Beneish M-score

The Beneish model is a statistical model that uses financial ratios calculated with accounting data of a specific company in order to check if it is likely (high probability) that the reported earnings of the company have been manipulated.

How to calculate
The Beneish M-score is calculated using 8 variables (financial ratios): (DSRI) DSRI = (Net Receivablest / Salest) / (Net Receivablest-1 / Salest-1) GMI = [(Salest-1 - COGSt-1) / Salest-1] / [(Salest - COGSt) / Salest] AQI = [1 - (Current Assetst + PP&Et + Securitiest) / Total Assetst] / [1 - ((Current Assetst-1 + PP&Et-1 + Securitiest-1) / Total Assetst-1)] SGI = Salest / Salest-1 DEPI = (Depreciationt-1/ (PP&Et-1 + Depreciationt-1)) / (Depreciationt / (PP&Et + Depreciationt)) SGAI = (SG&A Expenset / Salest) / (SG&A Expenset-1 / Salest-1) LVGI = [(Current Liabilitiest + Total Long Term Debtt) / Total Assetst] / [(Current Liabilitiest-1 + Total Long Term Debtt-1) / Total Assetst-1] TATA = (Income from Continuing Operationst - Cash Flows from Operationst) / Total Assetst
 * Days Sales in Receivables Index
 * Gross Margin Index (GMI)
 * Asset Quality Index (AQI)
 * Sales Growth Index (SGI)
 * Depreciation Index (DEPI)
 * Sales General and Administrative Expenses Index (SGAI)
 * Leverage Index (LVGI)
 * Total Accruals to Total Assets (TATA)

The formula to calculate the M-score is:
 * M-score = −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

How to interpret
The threshold value is -1.78 for the model whose coefficients are reported above. (see Beneish 1999, Beneish, Lee, and Nichols 2013, and Beneish and Vorst 2020).
 * If M-score is less than -1.78, the company is unlikely to be a manipulator. For example, an M-score value of -2.50 suggests a low likelihood of manipulation.
 * If M-score is greater than −1.78, the company is likely to be a manipulator.  For example, an M-score value of -1.50 suggests a high likelihood of manipulation.

Aggregate recession predictor
A 2023 research paper will use an aggregate score of many companies to predict recessions. It finds that the score in early 2023 is the highest in some 40 years.

Important notices

 * Beneish M-score is a probabilistic model, so it cannot detect companies that manipulate their earnings with 100% accuracy.
 * Financial institutions were excluded from the sample in Beneish paper when calculating M-score since these institutions make money through different routes. Sales and receivables which are two main ingredients that go into the Beneish formula are not used when analyzing a financial institution.

Example of successful application
Enron Corporation was correctly identified 1998 as an earnings manipulator by students from Cornell University using M-score. Noticeably, Wall Street financial analysts were still recommending to buy Enron shares at that point in time.

Further reading on financial statement manipulation

 * A sequence of articles on Alpha Architect blog.
 * An article on Investopedia about different types of financial statement manipulation ("smoke and mirrors", "elder abuse", "fleeing town", and others).