CAA Fellow Actuary (China) common mistakes
CAA Fellow Actuary (China) · Understand the traps and how to avoid them.
Calculating a life-insurance net premium or reserve with the wrong benefit-timing convention - e.g., paying the death benefit at the end of the year of death when the contract specifies payment at the moment of death, or discounting an annuity-due as if it were an annuity-immediate.
Candidates memorize a single formula per product and default to it under time pressure, without re-reading the benefit clause and payment timing stated in the question. The difference between A_x and A-bar_x, or between a-double-dot and a, is a single factor that is easy to drop.
What to do instead
Before computing, write down the payment timing explicitly (immediate vs. due, continuous vs. discrete) and the corresponding actuarial symbol. Build a one-page cheat sheet mapping each product to its symbol and timing, and practice converting between them using i and the force of mortality.
Using a single deterministic mortality/lapse assumption in a reserve or pricing calculation when the question asks for a sensitivity, scenario, or best-estimate-plus-risk-margin result, and reporting only the central figure.
Candidates treat the problem as a pure formula exercise and overlook the instruction to reflect uncertainty, or they confuse a deterministic reserve with a stochastic best-estimate liability.
What to do instead
Re-read the ask and underline every word about scenarios, margins, or sensitivity. Practice setting up a small scenario table (base, adverse mortality, adverse lapse) and reporting the range, not just one number.
Applying a reserve or premium basis that mixes best-estimate assumptions with prudent margins inconsistently - e.g., using best-estimate mortality but a prudential discount rate, or ignoring the prescribed valuation basis in favor of pricing assumptions.
Candidates conflate pricing, valuation, and solvency bases, which serve different purposes and use different margins. Under time pressure they grab whichever assumption set they remember.
What to do instead
For each problem, first identify the purpose (pricing, statutory reserve, solvency capital, IFRS 17 fulfilment cash flows) and select the corresponding basis. Keep a comparison table of bases and practice stating which basis applies before calculating.
Computing a health-insurance premium or reserve without separating claim incidence, claim severity, and the effect of deductibles/coinsurance, treating the product like a simple life contract.
Health products have frequency-severity structure and cost-sharing features that life formulas do not capture; candidates trained on life math skip the decomposition.
What to do instead
Always decompose expected claim cost into frequency × severity, then apply deductible and coinsurance adjustments explicitly. Practice a worked example with a deductible and a coinsurance percentage before attempting full premium calculations.
Estimating incurred-but-not-reported (IBNR) reserves with a development-factor method without checking that the development pattern is stable, and reporting the point estimate without a range or diagnostic.
Candidates apply the chain-ladder mechanically because it is the most familiar method, ignoring changes in case-reserving practice, mix of business, or claims inflation that distort the factors.
What to do instead
Before applying chain-ladder, test factor stability across accident years and note any outliers. Cross-check with an expected-loss-ratio or Bornhuetter-Ferguson method and present a range with the rationale for the selected estimate.
Calculating a pension liability using the projected unit credit method but discounting at the wrong rate or omitting the attribution of benefit to past service, producing a liability that does not reconcile to the accrued benefit obligation.
Candidates mix the accumulated benefit obligation, projected benefit obligation, and vested benefit obligation, and use the asset return instead of the discount rate.
What to do instead
Write the definition of each obligation before computing, confirm the discount rate is a high-quality corporate bond yield (or the prescribed rate), and reconcile the service cost, interest cost, and benefit payments line by line.
Ignoring the duration/convexity mismatch between assets and liabilities when asked to assess interest-rate risk, and recommending an asset allocation based only on expected return.
Candidates focus on return maximization and treat ALM as a portfolio-selection exercise, forgetting that surplus volatility is driven by the gap in interest-rate sensitivity.
What to do instead
For every ALM question, compute or estimate asset and liability durations, state the surplus duration, and evaluate the allocation against surplus volatility as well as return. Practice immunization conditions (duration match, convexity, and present-value match).
Treating an insurance contract's accounting result as equal to its economic or solvency result - e.g., reporting profit at inception under IFRS 17 or confusing statutory reserves with best-estimate liabilities.
Candidates conflate the three reporting frameworks and assume accounting profit equals economic profit. IFRS 17's contractual service margin and the deferral of acquisition costs are commonly missed.
What to do instead
Maintain a three-column comparison (statutory, IFRS 17, solvency) for recognition, measurement, and profit emergence. For IFRS 17, always identify the CSM, risk adjustment, and fulfilment cash flows separately before stating any profit figure.
Answering an ethics or conduct question by citing the legally permissible action rather than the professionally required one, or failing to identify the conflict of interest and the required disclosure.
Candidates default to legal compliance and overlook the actuarial profession's higher standards on integrity, objectivity, and disclosure of conflicts.
What to do instead
For conduct questions, apply a structured test: identify the duty owed, the conflict, the disclosure required, and the escalation path. Practice distinguishing 'legal' from 'professionally appropriate' and always state the disclosure obligation explicitly.
Computing a solvency capital requirement by summing standalone risks instead of aggregating them with a correlation matrix or copula, and ignoring diversification benefits.
Candidates treat the SCR as a simple sum of risk modules because aggregation formulas are easy to misremember, and they forget that the standard formula uses prescribed correlations.
What to do instead
Memorize the aggregation structure (correlation matrix across modules, then within modules) and practice a two-module aggregation by hand. Always state the diversification benefit as the difference between the sum and the aggregated figure.
Building a model with hard-coded assumptions scattered across cells and no separation between inputs, calculations, and outputs, then failing to document or validate it before reporting results.
Candidates prioritize getting a number quickly and treat the spreadsheet as a scratchpad; they underestimate how much of the exam's modeling credit comes from structure and validation.
What to do instead
Adopt a three-sheet structure (inputs, calculations, outputs), name ranges, and include a validation section with at least one independent check (e.g., replicate a simple case analytically). Practice rebuilding a small model under this structure.
Presenting a numerical result to a non-technical audience without stating the key assumption, the uncertainty, and the decision implication, or burying the conclusion under methodology.
Candidates are trained to show working and believe more detail signals competence, but exam communication questions reward clarity and decision-relevance.
What to do instead
Use a conclusion-first structure: state the answer, the one or two assumptions that drive it, the range of uncertainty, and the recommended action. Practice writing a 100-word executive summary for each practice calculation.