Beneish M-Score: what 3054 US filers actually report
Half of the 3054 companies with a filed beneish m-score sit at -2.54 or below. The bottom quarter is at -2.82 or below, the top quarter starts at -2.21, and the top tenth at -1.07.
Measured 2026-09-07 from accounts filed with the SEC, 0 days ago, one row per company. The cohort is rebuilt nightly and is flagged stale after 35 days. 4277 of the 5257 listings in the current pool hold this figure at all.
| Cohort | Companies | p5 | p10 | p25 | p50 | p75 | p90 | p95 |
|---|---|---|---|---|---|---|---|---|
| Every filer | 3054 | -3.96 | -3.29 | -2.82 | -2.54 | -2.21 | -1.07 | 1.05 |
| Healthcare | 600 | -5.13 | -3.92 | -2.85 | -2.42 | -1.63 | 0.37 | 5.00 |
| Technology | 581 | -3.82 | -3.32 | -2.80 | -2.48 | -2.06 | -0.62 | 2.40 |
| Industrials | 508 | -3.76 | -3.24 | -2.81 | -2.55 | -2.29 | -1.66 | 0.28 |
| Consumer Cyclical | 413 | -3.34 | -3.08 | -2.80 | -2.60 | -2.37 | -1.93 | -1.03 |
| Energy | 213 | -4.04 | -3.55 | -3.05 | -2.71 | -2.38 | -1.86 | -1.00 |
| Consumer Defensive | 176 | -3.37 | -3.05 | -2.77 | -2.55 | -2.22 | -1.35 | 0.47 |
| Basic Materials | 176 | -3.93 | -3.29 | -2.77 | -2.57 | -2.21 | -0.98 | 1.19 |
| Communication Services | 164 | -3.42 | -3.16 | -2.87 | -2.63 | -2.41 | -1.98 | -1.50 |
| Utilities | 94 | -3.21 | -2.97 | -2.73 | -2.59 | -2.47 | -2.21 | -2.09 |
| Financial Services | 83 | -3.26 | -2.90 | -2.60 | -2.41 | -2.28 | -2.01 | -0.58 |
| Real Estate | 45 | -4.42 | -3.17 | -2.70 | -2.53 | -2.09 | 0.35 | 5.00 |
3054 companies in the market row. How a cohort is counted: One deduplicated company per cohort row, drawn only from companies with filed statements on record. One row per company. Key: case-folded, whitespace-collapsed stocks.name. Representative listing: most archived fiscal years, then largest stored market cap, then ticker ascending.
Selection rule actually applied to the market row: tier=VERIFIED; one row per company; exact zeros excluded as coerced missing data; 740 excluded as not applicable (Bank, Capital markets, Equity REIT, Insurer, Mortgage REIT); 0 excluded as untraceable to a filing; winsorised at p1/p99; min cohort 30. The clamp ran between -7.74 and 5.00. Clamped to the observed values at p1 and p99 before the breakpoints were computed. Values are clamped, never dropped, so the cohort size is the cohort that produced the breakpoints.
Every sector this build published for cleared the 30-company minimum on this metric, so no row is withheld. Every sector this build published a distribution for, on any metric, at this coverage tier. A sector named as refused therefore had enough companies for some other metric and not for this one.
How this figure is computed
An eight-part score built to flag accounts that fit the pattern of manipulation.
The arithmetic, as the product prints it when a subscriber expands the figure: -4.84 + 0.92 x DSRI + 0.528 x GMI + 0.404 x AQI + 0.892 x SGI + 0.115 x DEPI - 0.172 x SGAI + 4.679 x TATA - 0.327 x LVGI.
It is one of the 118 indicators computed for every covered listing, filed under integrity. That count and that pillar are read from the product's own indicator registry when this page renders, so a metric that leaves the registry stops being described as one of its indicators here.
What the number hides
It detects a statistical pattern, not wrongdoing. Fast-growing companies trip several of its components for ordinary reasons: receivables rising with sales, margin falling as the business scales, assets shifting as it builds. A score on the wrong side of the usual cut-off is a reason to read the filing, and that is the only thing it is a reason to do. A lower reading is the cleaner one here, which is the opposite direction to most of this section.
What this page leaves out
It is one cross-section, not a history: these marks describe where filers stood on 2026-09-07 and say nothing about where they stood a year ago. Every figure behind it comes from a filed statement rather than from a price, so nothing here moves with the market. Listings in the current equity pool holding this figure at all, counted by GET /api/screener/fields. Listings, not deduplicated companies, and counted today rather than on the build date, so it is a coverage figure and not the cohort.
Cross-sectional statistics computed from filed accounts held on the build date shown. A breakpoint describes where a figure sat among comparable companies, not whether any company is worth owning. This is research, not advice.
This figure on named companies, against their filings
- No company page uses this figure yet. The set of companies this section publishes is a fixed authorised list rather than a product of the ticker universe, so a metric can have a market distribution before it has a single named example.