Specimen paper

It is 18:00 on working day 4 at Trentbank Care Group and the manager must decide whether the close moves on to trial balance review. Which statement about the bank reconciliation lets her make that decision without asking anybody?

• A

  Balance agrees to the statement, all items described and aged, none over 30 days unexplained.

  Correct: It states a condition on the output itself, covering the composition of the balance as well as its total, and you can test it from the file alone at 18:00 with nobody left to ask.

• B

  Balance agrees to the statement and the schedule has been signed and dated by its preparer.

  This is the most familiar gate wording in practice, and it attracts you because it names both a check and an approver. It tests only the total, so an unmatched item can sit inside a balancing reconciliation for months and never fail the gate. It would be sufficient only if agreement to the statement were the single assertion at risk, which stops being true the moment you accept that a reconciliation can balance while holding items nobody has explained.

• C

  Reconciliation is complete, filed in the close folder, with the bank statement attached to it.

  This substitutes evidence that the activity happened for a condition the output must meet, and it is where you land if you treat 'done' as meaning 'documented'. Filing tells the manager where the schedule is, not whether the number in it can be carried forward. It would be the right answer to a different question, namely how the performance of the task should be evidenced in the close file.

• D

  Reconciliation has been prepared by the cashier and reviewed by the financial controller.

  This names the people instead of the standard, and preparer plus reviewer sounds like assurance because in most functions it is how quality is described. Who performed the work says nothing about what the work found, so the manager still has to open the file to learn whether anything is unexplained. It would work as a gate only if naming individuals guaranteed a defined quality of output, which no allocation of names can do.

The agreed adjustment threshold at Solterra Renewables is USD 25,000 for a single item. On working day 4 the manager finds uncorrected differences of USD 18,000, USD 14,000, USD 12,000 and USD 21,000, all overstating a monthly profit of USD 300,000. What should he do?

• A

  Correct them, since in aggregate they pass the threshold and all push the same way.

  Correct: The aggregate of USD 65,000 against profit of USD 300,000 is material, and the one-way direction of all four differences means they compound rather than offset.

• B

  Leave them, since no single item passes the threshold, and log them on the schedule.

  This applies the policy exactly as written, item by item, which is why it is the commonest wrong answer and why the failure it describes survives in real functions. It ignores that a single-item threshold is a working convenience sitting underneath an aggregate test, and four one-way differences totalling more than a fifth of monthly profit fail that test. Logging without correcting would be right while the aggregate stayed within tolerance and the items pointed in both directions, which is not the position here.

• C

  Correct the USD 21,000 alone, as the nearest to the threshold, and carry the rest.

  This comes from treating proximity to the threshold as the criterion, and it feels like proportionate compromise because some work is done. It removes USD 21,000 of a USD 65,000 one-way distortion and leaves USD 44,000 still overstating profit, with the same bias and a better-looking schedule. It would be right only if that item were the sole difference and the others did not exist to be aggregated with it.

• D

  Raise the threshold to USD 75,000 this month, as profit is ahead of plan, and sign off.

  This is rationalisation rather than materiality: the threshold is moved during the close, in the direction that removes the work, and justified by the result rather than by the needs of the people who read the accounts. Materiality is set by reference to users and to the size of the reporting entity, not by how comfortable the month looks. A revision would be legitimate before the period, argued from the user's decisions, and documented.

Orinoco Coffee Traders' accrual register shows the freight accrual within 3% of actual for seven months, and the packaging accrual out by 28% and then 34% in the last two months. Both are rebuilt from scratch each close. What should the manager decide?

• A

  Roll both, now that the register records how each estimate has actually performed

  This treats the existence of the register as the evidence rather than reading what the register says, which is the error of confusing having a measurement with having a good result. Rolling an estimate that has just missed by roughly a third writes that error straight into next month and every month after, since a rolled figure carries forward until something disturbs it. It would be right if packaging's history looked like freight's, and the register is precisely what tells you it does not.

• B

  Rebuild both, since rebuilding remains the more prudent treatment at month end

  This is prudence by habit: rebuilding feels safer because it starts from source data, and nobody is ever criticised for doing more work. It spends critical path hours on an estimate already demonstrated accurate to within 3% over seven months, and each rebuild reintroduces variability from re-keyed inputs and fresh judgements. It would be the right default where no measured history existed, which is exactly the condition the register has removed.

• C

  Roll packaging with an uplift factor, and leave freight on its current rebuild

  An uplift assumes the error has a stable sign and a stable size, and two observations of 28% and 34% establish neither; it applies a correction factor to an estimate whose source data is the actual problem. It also leaves the reliable estimate consuming the rebuild hours, so it gets both decisions the wrong way round. It would be defensible if packaging missed by a consistent, measured margin traced to a known and unchanging cause, which would make the uplift part of the method rather than a patch on it.

• D

  Roll freight with a documented true-up, and rebuild packaging from a better source

  Correct: Each estimate is treated on its own measured performance: the accurate one is rolled with a true-up, the inaccurate one is rebuilt from data capable of supporting it.

Nordhavn Marine Supplies ranks its risks by expected annual exposure. Supplier bank detail diversion sits near the bottom at GBP 4,000 a year, but a single successful diversion against a large settlement would remove GBP 80,000 at once. How should the accounting manager treat it?

• A

  Control it despite the low expected value, because the loss arrives in one piece.

  Correct: A low expected value that conceals a single large and effectively irrecoverable event is the definition of a tail risk, and tail risks are controlled on their shape rather than their average.

• B

  Rank it above payroll leakage, because the single-event figure is far larger.

  This overcorrects by replacing expected value with worst case as the ranking rule, and it is tempting because the GBP 80,000 is the most vivid number in the stem. Ranking the whole register on single-event severity would push continuous exposures that leak far more cash each year down the list and starve them of attention. It would be right only if the register's purpose were to rank severity, whereas its purpose is to direct finite control effort at where the money actually goes.

• C

  Move it to the misstatement list, because the figure rests on an assumption.

  This confuses two registers: cash loss risks are measured in money out of the bank, misstatement risks are measured against materiality in the accounts, and the fact that a frequency has been assumed does not convert one into the other. Someone lands here by treating uncertainty in the estimate as a reason to file the item elsewhere rather than as a reason to control it. It would be right if the exposure could misstate the accounts without removing cash, which a diverted payment plainly does not.

• D

  Defer it until its expected value exceeds that of the petty cash exposure.

  This ranks strictly on expected value, which is the very error the tail distinction exists to correct, and it has an unpleasant property: the expected value only rises after a diversion has succeeded. Deferral therefore guarantees that the control arrives after the loss it was meant to prevent. It would be reasonable if losses of this kind arrived in small recoverable pieces, so that early experience could be bought cheaply and used to justify the spend.

Marchetti Ceramics sends every purchase invoice above GBP 200 to the Finance Director, about 1,050 items a month, which she clears in two sittings a week. What does this most likely imply?

• A

  The approval produces a record rather than a control, and a defence that will not hold.

  Correct: At that volume the approver cannot form a view on any individual item, so the ledger records an approval that was never exercised.

• B

  The approval is sound but inefficient, and should be batched into one weekly sitting.

  This reads a control failure as a diary problem, which is the natural reading if you assume that a documented approval is by definition operating. Batching 1,050 items into a single sitting makes the seconds available per item worse rather than better, so the proposed remedy deepens the fault it is meant to cure. It would be the right answer if the approver's judgement were being properly applied and only the scheduling were awkward for her.

• C

  The approval is effective for value, since it covers most of the ledger spend by amount.

  The coverage claim is arithmetically true and beside the point: a GBP 200 threshold does capture most of the ledger by value, but risk is not distributed like value, and the payment that hurts is often a modest one to a supplier created three weeks ago. This option comes from treating population coverage as evidence of effectiveness. It would be right if effectiveness could be demonstrated by what was included in the population rather than by evidence that judgement was applied to it.

• D

  The approval is weak because no second approver is required on the largest invoices.

  This reaches for a stronger version of a control that is not operating at all, an instinct that comes from grading controls by how many signatures they carry. Adding a second approver puts another name on a population nobody reads and doubles the volume of unexercised approval. It would be a fair criticism if the first approval were genuinely being performed and the remaining concern were concentration of authority over large payments.

Larkhill Brewing's monthly close needs 386 hours of effort, and the window of working days one to five holds 318 hours once business as usual is deducted. The financial controller proposes that each of the twelve staff work six extra hours in close week, producing 72 hours. What is the strongest objection?

• A

  Overtime hours cost more each than a contractor engaged for the five close days.

  This argues price when the objection is structural, and it will appeal if you have been taught to compare the unit cost of every resourcing option. Buying the hours more cheaply from somebody else still leaves a close that needs 386 hours inside a 318-hour window every month for ever. It would be the right answer to a different question, namely how to source a shortfall that has already been accepted as genuine and temporary.

• B

  The plan assumes full attendance, so one absence in close week would undo the cover.

  This is the most tempting wrong answer because it is true: 72 hours against a 68-hour gap has almost no slack, and one absence sinks it. It nevertheless treats a permanent design deficit as a staffing contingency, so the strongest version of this objection produces a slightly more robust plug and leaves the 68 hours to be found again next month. It would be the leading objection if the deficit were a one-off, arising from a cutover or an unusual reporting demand.

• C

  Overtime hours are less productive, so 72 booked hours deliver fewer usable ones.

  This invents a discount the measured hours do not show, and it argues about the size of the plug rather than about whether a plug is the right instrument at all. Even granting a fatigue effect, correcting for it would only mean asking for more overtime. It would be a useful contribution if the question were how many overtime hours to budget for a genuinely temporary peak.

• D

  The gap is structural, so standing overtime conceals the fault in the close design.

  Correct: A recurring 68-hour deficit is a property of the close design, and monthly overtime funds it out of goodwill while removing the signal that would have prompted redesign.

Two experienced accountants at Larkspur Pharmaceuticals have missed the intercompany reconciliation deadline in consecutive months. Both meet every other deadline they own, both are inside their hours, and both have asked which rate should be used for the same class of balances. What should the manager conclude?

• A

  The task lacks a specification, so the fix belongs to the task.

  Correct: Two capable people failing at the same task, and asking the same question about it, identifies the task rather than the people as the common factor.

• B

  Both need coaching on the intercompany rate rules ahead of the next close.

  This reads a clarity problem as a capability problem, which is the default diagnosis when a deadline slips and the most common wrong answer here. Coaching two experienced accountants on a rule that has never been written down transmits your answer verbally and records nothing, so the next owner of the task asks the identical question. It would be right if one of them had struggled while the other worked confidently from a documented standard, which is the pattern that actually indicates a development need.

• C

  The reconciliation should move to one owner so the confusion ends.

  This buys silence rather than resolution: consolidating the work removes the visible disagreement while leaving the ambiguity untouched, and one person will simply choose a rate and apply it consistently, whether or not it is the right one. It also manufactures a single-person dependency on a task that two people can currently attempt. It would be a sensible move if the difficulty were coordination between two preparers rather than an unanswered technical question.

• D

  Both are over capacity, since the same deadline slips every month.

  Capacity is the usual explanation for a missed deadline, and this option offers it, but the stem closes it off deliberately: both are inside their hours and meet every other deadline they own. Choosing it means the evidence in the stem has been read past rather than used. It would be right if the hours data showed overload, in which case the remedy would be reallocation of work and not a written rule.

Steinhaus Machine Tools reports trade debtors of EUR 6,300,000, which include sales tax at 20%. Revenue for the last twelve months was EUR 36,000,000 excluding sales tax. Using the simple average daily sales method on a 365-day year, what is the debtor days figure?

• A

  44.4 days

  This adjusts twice: it strips the tax out of debtors to give EUR 5,250,000 and also grosses revenue up to EUR 43,200,000, so the same mismatch is corrected in both directions and the ratio is understated by a wide margin. It usually arises when you spot the tax problem, apply one fix, and then apply the other for safety. It would be correct only if debtors were reported net of sales tax while revenue were reported gross, which is the reverse of how ledgers are kept.

• B

  52.5 days

  The gross-up is done correctly and the year is taken as 360 days, giving 6,300 divided by 43,200 multiplied by 360, so the only error is the convention. It would be right in a house that reports on a 360-day basis, but the question specifies 365. Note how little separates the two figures and how much that matters in practice: 0.7 days of movement caused entirely by convention is exactly the kind of change that must be disclosed rather than presented as collection performance.

• C

  53.2 days

  Correct: Revenue is grossed up to EUR 43,200,000 so that it matches tax-inclusive debtors, and 6,300 divided by 43,200 multiplied by 365 gives 53.2 days.

• D

  63.9 days

  This divides tax-inclusive debtors by tax-exclusive revenue with no adjustment at all, overstating the answer by roughly the tax rate, and it is the figure you get by reading both numbers off the face of the accounts and dividing. It is the single commonest error in this ratio and it is persistent, because the result looks plausible and nothing in the calculation fails. It would be correct in a jurisdiction where sales carried no such tax, so that both quantities were already on the same basis.

Wrenfell Distribution's Hull site reports own payroll GBP 18,000 favourable, agency GBP 21,000 adverse and subcontracting GBP 46,000 adverse, with seven driver posts vacant. The draft commentary lists the payroll figure under favourable movements. What is the main fault?

• A

  It presents as a saving the vacancy that caused the substitution cost.

  Correct: The payroll underspend and the GBP 67,000 of agency and subcontract cost are one operational event, and separating them invites the reader to bank the favourable half as performance.

• B

  It shows the three lines gross when the pack convention is to report each movement net.

  This invents a reporting convention and then finds the draft in breach of it, which is what happens when you look for a presentational rule to cite rather than asking what the reader will conclude. Netting would not cure the fault and might deepen it, since the substitution would disappear into a single number and the vacancy would never be named. It would be a valid criticism only if the pack genuinely mandated net reporting and the three lines were unconnected, which is exactly what they are not.

• C

  It omits the percentage variance beside each line, so the reader cannot judge the scale.

  This is the restating-the-table habit: asking for another column instead of for an explanation, on the assumption that commentary exists to describe the numbers rather than their causes. A percentage beside the GBP 18,000 favourable would state a misleading item more precisely. It would be a fair point if the reader's difficulty were judging the size of an otherwise well-explained movement.

• D

  It gives the site position without the group figures, so materiality cannot be judged.

  This asks for aggregation and context, which sounds rigorous but does nothing about the defect: the site page is read by people who run the site, and adding group columns would not stop any of them treating the vacancy as a saving. It comes from reaching for a general reporting principle when a specific causal error is in front of you. It would be relevant if the question were whether this variance deserved commentary at board level at all.

Ellerdale Haulage budgeted 15% of 30,000 loads to subcontractors at GBP 200 a load; actual was 24% of 30,000 loads at GBP 212. Subcontracting is GBP 626,400 adverse and the draft commentary proposes renegotiating rates. Which decomposition shows why that remedy is misdirected?

• A

  Rate accounts for GBP 54,000; the shift of 2,700 loads to subcontractors is GBP 572,400 of it.

  This prices the rate effect on budgeted subcontracted loads, GBP 12 on 4,500, instead of on the 7,200 actually subcontracted, which is an inconsistent convention. It is dangerous precisely because it still foots to GBP 626,400: the residual silently absorbs the rate increase on the extra 2,700 loads and inflates the mix figure by GBP 32,400. It would be acceptable only in a scheme that priced every effect at budget volumes throughout, and even then the residual could not be described as the shift alone.

• B

  Rate accounts for the whole GBP 626,400, since every load ran above the planned price.

  This repeats the assumption already made in the draft commentary, and it is seductive because its premise is true: every subcontracted load did cost GBP 12 more than planned. It ignores that the number of subcontracted loads rose by 60%, so it attributes to a GBP 12 price movement money that a decision about who carries the work actually caused. It would be right if the subcontracted share had held at 15%, in which case rate would indeed be the only variable that moved.

• C

  Rate accounts for GBP 86,400; the shift of 2,700 loads to subcontractors is GBP 540,000.

  Correct: Rate is GBP 12 on the 7,200 loads actually subcontracted and mix is 2,700 additional loads at the budgeted GBP 200, and the two foot to the reported GBP 626,400.

• D

  Rate accounts for GBP 86,400; a volume effect of 2,700 extra loads is GBP 540,000.

  This is the near miss: the arithmetic is right and the label is wrong, which makes it the most instructive distractor on the paper. Calling the effect volume tells the reader the business carried more work, whereas total loads were 30,000 against a plan of 30,000, so what changed was the share pushed outside the fleet. It would be correct if actual loads had exceeded plan by 2,700 at the budgeted 15% share, and the distinction is not academic, since the remedy for volume is capacity and pricing while the remedy for mix is finding out why the fleet is not carrying the work.

Halden Plant Hire holds a separate nominal code for each of its four yards under every cost heading. A fifth yard opens in March. What does the existing chart design imply about that opening?

• A

  Roughly a full set of new codes, plus changes to every report and rule

  Correct: Location encoded in the account code means a fifth yard needs a code under every heading, and every report, recurring journal, bank rule and mapping table built on those codes must be amended.

• B

  One new dimension value, with existing reports picking the yard up

  This describes the design the business ought to have rather than the one the stem gives it, and it is chosen by candidates who answer from good practice instead of from the facts in front of them. Under a dimensional model it would be exactly right, and the contrast is the point of the question: the cost of the March opening is the price of a design decision taken years earlier. It would be the correct answer if yard were held as an analysis dimension rather than as a segment of the account code.

• C

  A new block of codes above the current range, with reports unchanged

  This reads the difficulty as running out of numbers, so it solves the numbering and then asserts that reporting is unaffected. Reports, allocations and posting rules generally select accounts explicitly, so newly created codes are invisible to them until each one is amended, and the first sign of the omission is usually a report that no longer agrees to the trial balance. It would hold only if every report and rule were built on ranges or wildcards that automatically absorb new codes, which is rare and fragile in practice.

• D

  A revised mapping table only, since the cost headings are unaffected

  This is the least wrong of the three, since a mapping table is genuinely one of the things that must change, and it correctly observes that no new cost heading is needed. It stops short of the real implication: the codes themselves must be created under every heading, and reports and rules are built on codes rather than on the mapping. It would be right if the yard analysis lived in a dimension and only the reporting hierarchy needed a new value added to it.

At Northmoor Dairy an automated rule codes haulage invoices to a depot using the delivery reference. A new contract changed the delivery pattern in March, and well-formatted invoices have been coded to the wrong depot since. No exception has been raised. Which control detects this?

• A

  An exception queue with a named owner and a maximum item age

  This is a sound control and the right answer to most automation questions, which is why it is the leading distractor here. An exception queue collects the items the rule could not handle, and these invoices were handled: well formatted, matching the rule, coded with complete confidence to the wrong depot. It would detect the problem if the contract change had caused the delivery reference to fail the rule rather than to satisfy it wrongly, which is the loud failure mode rather than the silent one.

• B

  A review of the change log covering amendments to the coding rules

  This assumes that wrong output means somebody changed something, which is a reasonable first instinct and false here: nobody touched the rule, the contract moved underneath it. A change log review would return a clean result and provide false comfort, since the file shows no amendment since the rule was written. It would be the correct control if the coding had been altered without authority, which is a genuine risk and a different one.

• C

  Re-performing a sample of the coded invoices by hand each period

  Correct: Re-performance is the only control listed that tests whether the answer is right rather than whether the process ran, and a rule that has drifted out of step with the business produces no other signal.

• D

  A monthly reconciliation of capture tool totals to the purchase ledger

  This proves that value crossed the boundary intact, and it would do so successfully every month, because every invoice reached the ledger at the correct amount. Only the depot analysis inside the total is wrong, and a classification error nets to nil against a control total, so the reconciliation cannot see it. It would be the right control if the concern were completeness or value, such as invoices lost in transit or amounts transferred incorrectly.