Associate PRM common mistakes
Associate PRM · Understand the traps and how to avoid them.
Treating the four risk categories (market, credit, operational, enterprise/business) as interchangeable labels and assigning a scenario to the wrong category, for example classifying a rogue-trader loss as market risk because prices moved.
Candidates learn category definitions in isolation and then read scenario stems quickly, anchoring on the first risk word they see (a price, a default) rather than on the root cause of the loss or exposure.
What to do instead
For every scenario, identify the source of the loss first and then map it: loss from price/rate moves is market risk; loss from a counterparty failing to pay is credit risk; loss from failed processes, people, systems, or external events is operational risk. Practise a two-column drill of 'event described' versus 'root cause' until the mapping is automatic.
Selecting a control activity when the question asks for a governance element, or vice versa, such as answering 'hedging with derivatives' when the stem asks which board-level mechanism provides oversight.
Governance and control vocabulary overlaps heavily (policy, limit, committee, hedge), and candidates answer the topic they associate with the keyword instead of the specific function the question demands.
What to do instead
Build a three-tier map: governance (board and committee structure, risk appetite, policy setting, independent oversight), control (limits, approvals, segregation of duties, reconciliation), and mitigation (hedges, insurance, collateral). Before answering, restate the stem in one sentence naming which tier it asks about.
Confusing Value at Risk with expected shortfall or maximum loss, and asserting that a 99% one-day VaR of a stated amount means losses cannot exceed that amount.
VaR is taught with a memorable single number, and the confidence-level semantics are easy to blur under time pressure, especially when distractors restate the number as a worst-case bound.
What to do instead
Memorise the precise statement: VaR is a loss threshold not expected to be exceeded with a given probability over a given horizon, and it says nothing about the size of losses beyond it. Expected shortfall is the average loss conditional on exceeding VaR. Practise writing the correct interpretation from memory and rejecting any option containing 'maximum' or 'cannot exceed'.
Applying the wrong credit metric in a question, for example using probability of default when the stem asks about loss given default, or computing expected loss without multiplying exposure at default.
PD, LGD, EAD, and expected loss are introduced together, and candidates pattern-match on the acronym in the stem rather than on the quantity being requested.
What to do instead
Write the identity expected loss equals PD times LGD times EAD on scratch paper at the start of any credit calculation and label each input before substituting numbers. Check units: PD as a decimal, LGD as a percentage of exposure, EAD in currency.
Choosing a risk response (avoid, reduce, transfer, accept) that does not match the residual exposure described, such as recommending transfer through insurance for a risk the stem says is uninsurable or already retained deliberately.
Candidates default to 'transfer' or 'mitigate' as the safe-sounding answer without checking feasibility constraints stated in the stem, such as cost, availability, or the organisation's stated risk appetite.
What to do instead
Read the stem for explicit constraints before selecting a response, and test each option against them. Distinguish transfer (risk and often a premium pass to a third party) from reduction (frequency or severity lowered in-house) and from acceptance (retention, which may be deliberate and justified).
Misreading the direction of a hedging or duration question, for example stating that a rise in interest rates benefits a holder of a long bond position, or that buying a put protects a short position.
Directional relationships are memorised as phrases rather than derived, so under exam pressure the sign flips, particularly when the question is framed from the hedger's rather than the investor's perspective.
What to do instead
Derive direction each time from first principles: bond price moves inversely to yield; a long position loses when prices fall and is protected by buying a put or selling a future. Sketch a one-line payoff arrow for the position before choosing an answer.
Treating operational risk as synonymous with compliance or legal risk, and excluding events such as system outages, fraud, or process errors from the operational category.
In practice operational risk sits beside compliance, so candidates import an organisational boundary rather than the exam's broader definition covering people, processes, systems, and external events.
What to do instead
Use the broad definition and test each event against the four sources. Legal and compliance failures are subsets of operational risk, not separate categories, unless the question explicitly distinguishes them. Review a list of event types (fraud, cyber, model error, outsourcing failure) and classify each.
Answering ethics questions by choosing the option that protects the employer or client relationship rather than the option that satisfies the professional standard, such as staying silent about a material error to avoid reputational damage.
Candidates reason from workplace norms and loyalty, and the ethically correct answer often appears to conflict with short-term commercial interest, creating doubt under time pressure.
What to do instead
Apply a fixed hierarchy: integrity and client interest first, then escalation through internal channels, then external disclosure where required. Reject any option whose justification is reputational or commercial. Practise identifying the disclosure or escalation duty embedded in each scenario.
Confusing model risk with market risk because both involve quantitative models, and failing to flag model error as an operational risk event when a valuation model is misapplied.
The word 'model' triggers association with pricing and valuation, so candidates classify the consequence (a misstated value) rather than the cause (an inadequate or misused model).
What to do instead
Classify by cause: a flawed or misused model is an operational risk and a model risk issue, regardless of whether the resulting number is a price. Note that model risk can amplify market or credit exposures but is not itself a market or credit risk.
Ignoring the reporting audience and purpose in risk reporting questions, for example recommending the same detailed exposure data for the board as for a trading desk, or omitting escalation thresholds.
Candidates focus on what data exists rather than on aggregation level, frequency, and decision use, which are the actual discriminators in reporting questions.
What to do instead
For each reporting question, state the audience, the decision they must make, and the appropriate aggregation and frequency. Board reporting should be aggregated, forward-looking, and tied to risk appetite and limits; desk reporting should be granular and timely. Include escalation triggers and limit-breach protocols where the stem involves breaches.