Value-Trap Radar — Stocks Flagged by Beneish M-Score

What you’re looking at

Messod Beneish, an accounting professor, published a model in 1999 for spotting companies whose reported profits may have been dressed up. He took a sample of companies that had been caught manipulating earnings, compared their financial statements with those of ordinary companies, and found eight ratios that tended to look different in the year the manipulation happened. He combined those ratios into one number, the M-Score. A 1998 classroom exercise that ran the model on Enron is often cited as an early, real-world hit; the model has also missed well-known frauds, which is covered below.

This list shows the S&P 500 companies whose most recent M-Score is above -1.78, the line Beneish’s work associates with a higher chance of manipulated earnings. The table is sorted from the highest score down, so the top rows are the furthest above that line.

A high score does not mean fraud. It means the company’s reported numbers show the same patterns that past manipulators showed. Sometimes the reason is ordinary: a one-time accounting change, a real jump in growth, a big new customer who hasn’t paid yet. Sometimes it isn’t. The score is a reason to read the filings closely, not a conclusion.

The eight ratios, in plain words

Each ratio compares this year with last year. You don’t need to compute them yourself to use this list, but knowing what the model reacts to helps you read a flagged name.

• DSRI (receivables index). Are unpaid customer bills growing faster than sales? That can mean sales were booked before the cash arrived.
• GMI (gross margin index). Did gross margin fall? Companies under margin pressure have a stronger reason to make the top line look good.
• AQI (asset quality index). Did the share of “soft” assets (anything other than current assets and plant and equipment, such as goodwill and deferred costs) grow? Costs that should have been expensed sometimes end up here.
• SGI (sales growth index). How fast are sales growing? Fast growth isn’t wrong, but it brings pressure to keep the streak going, so the model treats it as a risk factor.
• DEPI (depreciation index). Did the depreciation rate slow? Stretching the assumed life of assets lifts this year’s profit.
• SGAI (overhead index). Did selling and administrative costs grow faster than sales?
• LVGI (leverage index). Did debt rise relative to assets? Companies near loan covenants have a reason to manage reported earnings.
• TATA (accruals to assets). How much of this year’s profit isn’t backed by cash? This is net income minus operating cash flow, divided by total assets.

Beneish gave each ratio a weight and added them up. TATA carries by far the largest weight, so profit that isn’t matched by cash moves the score the most; DSRI and SGI also carry a lot of the signal. Every input comes from two consecutive annual reports, so you can check the arithmetic yourself. The Beneish M-Score lesson walks through it with a made-up company.

How to read the score

• Below -2.22: the clean zone. The patterns the model looks for are absent or faint.
• -2.22 to -1.78: the middle zone. Worth a look at the notes, often nothing.
• Above -1.78: the elevated zone, the companies on this list. The higher the number, the more of the manipulator patterns are present at once. On a company page, the Beneish card shows this zone in red.

The M-Score is an accounting screen, not a valuation. It says nothing about whether a stock is cheap or expensive. A clean score doesn’t make a company a good investment, and a flagged score is not a reason to do anything on its own. It means the reported numbers deserve a closer read.

What this list can’t tell you

• Whether earnings were actually manipulated. The model was fitted to past cases. Many companies with high scores turn out to have an ordinary explanation.
• Anything reliable about banks and insurers. The ratios assume an industrial balance sheet. Scores for financial companies are computed with the same arithmetic but mean less.
• What a one-off event did. A large acquisition, a spin-off or a recapitalization can produce a high score that reflects a structural change rather than aggressive accounting. The segment notes in the 10-K usually explain it.
• Newly public companies. The model needs two full fiscal years of comparable data.
• Frauds that leave no trace in these ratios. Fictitious cash balances, round-trip deals with off-balance-sheet entities and invented customers can all score clean. Wirecard scored clean before it collapsed. The model has flagged several large frauds and missed others.

How to follow up on a name

1. Open the company page and read the other three score cards. A high M-Score next to a strong Piotroski F-Score and operating cash flow above net income reads differently from a high M-Score next to weak cash flow.
2. Open the latest 10-K from the filing timeline. Read the revenue-recognition note, the receivables trend in Management’s Discussion and Analysis, and any note about changes in accounting estimates.
3. Compare the M-Score with the prior year where the data exists. A sharp jump suggests something changed; a stable reading is more often a feature of the business.

Related lists

• Going-concern radar: companies whose Altman Z″-Score sits in the distress zone. A company on both lists deserves the most careful read on the site.
• Highest Piotroski: the strength side. A company that appears on both this list and that one is a reason to read the full 10-K rather than trust either number alone.

Common questions

A company I follow is on this list. What does it mean? That its latest annual numbers show patterns the model associates with manipulated earnings. It is a reason to read the filings closely, in particular the receivables trend and the cash-flow statement against the income statement. It is not a verdict on the company and not advice to do anything.

Why are growth companies so common here? The sales-growth index treats fast growth as a risk factor, and fast-growing companies also tend to sell more on credit and build inventory, which lifts DSRI and TATA. Clean, fast-growing companies do appear on this list for that reason.

Is this updated automatically? The scores are recomputed when a company’s data is refreshed, and this page updates at each site build. Each company page shows the filing date of the inputs behind its scores.

Where do I see the component ratios? Filing.fyi shows the overall M-Score on each company page, with a lesson panel explaining what goes into it. The individual ratios are not broken out on the site; the lesson shows how to compute them from a 10-K.