IAI SA1 common mistakes
IAI SA1 · Understand the traps and how to avoid them.
Answering the scenario question as if the examiner asked for a textbook definition of the technique, writing a neutral description of the model or method and never committing to a recommendation for the named client or scheme.
Candidates trained on earlier, more knowledge-recall papers default to describing what a technique is rather than applying it; the SA1 style question carries most of its marks for judgement, justification, and consequences, so a correct definition scores little.
What to do instead
Before writing, underline the decision the question asks for and the entity it concerns. Structure every answer as recommendation, then reasoning tied to the specific facts given, then risks of the alternative. Never write a paragraph that would still be true if the client's name and numbers were removed.
Ignoring the mark allocation and writing four pages on part (a) while leaving part (c) with two lines, so a high-value evaluation section is answered in a single sentence.
Under time pressure, candidates write what they know first and run out of time on the parts that require synthesis; the early descriptive part feels safer and expands to fill the available time.
What to do instead
In the first two minutes, convert marks into minutes (roughly one mark per minute of writing) and write the allocation beside each sub-part. Answer the highest-mark sub-part first if it requires the most thought, and leave a visible gap rather than overrunning a low-mark part.
Producing a list of bullet points with no connectives, so the reasoning linking cause to effect to recommendation is absent and the examiner cannot award application or evaluation marks.
Bullet lists feel efficient and are easy to generate under time pressure, but they present conclusions without the chain of argument that advanced marks depend on.
What to do instead
Use short paragraphs with explicit connectives ('because', 'which means that', 'however', 'therefore the scheme should'). If you use bullets, make each bullet a full causal sentence rather than a noun phrase.
Treating the question as a pure calculation and presenting a single number with no statement of assumptions, no comment on the sensitivity of the result, and no interpretation of what the number means for the decision.
Quantitative candidates are comfortable with the arithmetic and stop when the figure is produced; the professional-skills marks for interpreting and challenging the output are treated as optional.
What to do instead
After every calculation, add three sentences: the key assumptions made, how the answer changes if the main assumption moves (for example a change in the discount rate or mortality improvement), and what a decision-maker should conclude from it. State units and basis explicitly.
Failing to state the professional or regulatory context: recommending an approach without naming the applicable actuarial standard, the disclosure or capital requirement it interacts with, or the governance route through which the recommendation must be approved.
Technical training focuses on methods; candidates underweight the governance and conduct dimension and assume the examiner only wants the technique.
What to do instead
Keep a one-page checklist of the regulatory and professional framework relevant to your practice area and, for each answer, ask whether the recommendation triggers a disclosure, a sign-off, a peer review, or a conflict-of-interest consideration. Add a short 'governance and communication' paragraph to any recommendation question.
Writing an answer that is technically sound but unreadable: no headings, no underlined key terms, long unbroken paragraphs, and the recommendation buried in the middle of the final paragraph.
Candidates write in the order they think rather than the order the examiner reads, and underestimate how quickly a marker must locate awardable points across many scripts.
What to do instead
Adopt a fixed layout for every answer: a one-line recommendation in bold or underlined at the top, headed sections matching the sub-parts, and a short conclusion. Use the question's own terminology as headings so the marker can map your answer to the marking schedule.
Misreading the command word, for example writing an evaluation when asked to 'describe', or describing when asked to 'critique', or giving a balanced discussion when the question asks for a single recommended course of action.
Command words look similar and candidates skim them; the pressure to start writing overrides careful reading of the verb.
What to do instead
Circle the command word and the scope words (the named entity, the time horizon, the constraint) before planning. Check the verb against a short list: describe means state features; explain means give reasons; evaluate means weigh and conclude; recommend means choose and justify. Re-read the question after drafting to confirm you answered the verb asked.
Giving generic risk lists (market, credit, operational, liquidity) with no tailoring to the specific product, scheme, or transaction described, so the same answer could be reused for any question.
Memorised risk taxonomies are easy to reproduce and give a false sense of completeness, while tailoring requires the harder work of identifying which risks actually bite in the given facts.
What to do instead
For each risk named, add the specific transmission mechanism in this case and the party exposed: what triggers the loss, over what horizon, and who bears it. Delete any risk you cannot connect to the facts in the question.
Presenting a calculation or model output as definitive, with no acknowledgement of model risk, data limitations, or the possibility that the result is wrong, and no recommendation for how to validate it.
Candidates equate precision with quality and fear that caveats will look like weakness; in practice the advanced paper rewards recognising uncertainty and proposing controls.
What to do instead
End quantitative answers with a short 'limitations and validation' section: data quality, parameter uncertainty, model choice, and the checks or expert judgement that should accompany the result before it is relied upon.
Running out of time because the candidate writes full prose plans, drafts and redrafts the opening answer, and never practises under timed conditions, so the final sub-part is answered in note form or omitted entirely.
Untimed practice builds habits of perfectionism; candidates have never measured how long a 25-mark answer actually takes them, so their time plan is fictional.
What to do instead
Sit at least three full past papers under strict timing, and record the time taken per sub-part against the mark allocation. Practise a two-minute plan using bullet keywords only, and rehearse finishing with a conclusion even if the middle is compressed.