Calcady
Home / Financial / Beneish M-Score Earnings Manipulation Calculator

Beneish M-Score Earnings Manipulation Calculator

Calculate the Beneish M-Score to detect potential financial fraud and aggressive earnings manipulation in corporate financial statements using eight weighted fundamental ratios.

Financial Statement Data

Current Year

Previous Year

$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$

Beneish M-Score

-2.191
Manipulation Index
Low Manipulation RiskScore securely rests below the -1.78 threshold line. Accounting variables scale cohesively.

Index Sub-Components

DSRI:1.35
GMI:0.97
AQI:0.87
SGI:1.11
DEPI:0.92
SGAI:1.00
LVGI:1.00
TATA:-0.013
Email LinkText/SMSWhatsApp

Quick Answer: What does the Beneish M-Score tell you?

The Beneish M-Score is a probabilistic earnings manipulation screening model using 8 weighted financial ratios: M = −4.84 + 0.92·DSRI + 0.528·GMI + 0.404·AQI + 0.892·SGI + 0.115·DEPI − 0.172·SGAI + 4.679·TATA − 0.327·LVGI. Interpretation: M > −2.22 = likely manipulator; M < −2.22 = likely non-manipulator. Each index compares a current-year financial ratio to the prior year — values >1 generally indicate deterioration or unusual growth raising manipulation risk. The model flagged Enron at +0.765 in 1998 using only publicly available 10-K data, three years before its collapse.

The 8 M-Score Indices: What Each One Detects

Each index is derived from two consecutive years of audited GAAP financial statements. The model coefficient (weight) was determined by Beneish's logistic regression on 1,708 companies. Higher weight = more statistically significant in predicting manipulation. Always use GAAP figures — non-GAAP “adjusted” numbers will systematically understate manipulation signals.

Index Weight What it Detects Red Flag Value
DSRI+0.920Days Sales in Receivables growing faster than revenue — channel stuffing or premature revenue recognition> 1.3
GMI+0.528Gross margin declining year-over-year — margin pressure creates incentive to manipulate revenue or understate COGS> 1.2
AQI+0.404Non-current assets (ex PP&E) growing as fraction of total assets — off-balance-sheet cost capitalization> 1.25
SGI+0.892High revenue growth — growth companies have the most to lose if growth falters; manipulation risk peaks at 20–50% annual growth> 1.6
DEPI+0.115Depreciation rate declining — extending useful-life estimates to reduce D&A expense and inflate operating income> 1.0
SGAI−0.172SGA costs growing faster than revenue — negative weight (model found SGA bloat not predictive of manipulation vs. lean firms)Context-dependent
TATA+4.679Most predictive of all 8. (Net Income − Operating Cash Flow) / Total Assets. Large positive TATA = income far exceeds cash collected, hallmark of accrual manipulation> 0.05
LVGI−0.327Leverage increasing — negative weight (Beneish found highly leveraged firms manipulate differently; model adjusts for this)Context-dependent
Source all figures from GAAP income statement, balance sheet, and cash flow statement as reported. Do NOT use non-GAAP or adjusted figures — management-adjusted numbers omit exactly the items TATA and AQI are designed to detect. The model requires two consecutive fiscal years of data (year t and year t−1).

M-Score Interpretation Bands

M-Score Range Interpretation Recommended Action
> −1.49Strong manipulation signalFlag for forensic review; check TATA and DSRI drivers; cross-check cash flow statement
−2.22 to −1.49Likely manipulator (above threshold)Investigate further; review auditor notes and SEC comment letters
−2.50 to −2.22Borderline (grey zone)Monitor closely; re-run after next annual filing; check peer group scores
−3.00 to −2.50Likely non-manipulatorLow concern; routine due diligence sufficient
< −3.00Very unlikely manipulatorMinimal earnings quality risk from this model alone
The −2.22 threshold yields ~76% sensitivity and ~17.5% false-positive rate. The alternative −1.78 threshold reduces false negatives at the cost of more false alarms. A score above −2.22 does not prove fraud — it indicates the financial pattern is statistically consistent with companies that previously manipulated earnings. Always use M-Score as one input in a broader earnings quality framework.

Pro Tips & Common M-Score Mistakes

Do This

  • Use M-Score as a first-pass screening tool, then direct deeper analysis at high-scoring names. Run it across a portfolio or peer group, rank by score, then focus forensic work (cash flow reconciliation, SEC comment letter review, auditor change history, insider selling patterns) on the top flagged companies. Almost all professional short-sellers use earnings quality models as their first filter — not the investment thesis itself.
  • Always source inputs from GAAP financial statements, never from non-GAAP or adjusted figures. Management-adjusted EBITDA and adjusted EPS regularly exclude items that are exactly what TATA and AQI are designed to detect — impairment charges, restructuring costs, and stock-based compensation are frequently excluded from non-GAAP earnings while inflating cash flow adjustments. Using adjusted figures will systematically understate the manipulation signal from TATA.

Avoid This

  • Don't apply M-Score to banks, insurance companies, or financial institutions. The model was calibrated on industrial/commercial companies. Financial institutions have fundamentally different receivables structures (loans, not trade receivables), completely different accrual mechanisms, and regulatory capital ratios that make DSRI, AQI, and TATA meaningless as calibrated. You will get nonsense results. For financial-sector earnings quality analysis, use the Piotroski F-Score or bespoke credit model frameworks instead.
  • Don't compare a high-growth tech company's M-Score to the universal −2.22 threshold without peer context. High-growth technology and biotech companies structurally score above −2.22 due to rapid receivable growth (DSRI) and high capitalized R&D (AQI) — even when completely honest. Always compare a company's M-Score trend to its peer group using the same SIC code. A score of −1.8 for a 40% growth SaaS company may be unremarkable; the same score for a mature industrial company is deeply concerning.

Frequently Asked Questions

Why is the M-Score threshold −2.22 and what does it mean statistically?

The M-Score is the output of a logistic regression model estimated on 74 manipulator firms and 1,708 non-manipulators from SEC enforcement actions. The threshold of −2.22 was chosen by Beneish to minimize total misclassification cost, assuming the cost of missing a manipulator (false negative) is roughly twice the cost of a false alarm (false positive). At this cutoff, the model correctly classified 76% of actual manipulators with a 17.5% false-positive rate. The alternative −1.78 threshold reduces false negatives at the cost of more false alarms — appropriate for short-sellers where missing a fraud is more costly than a wasted investigation. The M-Score is a continuous variable: a score of −0.5 is far more alarming than −2.18 even though both technically exceed the threshold.

Why does TATA carry the highest weight (+4.679) in the formula?

TATA = (Net Income − Operating Cash Flow) / Total Assets is the most direct measure of accrual-based earnings manipulation. When a company reports income far above its cash generation, those earnings consist of accruals — revenue recognized but not yet collected, or expenses deferred into the future. Honest, healthy companies typically have TATA near zero or slightly negative (cash earnings match or slightly exceed reported income). The Sloan (1996) accrual anomaly — which predates Beneish — showed that high-accrual companies systematically underperform over the following year as accruals reverse. During an earnings manipulation, accruals are the mechanism: premature revenue recognition creates accounts receivable without cash; deferred expense recognition creates liabilities without cash payment. TATA captures both. Critically, TATA must be calculated using GAAP operating cash flow from the formal cash flow statement — not an estimated or adjusted figure — otherwise it loses most of its predictive power.

How does the Beneish M-Score differ from the Altman Z-Score?

These models answer fundamentally different questions. The Altman Z-Score predicts bankruptcy risk: “Will this company become insolvent within 2 years?” The Beneish M-Score detects earnings manipulation risk: “Has management distorted the financial statements to misrepresent performance?” A financially healthy company can have a high M-Score (manipulating from a solid foundation). A distressed company can have a low M-Score (honestly reporting its poor results). They are complementary: a company with a high M-Score and a deteriorating Z-Score is a particularly serious warning — the manipulation may be concealing a worsening fundamental position that will eventually force insolvency. Professional forensic analysts and activist short-sellers typically run both models as the first two filters in an earnings quality framework before conducting primary research.

Can the Beneish M-Score be used for private companies or non-US companies?

The model was developed on US public companies (GAAP/SEC reporting). For IFRS-reporting non-US companies: the key inputs (receivables, sales, PP&E, net income, operating cash flow) map consistently between GAAP and IFRS on most line items. The model can be applied directionally, but the −2.22 threshold has not been independently validated for IFRS reporters — use it as a relative ranking tool rather than a binary classifier. For private companies: the model requires two consecutive years of accrual-basis financial statements with a formal cash flow statement. Without a cash flow statement, TATA (the most predictive variable, weight +4.679) cannot be properly computed, and the model loses most of its predictive power. In this case, use the remaining 7 indices as a partial screen and supplement by manually inspecting accounts receivable aging, management representation letters, and auditor independence disclosures.

Related Calculators