A-F grading system

How We Grade Earnings Quality

Every stock in our coverage receives a letter grade from A to F based on 18 forensic accounting checks and three quantitative models. This page explains exactly how grades are assigned, what the thresholds are, and why some grades deserve more context than others.

The Grading Scale

Our grades reflect the number of failed checks, the severity of watch items, and the output of three quantitative models: Beneish M-Score, Altman Z-Score, and Dechow F-Score. Here is what each grade means:

APristine financial statements

0 fails, 0-1 watches, M-Score safe (< -2.22). The company passes every check cleanly. Cash flow backs up reported earnings, receivables track revenue, margins are stable, and no quantitative model flags concern. This is the cleanest bill of health our system can give.

BGenerally healthy with minor items to monitor

0 fails, 2-3 watches. No check has crossed a hard threshold, but a few metrics are moving in a direction worth watching. For example, days sales outstanding may be rising modestly, or SG&A is ticking up relative to revenue. Nothing actionable yet, but the trend bears monitoring in the next filing.

CThe most analytically interesting grade

1 fail or 4+ watches. One structural issue has crossed a threshold. This is often the grade that rewards careful reading. A single fail might be a temporary blip from an acquisition, a one-time restructuring charge, or the early sign of a real problem. Our report explains which it is.

DMultiple concerns compounding

2 fails or M-Score in the grey zone (-2.22 to -1.78). Two independent checks have crossed hard thresholds, or the Beneish M-Score has entered inconclusive territory. When two different forensic angles both flag concern simultaneously, the probability that something is genuinely wrong increases significantly.

FMajor red flags

3+ fails, a critical failure on any single check, or M-Score > -1.78. The financial statements show patterns that historically correlate with earnings manipulation, financial distress, or aggressive accounting. However, many F grades are structural rather than dangerous — see below.

What Counts as a Fail vs. a Watch

Each of our 18 checks produces one of four results: Pass, Watch, Fail, or N/A. The distinction between Watch and Fail matters because it determines how the check flows into the overall grade.

Fail

A check crosses a hard, quantitative threshold. These thresholds are calibrated to flag conditions that, in academic literature and historical cases, have preceded earnings manipulation or financial distress. Examples:

  • *Accounts receivable growth exceeds revenue growth for 2+ consecutive years, suggesting revenue may be recognized before cash is collected.
  • *Free cash flow is negative for 3+ years while the company reports GAAP profits, indicating a persistent gap between paper earnings and real cash generation.
  • *Goodwill exceeds 40% of total assets with no impairment testing, creating latent write-down risk.
  • *Beneish M-Score breaches -1.78, matching patterns seen in historically confirmed manipulators.

Watch

A check shows concerning movement but has not yet crossed the hard threshold. Watches are early-warning signals. Examples:

  • *Days sales outstanding is rising but still within a single year of revenue — not yet a fail, but the trend warrants attention.
  • *Gross margin compressed in a single quarter but remains above the 3-year average.
  • *Capital expenditure spiked in one period without a clear explanation in management commentary.
  • *M-Score sits between -2.50 and -2.22 — safe, but closer to the grey zone than typical.

N/A

A check does not apply to this company's business model. We exclude inapplicable checks rather than forcing them through a framework that would produce misleading results. Examples:

  • *Inventory checks for pure software companies that carry no physical inventory.
  • *Traditional cash flow from operations checks for banks and financial institutions, whose cash flow statements follow fundamentally different structures.
  • *Depreciation analysis for asset-light businesses with negligible PP&E.

Structural F Grades

Not every F grade means danger. Many F grades are structural — they reflect the normal financial characteristics of certain business models rather than manipulation or distress.

REITs

Real estate investment trusts are required by law to distribute 90%+ of taxable income as dividends, resulting in high leverage ratios that trigger our debt checks. This is by design, not a warning sign.

Utilities

Regulated utilities carry high debt-to-EBITDA ratios to finance long-lived infrastructure assets. Their revenue is regulated, making the debt far more predictable than it appears in isolation.

Buyback-heavy companies

Companies like McDonald's and Starbucks have repurchased so many shares that they report negative book equity. Our equity-based checks flag this, but negative equity from buybacks is a capital allocation choice, not a solvency concern.

Serial acquirers

Companies that grow through acquisitions naturally accumulate goodwill. A high goodwill-to-assets ratio triggers our checks, but whether it represents real risk depends on the quality of the acquired businesses and management's track record.

We mark structural F grades in our reports and explain the sector context. An override system exists for cases where automated checks produce misleading results — every override is documented with a rationale in the report.

Why Grade Alone Isn't Enough

  • *A Grade A stock can still be overvalued. Our grading system measures earnings quality, not whether the stock is a good buy. A company with pristine financials trading at 80x earnings may still be a poor investment.
  • *A Grade F stock can be a perfectly safe investment if the F is structural. A REIT with an F grade due to leverage may be one of the most stable income-producing investments in the market.
  • *A Grade C stock is often the most interesting. One failed check might be a temporary anomaly from an acquisition or restructuring, or it might be the first crack in an accounting facade. Only the report tells you which.
  • *Grades are backward-looking. They are based on the most recent 10-K filing and up to 5 years of historical data. They cannot predict future events, management changes, or market shifts.
  • *Always read the report, not just the grade. Our reports explain every check result with company-specific context, management commentary from the 10-K, and sector benchmarks. The grade is the summary; the report is the analysis.

Learn More

Example Reports

See our grading system in action:

Back to MethodologyBack to Home
How We Grade Earnings Quality — EarningsGrade