Key point
High-quality earnings are repeatable, cash-backed, and free from aggressive accounting choices. Low-quality earnings rely on accruals, one-time gains, or estimate changes that flatter the income statement without generating cash.
Compare
| Signal | High Quality | Low Quality |
|---|---|---|
| CFO vs Net Income | CFO ≥ Net Income | CFO << Net Income |
| Accrual Ratio | Below 5% | Above 10% |
| Revenue sources | Recurring, cash-collected | One-time gains, unbilled |
| Estimate changes | Stable, conservative | Frequent, income-boosting |
Formula
Accrual Ratio (CF) = (Net Income − CFO − CFI) / Avg Total Assets
Step through
The Beneish M-Score combines eight financial ratios into a single fraud-detection score. An M-Score above −1.78 indicates a high probability of earnings manipulation.
| Variable | What It Measures | Red Flag |
|---|---|---|
| DSRI | Receivables growing faster than revenue | > 1.4 |
| GMI | Gross margin deterioration | > 1.0 |
| AQI | Non-current assets rising (capitalizing expenses) | > 1.0 |
| SGI | Sales growth pressure | High growth + other flags |
| DEPI | Slowing depreciation to boost income | > 1.0 |
| SGAI | SGA rising disproportionately | > 1.0 |
| LVGI | Rising leverage (covenant pressure) | > 1.0 |
| TATA | Earnings are accrual-heavy, not cash-backed | > 0.05 |
M-Score = −4.84 + 0.920(DSRI) + 0.528(GMI) + 0.404(AQI) + 0.892(SGI) + 0.115(DEPI) − 0.172(SGAI) + 4.679(TATA) − 0.327(LVGI)
Try it
Check-in
Key point
Going Deeper — cookie-jar reserves and the small-profits kink. Empirically, far more firms report tiny profits than tiny losses (the "kink" in the histogram around zero) — evidence that managers systematically nudge marginal results across the threshold. The mechanism is reserves: in good quarters, management over-reserves for warranty, litigation, or restructuring; in weak quarters, the reserve is partially released, and the income-statement effect flatters EPS. The reserve becomes a cookie jar that opens when needed. Probing question: "Has this company beat consensus by exactly a penny more than 75% of the time across the last 16 quarters? That alone is a flag."
Key point
Going Deeper — the four-question earnings-quality framework. Apply four tests to any reported earnings number before you multiply by anything. (1) How recurring? Strip out one-time gains, restructuring charges, litigation settlements, and discontinued-operations effects to isolate the run-rate. (2) How certain? Reserve adjustments, deferred-tax true-ups, and goodwill impairments are judgment-heavy; treat them as a narrower subset of the recurring number. (3) How cash-backed? Compute the cash conversion ratio (operating cash flow ÷ net income); a CCR well below 1 over multiple periods is a warning that accruals are doing the heavy lifting. (4) What multiple does the residual deserve? The recurring + certain + cash-backed slice deserves the peer multiple; everything else deserves substantially less. Worked example — Westmoor Optical Q3 reported $48M net income decomposed: $42M operating (recurring, $36M cash-backed), $4M deferred-tax adjustment (judgment heavy, will reverse), $2M settlement (one-time). Core recurring cash earnings $36M, vs the headline $48M. Applied at a 16x peer multiple the gap between disciplined and naive valuation is roughly $190M of equity value — about a third of any small-cap mispricing edge. AI prompt: "Walk through [TICKER]'s most recent quarterly earnings using the four-question framework. Quantify the recurring vs one-time split, the judgment-heavy vs transactional split, and the CCR. Return a decomposed core earnings number."
Sit with the ideas.
A company reports net income of $180M, but operating cash flow is only $95M. Total accruals to total assets (TATA) is 0.08, and the Days Sales in Receivables Index (DSRI) is 1.35. What does this combination suggest?