Exam guide

ASTAM Reserving and Pricing

ASTAM reserving and pricing is the largest syllabus block. It joins outstanding-claims methods, stochastic reserve models, trend, overall rate indications, and risk-classification differential changes.

Credential side
SOA
Primary intent
ASTAM reserving and pricing
Best next page
Stochastic Reserving: Mack, ODP, and Bootstrap

Quick Answer

The Spring 2026 ASTAM syllabus gives reserving and pricing a 15-29% weight, the largest range on the exam. It names expected loss ratio, chain-ladder, Bornhuetter-Ferguson, Bayesian, frequency-severity reserving, Mack, Poisson, overdispersed Poisson, trend, loss cost, loss ratio, and risk-classification differential changes.

This block is where ASTAM looks most like day-to-day short-term actuarial work: incomplete triangles, pricing indications, trend assumptions, and explaining what an estimate can and cannot support.

Reserve Method Map

Expected loss ratio is prior-driven. Chain-ladder is experience-driven. Bornhuetter-Ferguson blends expected loss with reported emergence. Mack adds variance under distribution-free assumptions. Poisson and ODP models turn triangle development into a statistical model.

A good answer identifies which method is being used and why it fits the data maturity. Immature years usually need more prior weight; mature years usually let reported development carry more of the estimate.

Pricing Method Map

Loss cost and loss ratio methods are two ways to express the same rate-change problem. Trend moves historical losses or premiums to the future cost level. Risk-classification work then allocates the overall indication across classes while balancing back to the portfolio target.

ASTAM can ask for a numeric indication and a short explanation. The explanation should name whether the change is driven by loss level, exposure mix, premium adequacy, or classification relativities.

Original Practice Drill

Build a three-accident-year cumulative paid triangle. Estimate ultimates by chain-ladder, then repeat with Bornhuetter-Ferguson using an expected loss ratio assumption. Finally, state which estimate you would rely on for the newest year and why.

For a second pass, add one trend assumption and ask for an overall rate change by both loss cost and loss ratio methods. The two answers should agree if the exposure and premium bases are handled consistently.

Common Traps

Trap 1: treating Mack's standard error as a guarantee. It is conditional on the model assumptions and the observed triangle.

Trap 2: using chain-ladder mechanically when the newest accident year has little development or a known operational shift.

Trap 3: calculating an indicated rate change but never saying whether it is an overall rate-level change or a class differential change.

Practice

Original Source-Backed Practice

3 questions built from syllabus outcomes and released-exam patterns. The prompts and answers are original, so they train the skill without copying official exam text.

ASTAM Reserving and Pricing Drill

Written-answer practice for chain-ladder, Bornhuetter-Ferguson, Mack caveats, and rate indications.

ASTAM - 22 min
Source pattern: SOA ASTAM reserving and pricing outcomes; original reserve and rate prompts.
  1. Question 1/Written Answer

    Newest accident year

    For the newest accident year in a sparse paid triangle, why might Bornhuetter-Ferguson be preferable to pure chain-ladder?

    Solution And Grading Points

    The newest year has little reported development, so pure chain-ladder can overreact to sparse early data. Bornhuetter-Ferguson blends expected loss with emergence and usually gives a steadier estimate.

    • Identifies sparse maturity as the issue.
    • Explains chain-ladder sensitivity.
    • Explains the prior-plus-emergence logic of Bornhuetter-Ferguson.
  2. Question 2/Flashcard

    Mack standard error caveat

    What should be said when reporting a Mack reserve standard error?

    Solution And Grading Points

    It is conditional on the Mack model assumptions and the observed triangle. It is not a guarantee and does not cover every operational or model-risk source.

    • States conditional-on-model language.
    • Mentions the observed triangle.
    • Avoids treating the standard error as a guarantee.
  3. Question 3/Written Answer

    Loss cost versus loss ratio check

    When should loss cost and loss ratio rate indications agree?

    Solution And Grading Points

    They should agree when exposures, premium, trend, on-leveling, expenses, and profit provisions are handled on consistent bases. Disagreement usually signals a basis mismatch.

    • Names consistency of exposure and premium bases.
    • Mentions trend or on-leveling.
    • Interprets disagreement as a diagnostic.

References And Official Sources