Actuarial Notes

Claims Processing Changes

Study guide: Exam 5

Claims Processing Changes are changes in how claims are handled — staffing and caseloads, settlement authority, use of defence counsel, closure targets, outsourcing, reserving philosophy — that alter the timing of payments and case reserve changes, and therefore the shape of the development pattern.

\text{Disposal Rate}_n = \frac{\text{Closed claims at age } n}{\text{Ultimate claim counts}}

\text{Avg Case O/S} = \frac{\text{Case Reserves}}{\text{Open Claims}}

  • Processing changes hit the diagonal: they affect every open accident year in the calendar period when the change happens, which is what distinguishes them from a change in one cohort's experience.
  • The two diagnostics above separate the two kinds of change. A moving disposal rate is a settlement speed change and distorts paid development; a jump in average case outstanding is a case adequacy change and distorts reported development. Both call for the corresponding Berquist-Sherman adjustment.
  • Faster closing inflates the recent paid diagonal, so historical paid factors over-develop it. Slower closing does the reverse. The intuitive answer is usually the wrong sign, which is why the disposal-rate table is computed rather than reasoned about.
  • Unlike Claims Coding Changes, which move losses between categories, processing changes genuinely change the emergence curve — so post-change data is describing a different process, not a relabelled one.
  • Watch for closed-without-payment activity: a drive to clear old files can raise the closure rate without moving any money, which changes the disposal rate but not the paid pattern. Counting CWP separately reveals it.

A Fast-Track Settlement Programme An insurer introduces fast-track settlement for auto bodily injury in 2022, cutting average time-to-close from 18 months to 9. Historical 12–24 paid factors for AY 2018–2021 averaged 1.45; AY 2022–2023 are coming in near 1.20.

What happens if the historical factor is applied to the recent years?

Answer More of each recent year's ultimate is already paid at 12 months, so the 12-month paid figure is high relative to its ultimate. Applying 1.45 — a factor calibrated to a book that had paid much less by 12 months — overstates ultimate losses for AY 2022 onward.

The observed 1.20 is the correct factor for the new process, but two years of it is thin evidence and the change may still be bedding in.

Options: restate the historical paid triangle to the current disposal pattern (Berquist-Sherman) so the full history is usable at the new pace; or lean on reported development, which is largely unaffected by closing speed, as the primary method while the paid series accumulates.

Two Changes at Once Diagnostics at 24 months across four accident years:

AY Disposal rate Avg case O/S Paid ÷ reported
2021 56\% \$11{,}000 0.44
2022 57\% \$11{,}500 0.45
2023 66\% \$14{,}900 0.52
2024 71\% \$18{,}200 0.56

What is happening, and what does it do to the methods?

Answer Both things are happening at once, starting in 2023:

  • Disposal rate up from 57\% to 71\% — claims closing faster.
  • Average case outstanding up from \$11{,}500 to \$18{,}200, far beyond any plausible severity trend — case reserves strengthened.

That combination is common in practice (a new claims leader often does both) and it is the hardest case, because it distorts both triangles at once:

  • Paid development is overstated by the faster closing.
  • Reported development is overstated by the strengthening.

Neither triangle can be used as the clean check on the other, which is the usual escape route. What remains:

  1. Apply both Berquist-Sherman adjustments — restate the paid triangle to a common disposal rate and the reported triangle to current case adequacy — and compare the two restated results.
  2. Lean on methods that do not depend on the triangle's recent diagonal at all: the expected claims technique and BF with an externally derived ELR.
  3. Use frequency-severity on closed claims with a disposal-rate projection, which handles the settlement change explicitly.

And document heavily: with both distortions active, the estimate rests far more on judgment than usual, and the report should say so.