FPAC Part I common mistakes
FPAC Part I · Understand the traps and how to avoid them.
Treating a year-over-year change in a reported line item as a real economic change without checking whether the prior year was restated or reclassified.
Candidates read the current filing's comparative column as if it were the originally reported figure, so acquisitions, discontinued operations, or accounting-policy changes get absorbed into the growth rate.
What to do instead
Before computing any trend, scan the notes and MD&A for restatement, reclassification, and discontinued-operations disclosures; rebuild the prior year on the current basis or explicitly flag the series as non-comparable.
Computing a ratio with mismatched numerator and denominator, such as dividing flow items by a point-in-time balance without averaging, or mixing pre-tax and after-tax figures.
Formulas are memorized as shapes rather than as economic relationships, so the candidate does not ask what period or basis each input represents.
What to do instead
For every ratio, state in one sentence what the numerator and denominator each measure and over what period; where a flow is divided by a stock, use the average balance and say so.
Projecting free cash flow without tying capital expenditure and working-capital investment back to the revenue growth assumption.
Each line is forecast independently, so the model produces growth that is not funded and the candidate never notices the implied capex-to-sales or days ratios are implausible.
What to do instead
After building the forecast, back-solve the implied drivers (capex as percent of sales, DSO, DIO, DPO) and compare them to the company's history and peers; reconcile any divergence before accepting the output.
Reading a liquidity or coverage conclusion off a single ratio while ignoring the maturity profile and covenant terms.
Candidates treat the current ratio or interest coverage as a verdict, which is easy to compute and easy to defend, and skip the debt footnote where the actual refinancing risk sits.
What to do instead
Pair every ratio-based conclusion with a maturity schedule and covenant headroom check; if the near-term maturities cannot be met from projected cash flow plus committed facilities, the ratio is not the answer.
Mixing book values and market values inside one calculation, for example using a market equity value with a book debt figure in a cost-of-capital or leverage computation.
Inputs are pulled from whichever source is closest at hand, and the candidate does not check that the weights and rates share a consistent valuation basis.
What to do instead
Label every input as book or market before use; in WACC, weight market values of equity and debt (or state the book-value proxy and its limitation), and never blend bases silently.
Accepting a system-generated report total or dashboard figure without reconciling it to the underlying ledger or source extract.
Automated outputs carry an implicit authority, and the candidate assumes the extract, filter, or join was correct because the number looked plausible.
What to do instead
Perform a control total tie-out (record count and sum) between the report and the source, and document the reconciliation; treat any unexplained variance as a blocker rather than a rounding item.
Building a spreadsheet model with hard-coded numbers embedded in formulas and no separation between inputs, calculations, and outputs.
Models are built forward while solving the question, so speed is prioritized over structure and the candidate never returns to generalize the assumptions.
What to do instead
Adopt a three-section layout with a single input block, color-coded and formula-driven calculations, and one output area; any change to an assumption should flow through without editing a formula.
Judging a data set or model output as reliable without checking completeness, duplicates, and period alignment in the source data.
Analytical attention goes to the calculation, and data preparation is treated as a clerical step that someone else has already done.
What to do instead
Run a standard pre-analysis checklist on the extract: row counts versus source, duplicate key test, missing-value profile, and date-range alignment; record the results so the analysis can be defended.
Trusting an automated calculation, macro, or AI-assisted output without verifying the logic against a hand-worked example.
Tool outputs are fast and formatted convincingly, and the candidate has not internalized that the tool executes the logic it was given, including the errors.
What to do instead
Independently recompute one representative case by hand or by a second method and reconcile to the tool's result; if they differ, resolve the difference before using any of the output.
Presenting an analysis without stating the assumptions, data sources, and limitations that the conclusion depends on.
The candidate focuses on reaching a number and treats documentation as optional, so the reader cannot tell which sensitivities would change the answer.
What to do instead
Attach a short assumptions-and-sources note to every deliverable, list the two or three assumptions the conclusion is most sensitive to, and show the effect of varying them.