October 29, 2026 · BriMindInvest Research Team · 15 min read
Reported profit is the output of dozens of judgement calls — when a sale counts as revenue, how long an asset lasts, how large a reserve should be, which costs belong on the balance sheet. Cash collected involves far fewer. When those two numbers diverge for long enough, the cash is usually right. This guide covers the arithmetic that measures the gap, how to normalise working capital so growth does not disguise it, and the eight signals that precede most restatements — none of which require anything beyond the filings.
Accrual accounting exists for a good reason. If a company signs a three-year service contract and collects the cash up front, recognising all of it as profit on day one would misrepresent the business just as badly as recognising none of it. Matching revenue to the period in which it is earned, and costs to the period they support, produces a more faithful picture than raw cash movements do.
The cost of that faithfulness is discretion. Accrual accounting requires management to estimate things that are genuinely uncertain: how much of the receivable book will go bad, how long equipment will remain useful, what a warranty will cost, when a contract obligation has been satisfied. Each estimate is a range, and within the range management chooses.
None of that is fraud. It is the system working as designed. But it means reported profit is a considered opinion, and two honest teams running identical businesses can report materially different earnings. Cash collected from customers involves far less discretion, which is why the relationship between the two is the most informative thing in a set of financial statements.
Think of the gap between profit and cash as a loan the income statement takes from the future. Recognising revenue early, deferring a cost, or trimming a reserve all pull profit forward. The business has not changed; the timing has. Loans come due, and the repayment shows up as a quarter where profit falls without any obvious operating cause.
This is why the analysis below is about trends across years rather than any single number. One period of divergence tells you almost nothing. A direction sustained across five tells you a great deal.
The accrual gap is net income minus cash flow from operations. It takes thirty seconds and it is the highest-yield check available.
A single year like this is a question, not a verdict. A company building inventory ahead of a product launch, or signing large enterprise contracts with long payment terms, can legitimately show this shape. The follow-up is always the same: does it persist, and does the cash flow statement explain where the gap went?
A common objection is that operating cash flow should naturally exceed net income because depreciation is subtracted from profit but added back to cash flow. That is exactly right, and it is why the comparison is informative. For a stable business reinvesting roughly what it consumes, operating cash flow should run above net income. A company whose profit exceeds its operating cash flow is therefore not merely unusual — it is inverted relative to how the accounting normally works.
The main legitimate exception is a fast-growing company funding working capital out of operations: receivables and inventory both expand with the business, consuming cash the income statement has already recognised as profit. That exception is real, and it is also the exact cover story an aggressive reporter would use. Which is why the next section matters.
Growth rates cannot settle the question, because a company growing 40% a year should have receivables growing roughly 40%. Day counts remove the growth and leave the behaviour.
Receivables outgrowing revenue by almost three to one, with DSO extending twelve days, is the clearest version of this signal. The benign explanation — a shift toward larger customers on longer terms — is testable: it should appear in the revenue mix disclosure and it should stabilise once the mix shift completes. A DSO that keeps climbing year after year has run out of benign explanations.
DSO + DIO − DPO gives the cash conversion cycle: the number of days between paying for inputs and collecting from customers. Lower is better, and the direction over five years tells you whether operational discipline is improving or decaying. It is one of the few metrics where the trend is more informative than the level, because the right level differs enormously by industry.
Almost every company reports an adjusted earnings figure alongside the statutory one, and almost every analyst model is built on the adjusted number. Adjusted earnings are not inherently dishonest — excluding a genuine one-off makes the underlying trend clearer. The problem is what counts as one-off.
Restructuring is the usual offender. A single reorganisation is an event. Restructuring charges in five consecutive years are a permanent feature of the cost base being presented as five separate surprises. Stock-based compensation is the other common one: it is a real cost paid in ownership, and excluding it reports profit for labour the company did not pay cash for — the mechanics of which are covered in our guide to buyback and shareholder yield, where the same issuance shows up as dilution.
The honest exclusions tend to share a shape: genuinely unusual in origin, clearly disclosed, non-cash or one-time in substance, and absent from most years. The questionable ones recur, grow, and are described in language that gets vaguer the longer they persist.
Ordered by how much weight each deserves. None is conclusive alone; the value is in how many appear together.
The single most informative divergence in financial statements. Earnings can be shifted between periods by judgement calls on revenue timing, reserves and capitalisation. Cash collected is far harder to manufacture. One quarter of divergence is noise — working capital genuinely swings. Four or more consecutive quarters of profit growing while cash generation shrinks is a pattern, and the pattern almost always resolves toward the cash.
If sales are up 14% and receivables are up 38%, the company is booking revenue it has not collected. Sometimes that is a genuine mix shift toward larger enterprise customers with longer terms. Sometimes it is pulling future orders into the current quarter on soft terms, which borrows from next year to make this one. Days sales outstanding tells you which story is more plausible, because it normalises for growth.
Rising inventory ahead of a genuine demand surge is good planning. Rising inventory into flat or falling sales is a future writedown that has not been taken yet, and in the meantime it flatters gross margin, because unsold goods sit on the balance sheet at cost rather than passing through cost of sales. Days inventory outstanding is the normalised version of this check.
Restructuring, impairment, integration and legal settlement charges are legitimately excluded from adjusted earnings when they are genuinely non-recurring. A company that reports a restructuring charge in each of five consecutive years is not experiencing five unrelated surprises; it is running restructuring as a permanent cost of doing business while presenting adjusted profit that excludes it. Add those charges back and recompute the margin yourself.
Extending the useful life of an asset lowers annual depreciation and raises reported profit without any change in the business. Reducing a warranty or bad-debt reserve does the same. These changes are disclosed, usually in the critical accounting estimates section, and they are sometimes entirely justified. The question to ask is whether the revision is supported by experience or arrived conveniently in a quarter that needed help.
When a cost is capitalised it moves off the income statement onto the balance sheet and is amortised over years, so current profit rises. Software development, certain cloud costs and some customer acquisition spend sit in genuinely grey areas. The useful test is relative: if a company capitalises a materially larger share of a cost than its direct competitors, it is reporting better earnings for the same economic activity.
A company can improve reported operating cash flow simply by paying suppliers later. That is a one-off benefit dressed as an improvement in cash generation — days payable cannot extend forever, and eventually the comparison reverses. Read the working capital lines inside the cash flow statement rather than only the total.
None of these proves anything by itself, and small companies change auditors for mundane reasons. But a dismissed auditor combined with a delayed filing and a disclosed material weakness in internal controls is a cluster, and clusters of governance signals have a far better record of preceding restatements than any single ratio does.
One flag on an otherwise clean company is usually an industry quirk or a genuine business change. Three or more appearing together, especially when one of them is a governance signal, is the configuration worth acting on. Accounting problems are rarely isolated, because the pressure that produces one usually produces several.
Equally important: none of this establishes intent, and most companies showing one or two of these signals are reporting honestly within the range the rules allow. The output of this analysis is a decision about position sizing and further reading, not an accusation.
All of this comes from documents anyone can read. Our walkthrough of how to read a 10-K covers the overall structure; this is the earnings-quality path through it.
For the quarter-by-quarter version of the same work, our guide to reading earnings reports covers what to check when a release lands, and free cash flow goes deeper on which add-backs deserve scrutiny.
Not one year — five. Plot them side by side. Healthy businesses show cash flow tracking or exceeding profit over time, because depreciation is a real add-back and genuine profit converts to cash. A persistent and widening gap is the finding.
Compute days sales outstanding, days inventory outstanding and days payable outstanding across the same five years. Growth rates mislead because a fast-growing company's receivables should grow; day counts do not have that problem.
Take management's adjusted figure, add back every charge excluded in each of the last five years, and see what the trend looks like without the exclusions. If a category appears every year, it belongs in the cost base.
The release leads with the number management wants discussed. The cash flow statement is the one that is hardest to shape, and the working capital detail inside it explains most apparent improvements.
Critical accounting estimates is where the judgement calls are disclosed. Revenue recognition policy tells you when a sale is booked — on shipment, on delivery, on acceptance — and a change in that policy mid-series makes the series non-comparable.
Auditor changes, restatements, material weaknesses, late filings, heavy related-party transactions and unusual executive turnover in the finance function. Any one is weak evidence; three together is a reason to stop and read properly.
Earnings quality is a screen, not a thesis. It does not tell you whether a business is good, whether the moat is durable, or whether the price is reasonable — the work in how to analyze a stock and ROIC versus ROE answers those. What it does is tell you whether the numbers you are about to build a thesis on describe the company accurately. Running it first saves the effort of valuing earnings that are not there.
Every analysis page carries net income, operating cash flow, free cash flow, receivables and inventory with historical financials — so you can run the five-year comparison above on live figures instead of the illustrative ones here.
Earnings quality analysis is not forensic accounting and it does not require special access. It requires comparing two numbers the company publishes itself — profit and operating cash flow — across enough years for a direction to emerge, then normalising working capital into day counts so growth cannot hide the behaviour.
The discipline that does most of the work: never accept an adjusted earnings figure without adding back every exclusion that appears in most years. A cost that recurs is a cost, whatever the reconciliation table calls it.
When profit and cash disagree for long enough, the cash is usually right. The years in between are the opportunity to notice.
This article is educational and is not investment advice. All figures in the worked examples are illustrative and were chosen to make the arithmetic clear; they do not describe any specific company's results. It is also not accounting, audit or legal advice, and it does not allege wrongdoing by any company.
The signals described here indicate elevated risk of future earnings disappointment or restatement. They do not establish that any particular company has acted improperly, and most companies exhibiting one or two of them are reporting entirely within accepted standards. Accounting practice, materiality thresholds and disclosure requirements vary by jurisdiction and change over time.
Past performance does not guarantee future results. BriMindInvest and its contributors may hold positions in securities mentioned in this article. Nothing here constitutes a solicitation to buy or sell any security. Always conduct your own due diligence and consult a qualified financial advisor.
Get AI prediction signals, unlimited stock comparisons, portfolio analytics, and personalized watchlists — free for 14 days, no credit card required.
14-day free trial · No credit card required · Cancel anytime
Not ready to sign up? Get one free email a week instead.