How we grade earnings quality

Methodology

EarningsGrade screens US stocks using forensic accounting techniques. Every grade is backed by 18 quantitative checks, three academic scoring models, and 5 years of financial data from SEC filings.

Grading System

Each stock receives an A through F grade based on how many checks it fails and how severe the failures are.

A
Clean
0 fails, 0-1 watch items, M-Score clean
Pristine financial statements. No manipulation signals, strong cash flow, clean balance sheet.
B
Minor Concerns
0 fails, 2-3 watch items
Generally healthy. Watch items are typically explainable as strategic choices, not red flags.
C
Investigate
1 fail or 4+ watch items
Something needs attention. Could be structural (acquisition goodwill) or genuinely concerning (AR divergence).
D
Significant Issues
2 fails or M-Score grey zone
Multiple concerning signals. The financial statements require careful reading before any investment decision.
F
Major Red Flags
3+ fails, or critical fail, or M-Score triggered
Serious concerns. May indicate manipulation, financial distress, or extreme leverage. Does NOT mean the stock will go down.

18-Point Screening

Six categories, each targeting a different type of financial shenanigan (per Schilit's framework).

A. Revenue Quality

A1DSO Trend

Days Sales Outstanding change year-over-year. Rising DSO suggests loosening credit terms or channel stuffing.

A2AR vs Revenue

Accounts Receivable growth vs Revenue growth. AR growing faster than revenue for 2+ years is a classic manipulation signal.

A3Revenue vs CFFO

Revenue growth vs Cash Flow from Operations growth. Revenue growing faster than cash flow suggests non-cash revenue recognition tricks.

B. Expense Quality

B1Inventory vs COGS

Inventory growth vs Cost of Goods Sold growth. Inventory piling up faster than sales suggests obsolescence or demand problems.

B2CapEx vs Revenue

Capital Expenditure growth rate vs Revenue growth rate. CapEx growing much faster may signal over-investment or empire building.

B3SG&A Ratio

SG&A as percentage of Gross Profit. Rising ratios suggest deteriorating operating leverage.

B4Gross Margin

Gross Margin trend over 3+ years. Sustained decline indicates pricing pressure or cost inflation not passed through.

C. Cash Flow Quality

C1CFFO vs Net Income

Cash Flow from Operations / Net Income ratio. Below 1.0 for multiple years means reported profits aren't turning into cash.

C2Free Cash Flow

FCF = CFFO - CapEx. Negative FCF for 3+ years while reporting profits is a serious red flag.

C3Accruals Ratio

(Net Income - CFFO) / Total Assets. High positive accruals mean profits are paper-based, not cash-based.

C4Cash vs Debt

Cash & equivalents / Total Debt. Below 30% means limited liquidity buffer.

D. Balance Sheet Quality

D1Goodwill / Equity

Goodwill + Intangibles as % of Shareholder Equity. Above 40% means the balance sheet is built on acquisition promises, not hard assets.

D2Leverage

Debt/EBITDA ratio. Above 4x is stress territory. Above 6x is distress.

D3Soft Assets

Other assets growth significantly exceeding revenue growth. May indicate off-balance-sheet games.

D4Impairment

Significant goodwill or asset impairments in recent years.

E. Acquisition Quality

E1Acquisition FCF

Free Cash Flow after acquisitions. Negative for 3+ years means the company is a serial acquirer funding deals with debt, not earnings.

E2Goodwill Surge

Year-over-year goodwill growth. A sudden surge indicates a major acquisition that may be overpriced.

F. Manipulation Detection

F1Beneish M-Score

8-variable model. Below -2.22 = safe. Between -2.22 and -1.78 = grey zone. Above -1.78 = likely manipulator. 76% historical accuracy.

Scoring Models

Beneish M-Score

Developed by Professor Messod Beneish (1999). An 8-variable model that detects earnings manipulation by measuring distortions in financial statements.

M = -4.84 + 0.920×DSRI + 0.528×GMI + 0.404×AQI + 0.892×SGI + 0.115×DEPI - 0.172×SGAI + 4.679×TATA - 0.327×LVGI
Variables
  • DSRI — Days Sales in Receivables Index
  • GMI — Gross Margin Index
  • AQI — Asset Quality Index
  • SGI — Sales Growth Index
  • DEPI — Depreciation Index
  • SGAI — SG&A Index
  • TATA — Total Accruals to Total Assets
  • LVGI — Leverage Index
Thresholds

Below -2.22: Safe | -2.22 to -1.78: Grey zone | Above -1.78: Likely manipulator

Beneish, M.D. (1999). "The Detection of Earnings Manipulation." Financial Analysts Journal.

Altman Z''-Score

Non-manufacturing variant of the Altman Z-Score. Predicts financial distress within 2 years. Not applicable to financial companies (banks, insurance, REITs).

Z'' = 6.56×X1 + 3.26×X2 + 6.72×X3 + 1.05×X4
Variables
  • X1 — Working Capital / Total Assets
  • X2 — Retained Earnings / Total Assets
  • X3 — EBIT / Total Assets
  • X4 — Book Value of Equity / Total Liabilities
Thresholds

Above 2.60: Safe | 1.10 to 2.60: Grey zone | Below 1.10: Distress

Altman, E.I. (1968, revised 1993). "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy."

Dechow F-Score

Developed by Patricia Dechow (2011). A 7-variable predicted probability of financial misstatement. Used as a parallel indicator — does not affect the A-F grade.

PV = -7.893 + 0.790×rsst_acc + 2.518×ch_rec + 1.191×ch_inv + 1.979×soft_assets + 0.171×ch_cs - 0.932×ch_roa + 1.029×issue
Variables
  • rsst_acc — Richardson accruals (WC + NCO + FIN changes)
  • ch_rec — Change in receivables / avg assets
  • ch_inv — Change in inventory / avg assets
  • soft_assets — (Total Assets - PP&E - Cash) / Total Assets
  • ch_cs — Change in cash sales
  • ch_roa — Change in return on assets
  • issue — Whether securities were issued
Thresholds

Predicted probability > 50%: High misstatement risk

Dechow, P.M. et al. (2011). "Predicting Material Accounting Misstatements." Contemporary Accounting Research.

Data Sources

SEC EDGAR
10-K annual reports (the primary source). Every number in our screening comes from audited financial statements filed with the SEC.
Yahoo Finance
Historical financial data via yfinance. Cross-referenced with SEC filings for accuracy.
Hand-written Reports
Our deep analysis reports are written by reading the actual 10-K, not by feeding data into templates. Each report identifies company-specific risks that automated screening can't catch.

Important Disclaimers

  • *This is NOT investment advice. A failing grade does not mean sell. A passing grade does not mean buy.
  • *The screening framework has known limitations. Some business models (utilities, REITs, financial companies, asset-light SaaS) may trigger structural flags that are features, not bugs.
  • *Grades are based on the most recent annual filing. Financial conditions can change rapidly between filings.
  • *The M-Score has a 76% historical accuracy rate. It can produce both false positives and false negatives.
  • *Past financial health does not predict future performance.
Methodology — EarningsGrade