How to use these practice questions
The Actuaries Institute pathway combines university-accredited subjects with Institute-set assessments. Later stages commonly include written, problem-solving work and applied communication tasks. The exact assessment depends on your subject and stage, so check the Actuaries Institute’s official website and your subject materials for current requirements.
The questions below are original practice examples, not official or past-paper questions. They cover skills that matter across actuarial problems: choosing a method, stating assumptions, checking units and signs, and explaining what a result means. For each question, choose an option before reading its explanation. After answering, note whether you got the method right as well as the number.
A useful written-answer sequence is: identify what is being asked; state assumptions; show the method and workings; interpret the result; then note a relevant sensitivity or limitation. The examples are short, but apply that structure whenever the problem calls for explanation or advice.
MySummaries can build practice questions from a candidate’s own study material. This browser question screen shows the kind of calculation and explanation to practise:
A fund is expected to pay $10,000 exactly two years from now. What is its present value at an annual effective discount rate of 5%, assuming annual compounding?
Discount by (1.05)^2: 10,000 / 1.1025 = $9,070.29. $9,500 discounts for only one year; $11,025 compounds in the wrong direction. State the timing and effective annual rate assumption.
Calculation and modelling questions
These questions test whether you can match a formula to the quantity requested. Keep units visible in your workings. A numerical answer without a method can conceal a timing, sign or modelling error.
This first question checks the expectation of a compound insurance loss model.
Annual claim count N has expected value 120. Individual claim amounts have expected value $800 and are independent of the count. What is the expected annual aggregate claim amount S = sum of claim amounts?
For a frequency-severity model with count independent of severities, E[S] = E[N]E[X] = 120 × $800 = $96,000. The variance is needed for the variance of aggregate claims, not their expectation. $120,800 incorrectly adds the two inputs, which have different meanings and units.
Be precise about what information a calculation needs. In this example, independence matters for the usual frequency-severity model, while the count variance does not enter the expected value.
This question tests the zero-event probability under a Poisson frequency assumption.
The number of claims in a year is modelled as Poisson with parameter λ = 2.5. What is the probability of no claims in a year?
For a Poisson variable, P(N = 0) = e^(-λ) = e^(-2.5) ≈ 0.0821. 0.9179 is the probability of at least one claim, 1 − P(N = 0), so it is the strongest distractor but answers the complementary event.
Name the assumed distribution when explaining the result. If the data do not support a Poisson model, the calculation may be exact but the model choice still needs justification.
This question checks a weighted portfolio return and whether the weights are applied consistently.
A portfolio holds 60% in an asset returning 8% and 40% in an asset returning −2% over the same period. Ignoring rebalancing within the period, what is the portfolio return?
The weighted return is 0.60 × 8% + 0.40 × (−2%) = 4.0%. A simple average gives 3.0%, but wrongly gives both assets equal weight. The result assumes the stated weights apply over the period.
Show the negative return with its sign. A quick check is that the portfolio return should lie between −2% and 8%; that rules out several possible arithmetic errors.
This question tests the first-order duration approximation and the direction of a bond price change.
A bond has modified duration 4.2. Using the duration-only approximation, what is the approximate percentage price change if its yield rises by 0.50 percentage points?
ΔP/P ≈ −D_mod × Δy = −4.2 × 0.005 = −0.021, or approximately −2.1%. The yield change is 0.50 percentage points = 0.005, not 0.50. The positive option has the wrong sign: bond prices generally fall when yields rise. Convexity is omitted in this first-order approximation.
A stated approximation is part of the answer. For a larger yield move, explain that duration alone omits convexity rather than presenting the estimate as an exact repricing.
This question tests compounding claim inflation over a specified number of years.
A claim cost of $1,000,000 is projected forward three years using annual claim inflation of 6%. What is the projected cost?
Compound the annual rate: $1,000,000 × 1.06^3 = $1,191,016 (rounded to the nearest dollar). $1,180,000 adds three times 6% and misses compounding. The calculation assumes the 6% rate applies each year and is compounded annually.
From result to decision
Actuarial questions often ask for more than a calculation. The next examples require you to connect the number to assumptions, uncertainty or a decision.
This question tests how correlation affects the variance of a combined result.
Two risks have standard deviations 10 and 12, with correlation 0.5. What is the variance of their sum?
Var(X + Y) = 10^2 + 12^2 + 2(0.5)(10)(12) = 100 + 144 + 120 = 364. 244 omits the covariance term and implicitly treats the risks as uncorrelated. The inputs are standard deviations, so they must be squared for the variance terms.
When presenting a combined-risk estimate, state the correlation assumption and explain its effect. Here, positive correlation increases the variance compared with independence.
This question tests solving a premium equation when expenses and target profit are proportions of premium.
Expected claims are $600 per policy. Fixed expenses are $40 per policy. Variable expenses are 8% of premium P and the target profit is 5% of P. What premium satisfies expected claims + fixed expenses + variable expenses + target profit = P?
Set 600 + 40 + 0.08P + 0.05P = P. Thus 640 = 0.87P and P = $735.63 (rounded). $672 adds 13% to $600 but ignores fixed expenses and applies the percentages to the wrong base. The equation assumes all amounts are per policy and the stated loadings are proportions of premium.
The equation makes the premium base explicit. In a longer response, identify whether the target is a margin on premium or a loading on cost; these are not interchangeable.
This question tests a credibility-weighted estimate and interpretation of the credibility factor.
A credibility estimate is Z × observed rate + (1 − Z) × prior rate. If Z = 0.4, the observed rate is 560 and the prior rate is 500, what is the estimate?
The estimate is 0.4 × 560 + 0.6 × 500 = 224 + 300 = 524. 536 reverses the weights, giving the prior estimate more weight than specified. Explain that Z = 0.4 assigns 40% weight to the observed rate under this stated formula.
A useful answer distinguishes calculation from interpretation: the result is closer to the prior because the prior receives 60% weight. Do not claim that the weighting is statistically justified unless the question gives the basis for it.
This final question tests the difference between a model output and a conclusion that is adequately supported by evidence.
A projection shows a surplus under the central assumption set, but the result becomes a deficit when claim inflation is 2 percentage points higher. What is the most decision-useful conclusion?
A decision-useful response reports the central result and the material sensitivity, then explains what the change means and what uncertainty remains. The central projection alone does not prove affordability, while automatically replacing the assumption treats a sensitivity as a forecast. State the assumptions and any further analysis needed before recommending action.
Practise writing the explanation
After a set of calculations, practise turning a result into a concise written response. The example below is a practice marking screen, not an official Actuaries Institute marking scheme. Its score reflects a candidate who has the main method and conclusion but misses a limitation.
A projection gives a present value of $9.07 million at a 5% annual effective discount rate. Explain the result and identify one relevant sensitivity or limitation.
7/9Discounting the $10 million payment two years at 5% gives $10m / 1.05² = $9.07m. This is the amount today that is equivalent to the future payment at the assumed rate. A higher discount rate would reduce the present value.
The method, timing, units and direction of the rate sensitivity are clear. The response could distinguish the discount-rate assumption from other projection assumptions and say how the sensitivity would be quantified.
Missed
−Quantify the effect of a specified alternative rate or identify another material limitation, such as payment timing uncertainty.
Model answerAssumptions & method — Assume one $10 million payment occurs exactly two years from now and the annual effective discount rate is 5%. Present value is 10 / 1.05² = $9.07 million. Interpretation — This is the current value at the stated rate, not a guaranteed market value. Sensitivity & limitation — A higher discount rate lowers the present value; quantify the effect using the relevant alternative rate and test payment-timing or cash-flow uncertainty if material.
That is a MySummaries paper, filled with Actuaries Institute material. Yours is written from your own notes. Start free
Review your own answers in the same order: did you answer every part, state the assumptions, show a checkable method, and explain the result in context? For a real assessment, follow the instructions and conventions specified for that subject rather than assuming every paper uses the same response style.
The weakest area in a short practice set can point to what to revise next; the counts below refer only to these practice questions, not to official exam frequency.
If you lose the same mark twice, make it a small, retrievable question rather than rereading a whole topic. For example, practise identifying the missing decision-relevant sensitivity.
You lost this mark twice: when a central projection changes materially under a plausible assumption, what should your written conclusion report besides the central result?
How MySummaries helps
Build a revision board from your own actuarial notes, then generate question practice from that material. Use the marked response and repeated-error prompts to identify what to revisit, and return to the source notes to check any formula or assumption. Open MySummaries.