CPA (CPA Exam) common mistakes
CPA · Understand the traps and how to avoid them.
Answering with what the client or management should do, or with a permissive action, when the requirement asks what the auditor must do. Candidates lock onto the topic of a long scenario and pick the first choice that is a true statement about the topic, ignoring the actor ('the auditor' vs. 'management') and the modal ('must' vs. 'may' vs. 'should consider').
Stems are lengthy and the subject shifts between sentences, so the tested responsibility gets buried under the fact pattern that candidates are still processing.
What to do instead
Before looking at the choices, write two words on the scratch paper: the actor and the verb of the requirement (e.g., 'auditor / must communicate'). Eliminate any choice whose actor or obligation level does not match, even if the statement is otherwise true.
Concluding that controls can be relied on because they are well designed, without evidence that they operated effectively throughout the period. Candidates select 'assess control risk below maximum' after reading a description of a well-documented control activity.
Design and operating effectiveness are taught together and look similar in a narrative exhibit; a good flowchart feels like proof that the control works.
What to do instead
For every control question, force a two-step check: (1) Is there evidence of performance (sign-offs, reperformance, samples)? (2) Does that evidence cover the whole period of reliance? If either answer is no, control risk stays at maximum and substantive procedures drive the answer.
Computing an item under the wrong framework, for example adding a revaluation surplus, reversing an inventory write-down, or capitalizing development costs when the stem specifies U.S. GAAP.
Review materials place GAAP and IFRS rules side by side in the same table or list, so both rules get encoded with equal salience and the framework named in the stem is not treated as a filter.
What to do instead
In the first three seconds of any computational item, write 'GAAP' or 'IFRS' at the top of the scratch area and cross out every rule that belongs to the other framework before doing arithmetic.
Making a sign or roll-forward error inside multi-step computations, such as adding a premium amortization instead of subtracting it, or mixing a discount into interest expense in the wrong direction, and then selecting the answer that matches the flawed arithmetic.
These computations are done inline under time pressure, and one sign slip produces a number that appears among the distractors, so nothing signals the error.
What to do instead
Write the roll-forward as a labeled table before plugging in numbers (beginning carrying amount, plus accretion, minus amortization, equals ending carrying amount) and check that the ending figure is on the correct side of face value.
Applying memorized dollar thresholds from a prior tax year, such as the standard deduction, the Section 179 limit, a phase-out range, or the gift tax annual exclusion, instead of the figures supplied with the item.
Candidates drill these numbers as flashcards, and the memorized value is retrieved faster than the stimulus is read.
What to do instead
Never compute from a remembered threshold. If a dollar limit is not stated in the stem or the provided tables, the item is testing a concept, not the limit. Where tables or instructions are available, navigate to the figure rather than recalling it.
Deducting a flow-through loss after applying only stock basis, skipping the ordering of the remaining limitation layers (at-risk amount, passive activity rules, excess business loss).
Basis is the layer taught first and tested most, so it becomes the only filter candidates apply; each additional layer feels like a separate topic rather than a sequence.
What to do instead
Draw a staircase before computing: basis, then at-risk, then passive, then excess business loss. Carry the running balance down each step and state the carryforward at each rung, even when the item only asks for the final deduction.
Selecting 'substantial authority' as the standard for a position taken on a return that involves a tax shelter or reportable transaction, or discussing the standard of authority without addressing the related penalty and disclosure consequences.
Substantial authority is the phrase most associated with return positions, and the shelter/non-shelter distinction is a single line in the study material that gets compressed out.
What to do instead
Build a three-column grid before answering: position type, required confidence level (more likely than not for shelters and reportable transactions, substantial authority otherwise), and required disclosure. Read the choice against all three columns, not just the confidence level.
Reporting a variance or ratio as favorable or unfavorable, or as the total variance, without decomposing it into its price and quantity (or volume and mix) components, and without fixing which factor is held constant.
Spreadsheet habits carry over: a single total variance is easy to compute and feels complete, and the favorable/unfavorable convention flips depending on whether the item is revenue or cost.
What to do instead
Before choosing an answer, write the two-way decomposition and label which variable is held constant in each component. Then apply the sign convention explicitly: for revenue items, actual above standard is favorable; for cost items, actual below standard is favorable.
Choosing a SOC report by matching its number to a vague memory of the report's nickname rather than to the criteria being reported on, for example selecting a SOC 1 report for a service organization whose customers need assurance on security, availability, and confidentiality.
SOC 1, 2, and 3 are learned as a numbered list, and under exam pressure the number-to-label association is retrieved faster than the underlying criteria.
What to do instead
Pre-write a three-line mapping at the start of the exam section: SOC 1 equals controls relevant to user entities' internal control over financial reporting; SOC 2 equals trust services criteria with restricted use; SOC 3 equals trust services criteria in a general-use report. Read the item's stated user need, not the report's name.
Applying an entity-level rule to the wrong taxpayer type, such as imposing individual rates and net investment income tax thresholds on a trust or estate, or applying corporate treatment to an entity that has made a pass-through election.
Rate tables and threshold questions are practiced most often at the individual level, and candidates skim past the entity description in the fact pattern.
What to do instead
Make the first annotation on every tax computation the entity type and the tax year, then verify the applicable rate schedule, threshold, and filing requirement against those two labels before computing anything.