KASNEB CPA Advanced common mistakes
KASNEB CPA Advanced · Understand the traps and how to avoid them.
Computing profit, EPS or ratios straight off the trial balance and only then reading the notes about adjustments (revaluation depreciation, deferred tax on the revaluation surplus, finance-lease interest split, prior-period error). The arithmetic is internally consistent but the numbers are wrong.
Under time pressure candidates treat the trial balance as the finished statement and go to the notes only when they run out of rows to fill. The layout of the question paper encourages this: the trial balance is presented first and the notes sit underneath.
What to do instead
Force a two-pass discipline. Pass one: copy every note into a numbered working schedule W1 to Wn with its journal (Dr/Cr) and the statement line it hits. Pass two: only then build the face of the statement. Never write a figure on the face of a statement that is not traceable to a working.
Producing a 'consolidation' by adding the parent and subsidiary line by line, so pre-acquisition equity, intragroup balances, intragroup dividends and unrealised profit in inventory remain in the group figures, and goodwill/NCI are absent or unadjusted for fair-value step-ups.
Add-across is fast and looks like progress. The working papers (group structure, goodwill, NCI, group retained earnings) are seen as optional decoration rather than the computation itself.
What to do instead
Always draw the W1 group-structure diagram first (dates of acquisition, holdings, NCI method), then run the standard proformas W2 goodwill, W3 NCI, W4 group retained earnings, W5 intragroup eliminations. The consolidated statement is then just the sum of the proformas. Mark-scheme marks sit in the workings, not the face.
Answering 'identify the audit risks' with generic assertion language ('there is a risk of misstatement of receivables') without linking it to the specific entity, the financial statement line and amount affected, the assertion at risk, and the audit response.
Risk questions come early in the paper and candidates reproduce memorised ISA 315 assertion lists to bank quick marks, then run out of time for the scenario-specific application that carries the higher marks.
What to do instead
Answer in a fixed four-column risk log: Risk (scenario fact, e.g. 'receivables days rose from 45 to 90 while credit limits were unchanged') / Effect (overstatement of receivables and profit, valuation assertion) / Assertion / Response (specific substantive procedure, e.g. post-year-end cash receipts testing and aged-debt review). If a fact from the scenario is not in the row, delete the row.
Naming ISAs and quoting their requirements instead of describing the procedure to be performed, so answers read 'under ISA 540 the auditor shall obtain sufficient appropriate audit evidence' with no procedure, sample, or evidence source named.
Candidates equate citing the standard with demonstrating competence, and citing is easier than designing a test. It also feels safe because it cannot be 'wrong'.
What to do instead
Ban the phrase 'the auditor shall' from answers. Write every procedure as verb + population + technique + evidence: 'Vouch a sample of 25 additions selected from the fixed asset register to supplier invoices and delivery notes to confirm existence and cost.' Cite an ISA only where the question explicitly asks which standard applies.
Applying tax rates, bands, capital-allowance rates and thresholds from memory without pinning the year of income, so the computation mixes rates from two different Finance Acts within a single answer.
Rate schedules are revised annually and candidates revise from notes written for an earlier sitting. Under pressure the most recently memorised figure is used regardless of the year in the scenario.
What to do instead
Make the first line of every tax answer 'Year of income: 20XX; rates per Finance Act 20XX'. Check the rate schedule and thresholds against the stated year before any computation, and write the rate used beside each line of the workings so a mixed-rate answer becomes visible to you before it becomes visible to the marker.
Computing tax on a transaction without first classifying its legal form, e.g. taxing a benefit as ordinary income rather than applying the benefit-in-kind rules, ignoring VAT on imported services, or answering a tax-planning question purely as a computation with no deferred tax or timing consequence.
Classification feels like a detour when the numbers are visible. Candidates jump to the arithmetic and the classification is then reverse-engineered from the arithmetic rather than the other way round.
What to do instead
Before computing, run a four-question decision tree: (1) income or capital? (2) resident or non-resident? (3) VATable, exempt or zero-rated? (4) does the timing create a deferred tax asset or liability? Write the four answers down; the treatment follows mechanically and the planning points write themselves.
Delivering variance calculations or a limiting-factor ranking with no interpretation or recommendation, leaving the 'comment on the results' or 'advise management' marks unearned.
The calculation is where candidates feel confident and it consumes the time budgeted. The discussion part feels like padding and is squeezed into whatever time remains.
What to do instead
Budget the answer 50/50 from the start and structure it as Compute / Interpret / Act. In the Interpret block discuss only the two or three variances or constraints that are material and explain the likely operational cause (price versus usage, mix versus volume). In the Act block give a named, actionable recommendation addressed to a named manager.
Using absorption cost per unit, including apportioned fixed overhead in a one-off-order or make-or-buy decision, deducting sunk costs such as already-incurred research, and omitting the opportunity cost of scarce capacity.
The absorption cost is the figure printed in the question, so it is the one that gets used. Relevant costing requires rebuilding the cost stack, which candidates avoid when time is short.
What to do instead
Restate every cost into three buckets before deciding: incremental (include), avoidable (include), opportunity (include at contribution forgone). If a cost appears in the question but survives none of the three tests, it is excluded and you say so in one line: 'fixed overhead apportioned, not incremental, therefore excluded'.
Calculating WACC or a project discount rate using book values, the coupon rate as the cost of debt with no tax adjustment, or the current dividend instead of next year's dividend in the dividend growth model.
Book values and coupon rates are the figures given in the balance sheet extract, and the model formulas are recalled as symbols (D0, D1, g) without the timing convention attached.
What to do instead
Write the model and its inputs on one line before calculating: 'Kd = i(1-t) using market value of debt; Ke = D1/P0 + g where D1 = D0(1+g)'. State each input's source (market price, cum-div/ex-div, current versus next dividend) next to its value. If a figure has no stated source, stop and find it.
Ethics answers that name a threat category and then jump straight to 'report it', with no assessment of the threat's significance, no safeguard identified, and no named recipient of the report.
Threat labels are the memorable part of the framework, and 'report it' feels like a decisive conclusion. Candidates do not practise the middle steps because they seem procedural rather than substantive.
What to do instead
Use a fixed four-step template for every ethics scenario: Identify the threat and the parties / Assess significance (quantify the fee dependence, the relationship, the interest) / Identify safeguards and who applies them / State the course of action with a named recipient and a timeframe. If the answer has no amount and no named person, it is incomplete.