The Beneish M-Score

Messina Beneish built a statistical model in 1999 by comparing the financial statements of companies later confirmed to have manipulated earnings against a matched set of companies that had not. The result, the Beneish M-Score, combines eight accounting ratios — covering receivables growth, gross margin changes, asset quality, sales growth, depreciation rates, and leverage, among others — into a single score using weights estimated from that historical comparison.

A commonly cited threshold puts scores above roughly -1.78 into the "likely manipulator" zone, though the model was built on a specific historical sample and should be read as a screening flag rather than proof of wrongdoing — plenty of genuinely aggressive but legal accounting choices can push a score higher without any fraud being present.

Worked example. One of the eight inputs is the Days Sales in Receivables Index, which compares how fast receivables are growing relative to sales versus the prior year. If receivables grew 40% while sales grew only 10%, that ratio comes in well above 1, contributing a positive (manipulation-suggestive) push to the overall score — the classic pattern of a company booking sales faster than it is actually collecting cash for them.

Forensic-accounting screens like the M-Score are best used across a portfolio of holdings to flag names worth a closer manual read of the filings, not as a standalone signal to trade on.