Beneish M-Score
An eight-variable model designed to flag earnings manipulation. Above −1.78 indicates elevated manipulation probability; we treat above −1.78 as a quality-gate failure.
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 × LVGIThe Beneish M-Score, developed by Daniel Beneish in 1999, is an eight-variable model that flags companies with elevated probability of earnings manipulation. The variables capture suspicious patterns: rising days-sales-in-receivables (revenue may be inflated), declining gross margin (pressure to manipulate), rising asset quality index (capitalizing what should be expensed), rising sales growth (a documented manipulation correlate), declining depreciation (extending useful lives), declining SG&A intensity, rising total accruals (the textbook earnings-management signal), and rising leverage. The composite scoring threshold of −1.78 historically separates manipulators from non-manipulators with reasonable specificity; the M-Score famously flagged Enron, WorldCom, and several other accounting frauds well before they collapsed. We treat an M-Score above −1.78 as a hard quality-gate failure: regardless of headline metrics, valuation, or growth, a flagged stock requires a forensic-accounting drill-down before we will recommend it. The M-Score is a probabilistic screen, not a definitive verdict — false positives are common — but the asymmetry of the underlying risk (catastrophic loss on the rare true positive) justifies the conservative posture.