Beneish M-Score Screener

By Christian Richter | Last updated 2026-09-23

Messod Beneish published the M-Score in 1999 while at Indiana University. The idea: earnings manipulation leaves traces in the financial statements. Companies that inflate revenue, delay expenses, or capitalize costs they should not show predictable year-over-year shifts in eight financial ratios. Beneish combined these eight indices into a single probit score. A score above -1.78 flags a company as a likely manipulator. The heaviest-weighted term is TATA, total accruals to total assets, which catches persistent accruals that never convert to cash. Unlike the Piotroski F-Score (which measures whether a company is getting financially stronger), the M-Score measures whether the financial statements themselves are reliable. No free tool screens the entire US market by Beneish M-Score. The alternatives are single-company calculators and downloadable spreadsheet templates. On Shibui, you describe the screen you want and Claude computes all eight indices across the full market in one pass. For a step-by-step walkthrough of each index, see the Beneish M-Score screening guide.

What the screener checks

The eight manipulation indices

Each index compares this year's financial ratio to last year's. A value of 1.0 means no change. Values far from 1.0 in the flagged direction contribute to a higher (more suspicious) composite score. Beneish combined them using a probit model with these weights:

The composite formula:
M = -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

Above -1.78: manipulation likely. Between -2.22 and -1.78: grey zone. Below -2.22: unlikely manipulator.

Caveats worth knowing. The depreciation index uses depreciation and amortization combined, not pure depreciation, because most data sources do not separate the two. This is the standard approximation used in practice. Financial-sector companies (banks, insurers) often lack cost of revenue, which means the gross margin index cannot be computed for them. High-growth companies frequently trigger the score because rapid revenue growth and shifting asset mixes mimic manipulation signals. A flagged score is a reason to investigate, not a verdict.

Example screens

The forensic red-flag screen. The prompt spells out the formula so Claude computes it correctly from annual financial statements. Each index compares this year's ratio to last year's. The probit weights come from Beneish's original 1999 paper. Adding the Piotroski filter surfaces stocks that are both statistically suspicious and financially weak. This combination is what no other free screener can do in a single step.

Per-company forensic check. The prompt defines each index so Claude maps them to the right financial statement fields. You can see exactly which signals triggered and whether the overall score is driven by one outlier index (often SGI for high-growth names like NVIDIA) or a broad pattern across multiple indices.

Trend-change detection. This finds companies whose earnings quality deteriorated between fiscal years. A stock that was safely below the threshold last year but crossed above it this year shows a shift worth investigating. The prompt includes the formula so Claude can compute the score at each historical point. Spreadsheet templates and single-company calculators cannot run this across the full market.

How it compares to other Beneish tools

Most Beneish M-Score tools either calculate the score for one company at a time or provide a spreadsheet template you fill in manually. No other free tool runs the computation across the full US market or lets you combine it with other filters.

For financial strength scoring, see the Piotroski F-Score screener. For bankruptcy risk, see the Altman Z-Score screener. The three are complementary: Piotroski checks whether a company is improving, Beneish checks whether the improvement is real, Altman checks whether the company can pay its debts. For earnings momentum signals, see the earnings surprise screener. For tracking what insiders are doing with their own money, see insider buying signals. See the AI stock screener overview for how Shibui compares on other screening approaches.

Want to run this screen daily on a schedule? See the automated stock screener workflow for scheduling commands and prompt reliability patterns.

Frequently asked questions

What is the Beneish M-Score?

The Beneish M-Score is a mathematical model that uses eight financial ratios to estimate the probability that a company has manipulated its reported earnings. Messod Beneish published it in 1999 in the Financial Analysts Journal. The eight indices measure changes in receivables, gross margins, asset quality, sales growth, depreciation, SG&A expenses, leverage, and accruals relative to the prior year. A composite score above -1.78 flags a company as a likely manipulator. Shibui computes all eight indices from annual financial statements for nearly 10,000+ US stocks.

Is there a free Beneish M-Score screener?

Shibui Finance is a free market-wide Beneish M-Score screener. Connect it to Claude (the free plan works) and describe your screen in plain English. Most existing Beneish tools are single-company calculators or downloadable spreadsheet templates. No other free tool screens the entire US market by M-Score and lets you combine it with valuation, financial strength, or technical filters in one query.

What does a Beneish M-Score above -1.78 mean?

A score above -1.78 means the company's financial statements show patterns statistically associated with earnings manipulation. It is a probability flag, not proof of fraud. Beneish's original research found this threshold correctly identified about 76% of known manipulators. However, high-growth companies often trigger the score because rapid revenue growth and shifting asset composition mimic manipulation signals. Scores between -2.22 and -1.78 are sometimes called the grey zone. Below -2.22 is considered unlikely to be a manipulator.

How is the Beneish M-Score different from the Piotroski F-Score?

The Beneish M-Score detects earnings manipulation (are the financial statements reliable?). The Piotroski F-Score measures financial strength (is the company getting stronger?). They answer different questions and use different inputs. Beneish compares year-over-year changes in eight ratios to flag suspicious patterns. Piotroski checks nine binary signals for profitability, leverage, and efficiency. The two scores are complementary. A stock with a high Piotroski F-Score but a flagged Beneish M-Score deserves closer inspection: the strength may be artificial.

Can you backtest the Beneish M-Score on Shibui?

Yes. Shibui has annual financial data going back 20 or more years for most companies. You can ask Claude to compute the M-Score at any point in the past and check whether flagged companies later restated earnings, suffered price declines, or faced SEC enforcement. This is a historical backtest, not a forward-looking prediction. Survivorship bias applies since delisted companies are not in the dataset.

Does Finviz have a Beneish M-Score screener?

No. Finviz does not include the Beneish M-Score as a screening filter. Neither do most other free screeners. The only widely available Beneish tools are single-company calculators (like MetricGate) and downloadable spreadsheet templates. Shibui computes all eight indices across the full US market and lets you combine them with any other filter in a single plain-English query.