Claims Development and Reserving

Basic Chain Ladder

Calculate a deterministic chain-ladder reserve and expose the assumptions behind every step.

Basic Chain Ladder

The deterministic chain ladder estimates how a cumulative triangle will develop if historical age-to-age patterns continue.

1. Start with cumulative observations

Let Ci,jC_{i,j} be cumulative paid loss for origin year ii at development age jj.

AYDev 0Dev 1Dev 2Dev 3
2020100180240280
2021120220300
2022140260
2023160

Units are £m. The triangle is cumulative, paid, nominal, gross of reinsurance, and valued at year-end 2023.

2. Estimate age-to-age factors

The usual volume-weighted factor is a ratio of column sums over rows for which both ages are observed:

f^j=iIjCi,j+1iIjCi,j.\widehat f_j= \frac{\sum_{i\in I_j}C_{i,j+1}} {\sum_{i\in I_j}C_{i,j}}.

For the example:

TransitionCalculationFactorData supporting it
0 → 1(180+220+260)/(100+120+140)(180+220+260)/(100+120+140)1.83333 origin years
1 → 2(240+300)/(180+220)(240+300)/(180+220)1.35002 origin years
2 → 3280/240280/2401.16671 origin year

Volume weighting gives larger cells more influence. It is not automatically superior to an arithmetic or credibility-weighted selection; it is a modelling choice whose sensitivity should be shown.

fj>1f_j>1 is a link ratio or age-to-age development factor. Its reciprocal is a completion proportion only under the selected development pattern. Calling both quantities a “grossing-up factor” invites direction errors.

The cumulative development factor (CDF) from age jj to ultimate age JJ is

F^j=k=jJ1f^k,p^j=1F^j,\widehat F_j=\prod_{k=j}^{J-1}\widehat f_k, \qquad \widehat p_j=\frac{1}{\widehat F_j},

where pjp_j is the implied percentage paid by age jj.

Latest ageCDF to Dev 3Implied paid proportion
01.8333×1.35×1.1667=2.88751.8333\times1.35\times1.1667=2.887534.63%
11.35×1.1667=1.57501.35\times1.1667=1.575063.49%
21.16671.166785.71%
311100%

A “100%” at Dev 3 is an assumption of this example. Real long-tailed business generally requires a selected tail factor beyond the observed triangle.

4. Project ultimate loss and reserve

If the latest observed value for origin year ii is Ci,jiC_{i,j_i},

U^i=Ci,jiF^ji,R^i=U^iCi,ji.\widehat U_i=C_{i,j_i}\widehat F_{j_i}, \qquad \widehat R_i=\widehat U_i-C_{i,j_i}.
AYLatest ageLatest paidCDFUltimateReserve
20203280.01.0000280.00.0
20212300.01.1667350.050.0
20221260.01.5750409.5149.5
20230160.02.8875462.0302.0
Total1,000.01,501.5501.5

Why the youngest year dominates

AY 2023 supplies only £160m of observed paid loss yet contributes £302m of reserve. It is both the least mature and the most leveraged by assumptions: a 5% change in its selected CDF changes its ultimate by about £23.1m. This is why reserve review should focus on influence, not merely on the number of rows.

5. Reproduce the calculation

Py

Chain-ladder calculation and factor sensitivity

Idle

6. State the assumptions as testable claims

The chain ladder does not merely assume “the past repeats.” It requires more precise claims:

AssumptionObservable warning signResponse
comparable origin yearsexposure or mix changes sharplyon-level, segment, or use an exposure-based method
stable development by agelink ratios trend by origin yearinvestigate operations, inflation, or model calendar effects
no unmodelled calendar shockresiduals align on diagonalsisolate shock or use a model with calendar terms
adequate volumeone cell controls a factoruse credibility, benchmarks, or a range
latest cumulative value is meaningfullarge recoveries/reopeningsreconcile and model the mechanism
selected tail is adequateoldest years still developfit or benchmark a tail factor

7. Three quick diagnostics

Leave-one-origin-year-out

Re-estimate each factor after removing one eligible origin year. Large movement identifies influential observations; it does not automatically justify deleting them.

Backtest a historical diagonal

Pretend the latest diagonal was unavailable, project it using only older data, and compare projection with what actually emerged. Repeat across diagonals where possible.

Compare paid and incurred

Agreement is not proof, but divergence can reveal settlement-speed changes or case-reserving shifts. Define incurred consistently before comparing.

Practice

  1. If f0:1f_{0:1} is unchanged but f1:2f_{1:2} rises from 1.35 to 1.42, calculate AY 2022 ultimate and reserve.
  2. Why is the Dev 2→3 factor especially uncertain here?
  3. A £20m payment is moved from Dev 1 to Dev 0 without changing ultimate loss. Predict the direction of the early link ratios.
Answers
  1. Ultimate =260×1.42×1.1667430.7=260\times1.42\times1.1667\approx430.7; reserve 170.7\approx170.7.
  2. Only AY 2020 supplies a fully observed pair, so process variation and that year's peculiarities are inseparable from the selected factor.
  3. Dev 0 rises and Dev 1 cumulative is unchanged, so the 0→1 ratio falls. The change is payment timing, not necessarily ultimate severity.

Next, separate claim emergence from average claim cost.

Copyright © 2026