Why Case Reserve Changes Matter
An incurred link ratio reflects more than the payment of claims. It also reflects changes in the case reserves attached to claims that remain open. If claim handlers begin setting higher case reserves, or if an insurer releases reserves at a later valuation, incurred losses can develop differently even when paid losses follow the same pattern. A change in an incurred link ratio can therefore be a signal to investigate reserving practices, not necessarily evidence that the underlying claims have changed.
The central diagnostic is to compare incurred development with paid development over the same valuation ages and experience. That comparison is useful only when it is interpreted alongside the movement in case reserves and relevant changes in the claims being observed.
A case reserve is an estimate of the unpaid amount associated with a reported claim. It is not the same as a payment, and it may change as information about the claim develops or as reserving practices change. Incurred losses can consequently increase or decrease between valuations: a release of case reserves may outweigh new payments during that interval.
The examples below use deterministic hypothetical loss triangles, with dollar amounts stated in thousands. No interest, mortality, or probability-distribution assumption is needed for these link-ratio calculations.
Decomposing Incurred Development
For one accident year, write the changes from age \(j\) to age \(j+1\) as \(\Delta P=P_{j+1}-P_j\) and \(\Delta R=R_{j+1}-R_j\). Since incurred losses equal paid losses plus case reserves at both ages, the change in incurred losses is the sum of the paid emergence and the change in case reserves.
Dividing by \(I_j\), when \(I_j>0\), gives the incurred link ratio:
This identity explains why a reserve change affects incurred development. A positive \(\Delta R\) adds to incurred emergence; a negative \(\Delta R\) offsets paid emergence. It also shows why the same reserve movement can have different effects depending on the size of incurred losses at the starting age: the reserve change is measured relative to that starting incurred amount.
The comparison with the paid link ratio gives a more specific diagnostic. Suppose \(P_j>0\), \(R_j>0\), and the relevant totals are positive. Then the paid link ratio is \(1+\Delta P/P_j\), while the incurred ratio is \(1+(\Delta P+\Delta R)/(P_j+R_j)\). Cross-multiplication shows that the incurred ratio exceeds the paid ratio exactly when \(P_j\Delta R>R_j\Delta P\). Equivalently, the proportional change in case reserves exceeds the proportional paid emergence. This is an algebraic comparison, not a claim that reserve changes are the only possible cause of a difference between paid and incurred development.
The equivalence requires positive starting paid losses and case reserves. If either is zero, or if the data include negative values such as recoveries or adjustments, use the original ratios and inspect the underlying amounts rather than relying on this proportional-change version.
Worked Example: A Reserving Change Raises an Incurred Link Ratio
A hypothetical insurer reviews two accident years at ages 12 and 24 months. At age 12, aggregate paid losses are $1,000 and case reserves are $500. At age 24, paid losses are $1,500 and case reserves are $1,000. Find the paid and incurred link ratios, decompose incurred emergence, and interpret the difference.
Method. Calculate cumulative incurred losses at each age as paid losses plus case reserves, then calculate both age-to-age ratios. The data are aggregate amounts in thousands.
Age 12 incurred losses: \(I_{12}=1{,}000+500=1{,}500\).
Age 24 incurred losses: \(I_{24}=1{,}500+1{,}000=2{,}500\).
Paid link ratio: \(1{,}500/1{,}000=1.500\).
Incurred link ratio: \(2{,}500/1{,}500=1.667\), rounded to three decimals.
Paid emergence is \(1{,}500-1{,}000=500\). The case reserve change is \(1{,}000-500=500\). Therefore, incurred emergence is \(500+500=1{,}000\), and \(1{,}500+1{,}000=2{,}500\), which checks against the age-24 incurred total. Here the increase in case reserves contributes to the higher incurred ratio relative to the paid ratio.
Common mistake avoided: Using only the paid increase to explain incurred emergence would omit the additional $500 of case reserves. The ratio difference is a reason to investigate the reserve movement; it does not, by itself, prove that a reserving change caused the difference.
Reading Paid and Incurred Development Together
Paid and incurred link ratios measure related but distinct patterns. Paid development reflects the timing and amount of payments, while incurred development reflects paid amounts plus current estimates of unpaid claims. The comparison is most useful when the ratios use the same accident years, valuation dates, coverage, and loss basis. Otherwise, differences in the populations or definitions can masquerade as a reserve signal.
A higher incurred link ratio than paid link ratio may be consistent with case reserves increasing relatively rapidly between the two ages. A lower incurred ratio may be consistent with case reserves declining, or simply increasing less rapidly than paid losses. But the ratios alone cannot identify why the reserves changed. Claim mix, large claims, claim closures, reopened claims, coverage changes, and valuation timing may also matter.
Worked Example: A Reserve Release Produces Low Incurred Development
For a hypothetical accident year, cumulative paid losses at age 18 months are $700 and case reserves are $500. At age 30 months, cumulative paid losses are $980 and case reserves are $100. Calculate both link ratios and reconcile the incurred change. Amounts are in thousands.
Age 18 incurred: \(I_{18}=700+500=1{,}200\).
Age 30 incurred: \(I_{30}=980+100=1{,}080\).
Paid link ratio: \(980/700=1.400\).
Incurred link ratio: \(1{,}080/1{,}200=0.900\).
Paid emergence is \(980-700=280\). The case reserve change is \(100-500=-400\), a decrease of $400. Incurred emergence is therefore \(280-400=-120\). Adding this change to the starting incurred amount gives \(1{,}200-120=1{,}080\), matching the age-30 total.
The incurred ratio is below 1 even though cumulative paid losses increased. The reserve release more than offset the paid emergence. The calculation is internally consistent: \(1{,}080/1{,}200=0.900\).
Common mistake avoided: Assuming every incurred link ratio must be at least 1 confuses incurred losses with cumulative paid losses. Incurred losses can decline when case reserves are reduced enough; the ratio should be interpreted with the reserve movement in view.
A Pooled Triangle Diagnostic
For a selected group of accident years, a volume-weighted link ratio is calculated as the ratio of aggregate losses at the later age to aggregate losses at the earlier age. This makes it possible to reconcile the pooled incurred ratio with total paid emergence and total case reserve change. The group should be kept consistent across the two ages.
For each age pair, divide the later-age aggregate paid losses by the earlier-age aggregate paid losses.
Add case reserves to paid losses at each age, then divide the later-age incurred total by the earlier-age total.
Check whether the change in incurred losses equals paid emergence plus the change in case reserves.
Use the comparison to identify a possible reserve signal, then examine the underlying claims and valuation context before attributing a cause.
Worked Example: Comparing Pooled Ratios Across Two Age Intervals
A hypothetical two-accident-year triangle has the following aggregate totals, in thousands. At ages 12, 24, and 36 months, paid losses are respectively $1,000, $1,500, and $1,800. Case reserves at those ages are respectively $500, $1,200, and $1,090. Calculate paid and incurred link ratios for both intervals and use the case reserve movements to interpret them.
Step 1: Calculate incurred totals. At age 12, incurred is \(1{,}000+500=1{,}500\). At age 24, it is \(1{,}500+1{,}200=2{,}700\). At age 36, it is \(1{,}800+1{,}090=2{,}890\).
Step 2: Calculate age 12-to-24 ratios. The paid ratio is \(1{,}500/1{,}000=1.500\). The incurred ratio is \(2{,}700/1{,}500=1.800\).
Step 3: Reconcile the first interval. Paid emergence is \(1{,}500-1{,}000=500\), and case reserves increase by \(1{,}200-500=700\). Total incurred emergence is \(500+700=1{,}200\), which equals \(2{,}700-1{,}500\).
Step 4: Calculate age 24-to-36 ratios. The paid ratio is \(1{,}800/1{,}500=1.200\). The incurred ratio is \(2{,}890/2{,}700=1.070\), rounded to three decimals.
Step 5: Reconcile the second interval. Paid emergence is \(1{,}800-1{,}500=300\). Case reserves decrease by \(1{,}090-1{,}200=-110\). Incurred emergence is therefore \(300-110=190\), and \(2{,}700+190=2{,}890\).
The first interval’s incurred ratio is substantially above its paid ratio, alongside a $700 increase in case reserves. In the second interval, the $110 reserve decrease offsets part of the $300 paid emergence, so incurred development is lower than paid development. These movements are consistent with reserve changes influencing the incurred ratios; further claim-level investigation would be needed to establish the reasons.
Common mistake avoided: Comparing one incurred age pair with a paid age pair based on a different set of accident years would not isolate the reserve signal. Here each comparison uses the same pooled population and the same valuation interval.
Limits of the Diagnostic
A paid-to-incurred comparison is a screening tool, not a mechanical adjustment rule. Replacing an incurred factor with a paid factor solely because the two differ can be inappropriate: payment timing may be volatile, and paid development can have its own distortions. Nor does a high incurred ratio automatically mean that case reserves were strengthened. The ratio is affected by both its starting denominator and the changes in paid losses and case reserves.
A useful review checks whether the apparent pattern persists across accident years and valuation periods, and whether a change in claims handling or case-reserving practice occurred at a relevant time. The review should also reconcile aggregate ratios to claim-level or reserve-level movements where possible. If large claims or changes in claim composition drive the result, an aggregate ratio may conceal the underlying explanation.
A particularly important denominator trap is to compare reserve changes in dollars without considering the starting incurred amount. A $100 reserve increase has a larger proportional effect on an incurred base of $500 than on a base of $5,000. Another trap is to assume that a change affecting both ages equally must change the ratio. If both incurred amounts are multiplied by the same positive factor, their ratio is unchanged. More generally, a change to both valuation-age amounts alters the ratio according to its relative effect at each age, not merely because the amounts changed.
Check Your Understanding
For each hypothetical case, calculate or interpret the ratios using the stated amounts. All amounts are in thousands.
- Paid losses rise from $500 to $700, while case reserves fall from $300 to $250. Find the paid and incurred link ratios. (Answer: Paid ratio \(1.400\); incurred ratio \(1.188\), rounded to three decimals.)
- Paid losses rise from $450 to $675, while case reserves rise from $150 to $300. Calculate both ratios and the amount of incurred emergence attributable to case reserve change. (Answer: Paid ratio \(1.500\); incurred ratio \(1.625\); case reserve increase \(150\).)
- Aggregate paid losses rise from $900 to $1,170 and case reserves fall from $360 to $270. Calculate the paid and incurred link ratios and reconcile incurred emergence. (Answer: Paid ratio \(1.300\); incurred ratio \(1.143\), rounded to three decimals; incurred emergence \(180=270-90\).)
- Incurred losses at two ages are $1,000 and $1,500. A revised reserving method would increase those amounts by $200 and $300, respectively. Calculate the revised ratio and state whether it changes. (Answer: Revised ratio \(1{,}800/1{,}200=1.500\), unchanged from \(1{,}500/1{,}000=1.500\).)
- Starting incurred losses are $1,000. Paid losses emerge by $250, while case reserves decrease by $290. Calculate the ending incurred amount and incurred link ratio, and explain why the ratio is below 1. (Answer: Ending incurred \(960\); ratio \(0.960\); the $290 reserve release exceeds the $250 paid emergence by $40.)