Every reserving method assumes the future looks like the past. When the company changes how it settles claims, who it insures, or how it recovers money, that assumption breaks, and your unpaid loss estimate quietly goes wrong unless you spot the change.
A chain-ladder (development) method multiplies immature losses by factors derived from history. A Bornhuetter-Ferguson (BF) method blends those factors with an a priori expected loss. Both trust that the data-generating process is stable. An operational change is a break in that process. Your job is to detect it, judge its direction, and either adjust the data or switch to a method that is immune.
Claims departments move two things that reserving methods care about: how adequately open claims are reserved, and how fast claims close.
Case reserve adequacy. If adjusters strengthen case reserves, reported (incurred) losses at recent maturities jump for reasons unrelated to real loss cost. Incurred age-to-age factors were built when reserves were weaker, so applying them to the newly inflated incurred losses double-counts the strengthening.
Common mistakes
- Trusting incurred after a case-reserve strengthening. The $11.68M incurred answer looks precise but embeds the 24% strengthening twice. The paid method or a restated incurred is correct.
- Reading one diagnostic as a clean signal. A flat paid-to-reported ratio can mask a strengthening and a speed-up that offset, and a rising average case can be a shifted open-claim mix, not stronger reserves. Confirm the other levers held still first.
- Restating the wrong diagonal in Berquist-Sherman. The formal method holds the latest diagonal fixed and lifts the older, under-adequate diagonals up to it, then rederives lower factors. Pushing the latest diagonal down is only a rough consistency check.
Bottom line
- Development methods assume stable patterns; any operational change that shifts the pattern distorts the estimate.
- Case reserve strengthening inflates incurred losses, so an unadjusted incurred (development) method overstates ultimate and IBNR.
- Faster claim settlement accelerates the paid pattern, so an unadjusted paid method overstates ultimate; slower settlement understates it.
- Diagnose case adequacy with average case reserves by maturity; diagnose settlement speed with disposal (closure) rates.
Exam shortcut
If average case reserves outrun severity trend, suspect strengthening: the incurred method overstates, so lean paid or Berquist-Sherman restated. Before you quote a direction from any single diagnostic, ask what else moved: offsetting strengthening and speed-up can flatten the paid-to-reported ratio, and a speed-up can raise average case through open-claim mix. Confirm a consistent claim-count definition before trusting count diagnostics.
The full lesson (about 2,239 words, 15 min read) adds 2 worked examples, all 8 common mistakes, a self-check, free in the app.
Learning objectives
- B10
Browse all free Exam 5 lessons or jump into free Exam 5 practice questions.