratio Diagnostic

The ratio diagnostic follows directly from the definition of incurred loss. Suppose paid loss is positive. Since $I=P+C$, divide both sides by $P$:

\frac{I}{P}=\frac{P+C}{P} =\frac{P}{P}+\frac{C}{P} =1+\frac{C}{P}.

Thus, at a common age, a change in $I/P$ is exactly a change in $C/P$, not an independent piece of evidence. If paid loss is zero, the ratios are undefined; if paid is very small, they can be highly volatile. In either situation, compare dollar amounts, open-claim information, or other stable measures rather than relying on the ratio.

A useful normalization can be made when a reference period provides a credible case-reserve relationship for a comparable population. If the reference case-to-paid ratio is $r$, then the benchmark case outstanding for another period is $r$ times that period’s paid loss. Adding this benchmark to paid loss gives incurred on the reference adequacy basis. This is a diagnostic reconstruction, not an automatic replacement for recorded reserves: the reference relationship must be appropriate and the adjustment must not remove genuine differences in the losses.

Worked Diagnostics

Worked Example: A Higher Incurred Ratio at the Same Age

A hypothetical insurer compares two accident periods at 12 months. Losses are in thousands of dollars, and the paid amounts have already been standardized to a common settlement basis. Period A has paid loss of $400 and case outstanding of $200. Period B has paid loss of $400 and case outstanding of $320. Calculate incurred loss and the two ratios for each period, then interpret the pattern.

Method. Calculate $I=P+C$, then calculate $C/P$ and $I/P$ separately for each period. The common development age and standardized paid basis make the comparison more informative.

Period A. Incurred loss is $400+200=600$. The case-to-paid ratio is $200/400=0.50$. The incurred-to-paid ratio is $600/400=1.50$.

Period B. Incurred loss is $400+320=720$. The case-to-paid ratio is $320/400=0.80$. The incurred-to-paid ratio is $720/400=1.80$.

Interpretation and check. Standardized paid loss is unchanged, while case outstanding is $120,000 higher in Period B. The identity checks: $1+0.50=1.50$ for A and $1+0.80=1.80$ for B. The pattern is consistent with higher case reserves relative to paid loss in B, rather than a paid-loss emergence increase. It does not establish that reserve adequacy is the only difference.

Common mistake avoided: Comparing incurred losses alone and calling the 20% increase from 600 to 720 a 20% increase in ultimate loss. The increase is entirely in case outstanding in these data; further evidence is needed before attributing it to ultimate losses.

Worked Example: Normalizing to a Reference Adequacy

At 18 months, a hypothetical portfolio has a reference accident period with standardized paid loss of $500,000 and case outstanding of $250,000. A later period has standardized paid loss of $600,000 and recorded case outstanding of $480,000. Assume, solely for this diagnostic, that the reference case-to-paid ratio is suitable for the later period. Find the reference ratio, benchmark case outstanding, adequacy-normalized incurred loss, and recorded incurred loss.

Step 1: Find the reference ratio. The reference case-to-paid ratio is $250/500=0.50$.

Step 2: Apply the reference ratio to later-period paid loss. Benchmark case outstanding is $0.50\times600=300$, or $300,000.

Step 3: Calculate normalized and recorded incurred amounts. Adequacy-normalized incurred loss is $600+300=900$, or $900,000. Recorded incurred loss is $600+480=1{,}080$, or $1,080,000.

Step 4: Compare. Recorded case outstanding exceeds the benchmark by $480-300=180$, or $180,000. The recorded incurred amount is consequently $180,000 above the diagnostic normalized amount.

Check and interpretation. The reference ratio reproduces the reference incurred-to-paid ratio: $1+0.50=1.50$. The later normalized ratio is also $900/600=1.50$, while recorded later incurred-to-paid is $1{,}080/600=1.80$. The excess is in case outstanding. This suggests a possible adequacy difference; it does not show that $900,000 is the correct ultimate loss.

Common mistake avoided: Applying the reference ratio to recorded incurred loss instead of paid loss. The ratio is $C/P$, so the benchmark is $0.50\times600$, not $0.50\times1{,}080$.

Worked Example: Coordinated Paid and Incurred Increases

Two hypothetical accident periods are compared at 12 months on a common settlement basis. Period A has paid loss of $400,000 and case outstanding of $200,000. Period B has paid loss of $500,000 and case outstanding of $250,000. Calculate incurred loss and compare the scale of change and the case-reserve relationship.

Period A. Incurred loss is $400+200=600$, or $600,000. The case-to-paid ratio is $200/400=0.50$, and incurred-to-paid is $600/400=1.50$.

Period B. Incurred loss is $500+250=750$, or $750,000. The case-to-paid ratio is $250/500=0.50$, and incurred-to-paid is $750/500=1.50$.

Compare changes. Paid loss rises by $100/400=25\%$. Case outstanding rises by $50/200=25\%$. Incurred loss rises by $150/600=25\%$. Both periods have the same case-to-paid and incurred-to-paid ratios.

Check and interpretation. In each period, $I=P+C$, and $I/P=1+C/P=1.50$. The coordinated 25% increases are more consistent with a change in loss emergence than with a reserve-adequacy shift alone, because case outstanding has not increased relative to paid. The data still cannot identify the underlying cause of the emergence change.

Common mistake avoided: Concluding that any increase in incurred loss signals weaker case adequacy. Here paid loss rises by the same proportion and the reserve relationship is unchanged.

Control the Comparison Before Drawing a Conclusion

A sound diagnosis compares like with like. Use the same development age, valuation basis, and coverage definition. Check whether paid amounts are affected by known settlement-rate changes; if they are, a raw paid comparison may be misleading. The earlier material on rebuilding an adjusted incurred triangle provides the framework for placing case estimates on a common adequacy basis when the evidence supports that adjustment.

Also distinguish a change in reserve adequacy from a change in the level of losses. If later periods have more paid loss and more case outstanding in similar proportions, the evidence does not point to a case-reserve shift by itself. If incurred changes substantially while standardized paid stays comparatively stable, investigate whether case estimates have changed. If the available patterns conflict, report the ambiguity rather than forcing a single explanation.

Align the observations.
Compare accident periods at the same development age and on a consistent valuation and coverage basis.

Check the paid basis.
Determine whether settlement-rate differences could explain changes in paid loss; use standardized paid amounts when appropriate.

Separate paid from case outstanding.
Calculate incurred as paid plus case outstanding and inspect how each component changed.

State a qualified diagnosis.
Describe whether the pattern is more consistent with adequacy distortion or changed emergence, and state what the triangle cannot establish.

Common Exam Trap

A frequent error is to compare different maturity ages and interpret the older period’s higher paid amount as evidence of greater ultimate emergence. For instance, a period with $700,000 paid at 24 months cannot be directly compared with another period’s $500,000 paid at 12 months to diagnose adequacy. The extra development time alone can explain part or all of the difference. Make the age comparison first; then examine whether paid and case components differ.

A second trap is treating $I/P$ and $C/P$ as two independent confirmations. Since $I/P=1+C/P$, they express the same relationship. Use the ratios to communicate the pattern, but seek separate evidence from the standardized paid triangle, claim information, or other relevant context.

Key takeaway: At a common development age, isolate the case-reserve component and compare it with paid loss on a consistent basis. A divergence between incurred and standardized paid patterns is a reserve-adequacy warning; coordinated movement with a stable case-to-paid relationship points more toward changed loss emergence. Treat both as diagnostics, not proof.

Check Your Understanding

For each problem, state what the observed pattern supports; do not claim that a triangle alone proves the cause.

At 12 months, two hypothetical periods have standardized paid losses of $300,000 each. Case outstanding is $120,000 for the first and $210,000 for the second. Find each incurred-to-paid ratio and identify the diagnostic signal. (Answer: 1.40 and 1.70; higher case reserves relative to unchanged paid loss in the second period, consistent with an adequacy difference.)
At 24 months, a reference period has paid loss of $800,000 and case outstanding of $400,000. A later period has standardized paid loss of $1,000,000 and case outstanding of $500,000. Find each incurred amount and case-to-paid ratio, then interpret. (Answer: incurred losses are $1,200,000 and $1,500,000; both case-to-paid ratios are 0.50, so the proportional increase is more consistent with changed emergence than an adequacy shift alone.)
A reference period has paid loss of $450,000 and case outstanding of $180,000 at a common age. Another period has standardized paid loss of $600,000 and recorded case outstanding of $360,000. Using the reference relationship, calculate benchmark case outstanding and normalized incurred loss. (Answer: reference ratio 0.40; benchmark case outstanding $240,000; normalized incurred loss $840,000.)
At the same development age, standardized paid loss rises from $500,000 to $550,000, while case outstanding rises from $250,000 to $400,000. Calculate the incurred amounts and the case-to-paid ratios. What is the main diagnostic concern? (Answer: incurred losses are $750,000 and $950,000; case-to-paid ratios are 0.50 and approximately 0.7273; the case component has risen more than paid loss, raising a reserve-adequacy concern.)
One accident period has paid loss of $420,000 and case outstanding of $210,000 at 12 months. Another has paid loss of $700,000 and case outstanding of $280,000 at 24 months. Can these figures alone establish a reserve-adequacy difference? Explain briefly. (Answer: No; the ages differ, so ordinary development is confounded with accident-period differences.)

What You'll Learn