CMA Part 1 common mistakes
CMA Part 1 · Understand the traps and how to avoid them.
When a revenue recognition scenario is presented, the candidate recites the five-step model from memory but never actually applies the transfer-of-control test to the specific facts, defaulting instead to 'revenue is recognized when the goods ship.'
The five-step model is studied as a list to memorize rather than as a decision procedure, so under time pressure the candidate falls back on a shipping-point heuristic and misses bill-and-hold, consignment, principal-versus-agent, and over-time recognition cues embedded in the fact pattern.
What to do instead
Before choosing an answer, write the words 'control transferred when?' and force yourself to cite the single fact in the stem that establishes transfer (customer acceptance, physical possession, legal title, significant risks and rewards, right to payment). Drill 15 short scenarios where shipping occurs but control has not transferred, and 15 where it has, and state the deciding fact for each.
On topics where IFRS and U.S. GAAP diverge, the candidate applies the U.S. GAAP rule automatically, even when the stem is silent about the framework or explicitly says IFRS.
Most study time is spent on U.S. GAAP and the differences are learned as a late-stage 'comparison table' that is memorized passively; the divergence points that appear most often (LIFO prohibition, inventory write-down reversal, development costs, revaluation of PP&E, component depreciation) decay fastest.
What to do instead
Build a one-page divergence sheet limited to six items and rehearse it as paired flashcards with the stem phrased both ways ('under IFRS...' / 'under U.S. GAAP...'). When a stem names no framework, treat the U.S. GAAP answer as the default only after confirming no IFRS-only option exists.
In a cash budget or purchases budget, the candidate computes cash collections equal to the current month's sales, ignoring the collection pattern and the opening accounts receivable balance.
The candidate anchors on the sales forecast as the headline number in the exhibit and treats the collection percentages as decoration rather than as the actual computational engine; the beginning receivables balance is mentally filed as 'already collected.'
What to do instead
Before computing anything, redraw the budget as a matrix with months across the top and 'collected from month X sales' rows down the side, then fill the cells from the collection percentages. Always add a separate top row for beginning accounts receivable and a bottom row for the ending balance, and finish by checking that total collections over the horizon equal total sales plus opening receivables minus closing receivables.
When a forecast question involves a learning curve, the candidate applies the unit-time formula to the second unit and then multiplies linearly, or averages the incremental and cumulative results.
The candidate confuses the cumulative-average model with the incremental-unit model because both use the same percentage, and exam questions are often written to be solvable only under one convention.
What to do instead
Commit to two separate worked templates and label them at the top of your scratch page before starting: cumulative-average (double the cumulative output, take the percentage of the cumulative average time, multiply by cumulative units) and incremental-unit (percentage applies to the marginal unit time). Practice by computing the total time for 4 units under both models from identical inputs and observing that the answers differ; this contrast cements which model a stem is describing.
The candidate computes ROI using the division's net income and year-end asset balances, and then uses that same figure to judge a manager's performance, without checking whether the question asks about the division as an economic investment or the manager as a steward of controllable resources.
Textbook formulas are memorized with one set of inputs; the distinction between operating income and net income, and between average and ending invested capital, is treated as a definitional nicety rather than as the point of the question, and controllable versus noncontrollable items are blended.
What to do instead
When you see ROI, RI, or EVA, first write 'who/what is being evaluated?' and 'controllable?' at the top of your scratch space. Then select inputs deliberately: operating income and average invested capital for divisional performance, and strip out allocated corporate charges and items the manager cannot influence. Recompute the same division under both an ending-balance and average-balance assumption so you can see which one the answer choices reward.
In variance analysis, the candidate computes a variance correctly in magnitude but labels it favorable or unfavorable by intuition, and assigns the price variance to the production supervisor rather than to purchasing.
Sign conventions (cost minus actual, or actual minus standard) are learned inconsistently across materials, and responsibility assignment is memorized as a table rather than derived from who controls the driver of the variance.
What to do instead
Adopt a single rule and never deviate: for a cost, a positive (standard minus actual) result is favorable. Write the rule in words at the top of every variance question before computing. For responsibility, ask 'who controls the input that moved?' - material price is purchasing, material quantity and labor efficiency are production, labor rate is usually a mix of hiring and scheduling. Redo five problems reading only the sign of your computed number, without looking at the label.
In a process-costing question, the candidate computes equivalent units using total units started rather than units completed plus equivalent units in ending work in process (and, under FIFO, subtracts nothing for beginning inventory).
Physical units and equivalent units get conflated because both appear in the same production report, and the weighted-average versus FIFO distinction is treated as an alternative method rather than as a rule about which units are counted in this period's denominator.
What to do instead
Sketch the physical flow first as a T-account: beginning WIP plus started equals completed plus ending WIP. Then, under weighted average, build equivalent units from completed units plus ending WIP percentage complete; under FIFO, use completed units plus ending WIP minus beginning WIP percentage complete. Always state which method you are using in writing before touching the cost per equivalent unit calculation.
The candidate allocates all overhead using a single plantwide rate and then disposes of over- or underapplied overhead by writing it off entirely to cost of goods sold without considering whether the amount is material or how the question asks for it to be prorated.
Plantwide allocation is the first method learned and the simplest to execute, so departmental rates and the materiality judgment on the disposition of the balance get skipped; the concept that underapplied overhead raises cost of goods sold is memorized as a direction rather than as a journal-entry consequence.
What to do instead
When overhead appears, list the allocation bases available in the stem and check whether different departments have different drivers; if so, use departmental rates. Compute the applied-versus-actual difference, then explicitly write 'immaterial -> close to COGS' or 'material -> prorate to WIP, finished goods, COGS' based on the magnitudes given. Rehearse the four-line journal entry for both closing and proration so the direction of the effect is automatic.
When analyzing an internal control scenario, the candidate concludes that the presence of a control activity (a reconciliation, a segregation of duties, an approval) means the risk is addressed, without identifying which COSO component is deficient or whether the control is designed versus operating effectively.
The framework is memorized as five names rather than as an evaluation sequence, and questions that describe a well-designed control that was overridden or never performed are answered on design alone.
What to do instead
Answer control questions in a fixed three-beat sequence: identify the assertion or objective at risk, map it to the specific COSO component and principle (control environment, risk assessment, control activities, information and communication, monitoring), then separate 'is the control designed appropriately?' from 'did it operate as designed and is it monitored?'. Treat any fact about override, lack of monitoring, or unaudited management estimates as a tone-at-the-top or monitoring defect regardless of how good the control description sounds.
The candidate answers analytics questions by matching a keyword in the stem to a term ('predictive', 'AI', 'robotic') rather than by describing what the tool actually does with the data.
The technology domain feels unfamiliar and low-priority relative to computational topics, so it is studied as vocabulary lists; the distinctions between descriptive, diagnostic, predictive and prescriptive analytics, and between rule-based RPA, machine learning, and generative AI, are learned as labels without operational meaning.
What to do instead
Rewrite each term as a verb phrase you can defend: descriptive reports what happened, diagnostic explains why, predictive estimates what will happen, prescriptive recommends an action; RPA follows deterministic rules on structured data, machine learning infers a model from data, generative AI produces new content. For every practice item, state the verb phrase before selecting the option, and specifically flag that RPA cannot handle unstructured judgment and that predictive output requires human governance and monitoring for bias and drift.