IFoA SP10 common mistakes
IFoA SP10 · Understand the traps and how to avoid them.
When asked to recommend a strategy for a life insurer or pension scheme, the candidate jumps straight to an asset allocation (e.g. '60% bonds, 30% equities, 10% property') without first stating the scheme's objectives, liabilities, time horizon, liquidity needs, regulatory capital and risk appetite, so the recommendation cannot be justified against the specific client.
Candidates memorise model portfolios and typical allocations and treat the question as a recall test of asset mixes, rather than recognising that SP10 marks are awarded for linking the recommendation to the stated liability profile and constraints. Under time pressure, listing percentages feels efficient and 'complete'.
What to do instead
Before writing any allocation, spend two or three sentences explicitly setting out: the nature of the liabilities (real vs nominal, duration, uncertainty), the funding position and surplus, the time horizon, liquidity requirements, the regulatory/capital regime, tax position and the trustees'/board's risk appetite. Then justify each asset class by reference to those stated needs, and state the expected return/risk contribution. Practise by writing a one-page 'objectives and constraints' template and applying it to every past-style scenario before touching numbers.
The candidate describes a derivative (e.g. an interest rate swap, equity put, or total return swap) and its mechanics correctly but omits the explicit hedge objective, the hedge ratio/notional, the basis risk and counterparty/collateral implications, and never states what happens to the hedge under a stress scenario.
Derivative questions look like product-recall questions, so candidates spend their words on defining the instrument and run out of space before addressing why the hedge is being put on and how effective it will be. Many candidates are also weak on practicalities such as collateral, margin calls and basis risk because they have only studied payoff diagrams.
What to do instead
Structure every hedging answer as: (1) the risk being hedged and its measurement; (2) the instrument and why it is preferable to alternatives; (3) notional/hedge ratio and how it is set; (4) residual risks (basis, counterparty, liquidity, roll, collateral/margin); (5) a stress case showing the net outcome. Work through a numerical example with a swap and a put option, computing the hedged outcome in two or three scenarios, so the mechanics are anchored to a purpose.
In equity valuation questions the candidate recites the dividend discount model formula and then applies it with a single growth rate, failing to discuss the sensitivity of the value to the assumed growth rate, the required return, the payout/retention assumption and the appropriateness of the model for the company described (e.g. a non-dividend-paying growth company or a cyclical).
Candidates treat valuation as formula application and are anxious to get a number down. They underestimate how heavily SP10 rewards commentary on assumptions, sensitivity and model choice relative to the arithmetic itself.
What to do instead
For every valuation, state the model chosen and why it suits the company, list each input with its source/justification, produce the value, then run at least a two-way sensitivity (e.g. growth ±1%, discount rate ±1%) and comment on the range. Explicitly discuss when the model breaks down (negative earnings, no dividend, terminal value dominance) and what alternative (multiples, free cash flow, sum-of-parts) you would use. Practise writing the commentary before the calculation so it is not squeezed out.
The candidate answers a bond question using only yield to maturity and duration, ignoring credit spread, convexity, the shape of the yield curve, liquidity and the possibility of default, and so mis-states the risk of a credit or long-dated bond portfolio.
Duration is the single most rehearsed bond concept, so it becomes the default answer to every bond risk question. Candidates conflate modified duration with a complete risk measure and forget that spread duration and convexity matter materially for corporate and long-dated paper.
What to do instead
Build a checklist for every bond question: yield curve level and shape, credit spread and spread duration, modified duration and convexity, default/recovery, liquidity, inflation linkage and currency. When calculating a price change, use both duration and convexity and show the difference. Work a numerical example on a long-dated and a high-yield bond to see how much duration alone misleads.
When asked to construct a portfolio or assess diversification, the candidate lists asset classes and correlation coefficients but does not address how the allocation is actually implemented: rebalancing policy, transaction costs, liquidity, currency exposure, benchmark choice, constraints and the difference between strategic and tactical allocation.
Portfolio construction is taught as mean-variance optimisation and candidates reproduce the efficient frontier narrative, but exam scenarios are about practical implementation. Candidates also under-weight costs and liquidity because they are not in the textbook formulas.
What to do instead
For any allocation question, separate strategic (long-term, driven by liabilities and risk appetite) from tactical (short-term, driven by valuation/views), then cover implementation: rebalancing rules and bands, cost and tax drag, liquidity and dealing, currency hedging policy, benchmark and tracking error, and constraints (regulatory, mandate, ESG). Quantify the effect of costs and rebalancing on expected return where the question gives data.
In performance measurement questions the candidate computes a total return or a simple difference from benchmark and calls it attribution, without decomposing the excess return into allocation, selection, currency and interaction effects, and without checking whether the return is time-weighted or money-weighted and gross or net of fees.
Attribution requires a disciplined multi-step method and candidates default to the easier arithmetic. They also overlook that the choice of return basis (TWR vs MWR, gross vs net) changes the conclusion, particularly for schemes with large cash flows.
What to do instead
Always state the return basis first (TWR for manager skill, MWR for investor experience; gross vs net) and then run a full Brinson-style attribution: allocation, selection, interaction, plus currency and any residual. Reconcile the components to the total excess return so nothing is unexplained. Practise with a two-sector, two-manager example and show the reconciliation explicitly.
The candidate discusses investment governance generically ('good governance is important', 'the board should oversee risk') without naming the specific bodies, documents, controls or escalation routes: statement of investment principles, investment strategy statement, mandate and guidelines, investment committee, delegated authority, custodian, compliance monitoring, conflicts of interest, and the role of the actuary and auditors.
Governance feels like common sense, so candidates write plausible-sounding prose that contains no testable content. They have often not memorised the UK-specific framework (trustee duties, SIP, scheme funding, TPR guidance) and so avoid specifics.
What to do instead
Learn a concrete governance map: who sets strategy (trustees/board), who advises (actuary, investment consultant), who executes (managers), who safeguards assets (custodian), who monitors (compliance, internal audit), and which documents record it (SIP/ISS, mandate, guidelines, risk register, conflicts policy). In answers, name the document or body and say what it does and how breaches escalate. Build this as a one-page diagram and reproduce it in practice answers.
The candidate answers an asset-liability modelling question by describing the model's outputs (projected surplus, probability of funding shortfall) without specifying the model's inputs, assumptions and limitations: asset return distributions, liability model, correlation assumptions, time step, number of simulations, treatment of inflation and longevity, and the danger of over-fitting to historical data.
Candidates focus on the impressive output and treat the model as a black box. They also lack confidence critiquing assumptions, so they skip the limitations paragraph which carries significant marks.
What to do instead
Answer ALM questions in four blocks: purpose and scope of the model; inputs and key assumptions (with justification and sensitivity); outputs and how they are interpreted; limitations and how they are mitigated (model risk, parameter uncertainty, tail events, data quality, governance of the model). Always include at least one sensitivity run, e.g. a 1% fall in asset returns or a 10% increase in longevity, and state the effect on the funding level.
The candidate writes a long, unstructured answer to a 'discuss' or 'recommend' question, mixing points from several domains, so that the examiner cannot see which requirement each paragraph addresses and marks are lost even where the content is correct.
Under exam time pressure candidates write continuously rather than planning, and SP10 questions often combine several domains (e.g. a pension scheme's bond allocation, hedging and governance). Without signposting, valid points are buried and the answer reads as a stream of consciousness.
What to do instead
Spend two minutes planning: read the verb (state, explain, discuss, recommend, calculate), split the question into its parts, and allocate a paragraph to each with a heading. Use the structure 'point - explanation - application to the scenario - consequence'. Leave a line between sections and make the first sentence of each paragraph state the point. Practise timed answers and mark your own against the command words.
The candidate treats property and alternative assets (private equity, infrastructure, hedge funds, commodities) as if they behave like listed equities: quoting a single expected return and volatility, ignoring valuation frequency and smoothing, illiquidity, leverage, manager selection risk, high fees and the difficulty of measuring risk with appraisal-based indices.
Standard deviation and correlation from published indices are the familiar tools, and candidates do not appreciate that smoothed appraisal valuations understate volatility and overstate diversification. They also under-weight fees and illiquidity premia, which are central to these asset classes.
What to do instead
For each alternative asset, state: return sources and drivers, risk measures and why reported volatility is understated (valuation smoothing, stale pricing), liquidity and lock-up terms, leverage, fees and alignment, manager/selection risk, and how it fits the liability profile. Where relevant, discuss de-smoothing or using cash-flow-based risk measures. Compare the net-of-fee return to the required return rather than quoting gross figures.