Piotroski F-Score Stocks: Financially Improving Value Companies

The Piotroski F-Score is a 9-point integer score, developed by Stanford accounting professor Joseph Piotroski in his 2000 paper Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers. It was designed to solve a specific problem: within the universe of value stocks (cheap on price-to-book), how do you separate the companies that will recover from the ones that will continue deteriorating?

Piotroski's insight was that investors had historically been unable to distinguish between two types of cheap-on-book-value companies: (1) those that were cheap because the market overreacted to temporary bad news and the underlying business was actually improving; and (2) those that were cheap because the business was genuinely deteriorating and the low P/B reflected real asset impairment. The F-Score uses nine binary accounting signals to identify which type applies.

The 9 signals, grouped by dimension:

Profitability (F1–F4):

• F1 — ROA Positive: Return on assets is positive (company is earning on its asset base, not destroying value)
• F2 — Operating Cash Flow Positive: OCF > 0 (earnings are backed by actual cash generation)
• F3 — ROA Improving: Current year ROA > prior year ROA (profitability is trending up)
• F4 — Accruals Quality: OCF/Assets > ROA (earnings quality check — cash earnings exceed reported earnings, low accruals)

Leverage & Liquidity (F5–F6):

• F5 — Leverage Declining: Long-term debt ratio lower than prior year (balance sheet repair in progress)
• F6 — Liquidity Improving: Current ratio higher than prior year (improving short-term financial stability)
• F7 — No Dilution: No new common shares issued in the past year (management not diluting shareholders)

Operating Efficiency (F8–F9):

• F8 — Gross Margin Expanding: Current year gross margin > prior year (pricing power or cost control improving)
• F9 — Asset Turnover Improving: Revenue/Assets ratio higher than prior year (getting more revenue per dollar of assets)

Each passing signal awards exactly 1 point; failing awards 0. The maximum score is therefore 9. Piotroski's original paper showed that stocks scoring 8–9 averaged 13.4% annual returns vs. 3.4% for the general value universe — an alpha of over 10 percentage points per year. Scores of 0–2 actually predicted continued poor performance and deterioration.

The Piotroski F-Score occupies a specific niche in factor investing that is distinct from other well-known qualitative screening approaches. Understanding those differences helps you use it correctly — and avoid double-counting with other screens.

F-Score vs. Buffett Screen:

The Buffett screen requires established high-quality franchises: 5-year average ROE ≥ 15%, gross margin ≥ 30%, D/E ≤ 0.5, FCF margin ≥ 10%. These are all absolute level tests. A company must already be excellent to pass. The F-Score instead measures directional improvement: is ROA rising? Is leverage declining? Is gross margin expanding? A company coming out of a difficult period with ROA improving from 2% to 4% passes F-Score signals F1 and F3 even though its ROA is far below what the Buffett screen requires. This makes F-Score stocks fundamentally different holdings: they are often cyclical recovery candidates, asset-heavy businesses in early turnaround, or conservative businesses that hit a temporary rough patch. Buffett screen stocks are enduring franchise compounders. Both have documented alpha; they are complementary rather than redundant.

F-Score vs. GARP Screen:

GARP (Growth at a Reasonable Price) focuses on forward earnings growth relative to the P/E paid — the PEG ratio. GARP stocks combine positive revenue and earnings momentum with reasonable valuation (low PEG). F-Score stocks need not show strong revenue growth (signal F9 tests asset turnover improvement, not revenue growth per se) and often trade at low P/E and P/B simultaneously. The GARP investor pays for accelerating growth; the F-Score investor pays for financial recovery from a depressed starting point. F-Score requires value (P/B ≤ 2); GARP has no valuation anchor requirement beyond PEG.

F-Score vs. Simple Value Screen:

Sorting by lowest P/B or P/E without a quality filter is the classic value trap — you find cheap companies but cannot distinguish recovering businesses from deteriorating ones. Piotroski's paper exists precisely because buying the cheapest quintile of stocks (on book-to-market) produced disappointing returns when done naively. The F-Score is the quality filter that turns a naive P/B screen into a predictive strategy. Combining low P/B with a high F-Score has generated consistent positive alpha; using P/B alone has not within modern developed markets.

Each of the nine F-Score signals is designed to test a specific, independently verifiable dimension of financial health. Here is what each signal captures and why it matters:

F1 — Positive Return on Assets (ROA)
ROA = Net Income ÷ Average Total Assets. F1 simply requires ROA > 0: the company generates positive returns on its asset base. This screens out businesses in active financial distress where the asset base is generating losses. It is the minimum viability test for the profitability dimension.

F2 — Positive Operating Cash Flow
OCF > 0. Earnings can be positive while cash flow is negative if management uses aggressive revenue recognition, capitalizes R&D, or defers receivables. F2 requires actual cash generation. Combined with F1, it distinguishes real profitability from accounting-supported profitability.

F3 — Rising ROA
Current ROA > Prior Year ROA. The trend matters more than the level. A business recovering from a trough year (ROA improving from −4% to +2%) passes F3 even though its absolute ROA is still below average. This captures early-stage turnarounds before the market prices the recovery.

F4 — Accruals Quality
OCF ÷ Average Total Assets > ROA. This is an earnings quality test: companies earning more in cash than they report in net income have low accruals and higher earnings reliability. Companies with high accruals (reported income exceeds cash income) often disappoint in subsequent periods as accrual reversals materialize.

F5 — Declining Leverage
Long-term Debt ÷ Average Total Assets lower than prior year. Financially stressed value companies often carry high debt — deleveraging is a signal that management is using cash generation to repair the balance sheet rather than paying for operations with new borrowing. F5 rewards balance sheet recovery.

F6 — Improving Liquidity
Current Ratio higher than prior year. Current Ratio = Current Assets ÷ Current Liabilities. A rising ratio means the company is building short-term financial cushion — either growing current assets (receivables collection improving, inventory building for growth) or reducing current liabilities (paying down short-term debt). Both are positive signals.

F7 — No New Share Issuance
No new common shares sold in the past 12 months. Equity dilution at low P/B is inherently value-destructive: selling shares below book value transfers wealth from existing shareholders to new ones. Management that issues shares when the company is cheap either needs external capital urgently (a financial distress signal) or lacks alignment with shareholder interests. F7 rewards restraint.

F8 — Expanding Gross Margin
Current year gross margin > prior year gross margin. Improving gross margin signals either better pricing power (revenue outgrowing cost of goods) or improving supply chain efficiency (cost of goods declining as a percentage of revenue). Both are operational improvements. This signal detects early operating recovery before it flows through to net income after overhead.

F9 — Improving Asset Turnover
Revenue ÷ Average Total Assets higher than prior year. Asset turnover improvement means the company is extracting more revenue from each dollar of assets — a sign of improving capital discipline and operational efficiency. Alongside F8 (gross margin), F9 captures the operating efficiency dimension comprehensively: more revenue per asset dollar AND better margin on that revenue is the ideal combination.

The Piotroski F-Score has accumulated more independent academic validation than almost any other single-factor screening system. The original results were published in the Journal of Accounting Research in 2000, and subsequent independent studies have confirmed, extended, and refined the findings across multiple time periods and markets.

Piotroski's original findings (1976–1996, US data):

• High F-Score stocks (8–9) earned a mean annual return of 13.4%
• Low F-Score stocks (0–2) earned a mean annual return of 3.4%
• A long-short strategy (long high F-Score, short low F-Score) generated 7.5% annual alpha
• The strategy was strongest among small-cap value stocks (where market coverage is thinnest)
• Results were consistent across subperiods, not concentrated in one era

Independent replications:

Hyde (2003) replicated Piotroski's methodology in European markets and found similar, though slightly smaller, alpha — consistent with the view that the alpha arises from fundamental investor inattention rather than a market anomaly specific to the US. Walkshäusl (2013) extended the analysis to 49 countries and found an annual premium of 6.5% globally for a long-only high F-Score strategy within the value universe.

Why the alpha has persisted:

Several structural reasons suggest the F-Score alpha is not purely a historical artifact:

• Investor attention asymmetry: Low P/B stocks are covered by fewer analysts, meaning fewer investors synthesize the nine-signal composite. The market underreacts to improving fundamentals in neglected value stocks.
• Behavioral anchoring: Once a stock is labeled a "distressed value" name, investors anchor to that classification even as fundamentals improve. The F-Score scores the improvement; market prices lag.
• Short-side constraint: Low F-Score stocks are hard to short efficiently (high borrow costs for illiquid small-caps), so the long-short theoretical alpha is never fully arbitraged away.
• Complexity barrier: Most retail investors do not calculate all nine signals from financial statements. This keeps the strategy's edge alive even in an era of highly efficient large-cap markets.

Limitations: The F-Score has shown reduced alpha among mega-cap stocks (where analyst coverage is exhaustive and mispricing is rare) and in pure growth companies (where the P/B anchor is absent). Its strongest results remain concentrated in mid-cap and small-cap value names — exactly the universe this screen targets with its $500M minimum cap and P/B ≤ 2 filter.

A high F-Score alone is not a buy signal — it is a strong filter that narrows the universe to stocks where the financial evidence supports further investigation. Here is a practical framework for turning a screen output into a researched investment decision:

Step 1 — Confirm the score is not a one-year artifact
F-Score signals are calculated on annual financial data. If a company posted an exceptional one-time revenue year (e.g., a government contract, asset sale, or insurance recovery), multiple signals may have triggered without representing ongoing improvement. Review 2–3 years of the key ratios — ROA trend, leverage trend, gross margin trend — to verify the improvement is real and sustained.

Step 2 — Understand why the stock is cheap (low P/B)
Low P/B can arise from three sources: (a) genuine temporary distress — the reason F-Score alpha exists; (b) structural industry decline — where cheapness is rational; or (c) accounting book value inflation — where stated assets overstate true asset value. Cases (b) and (c) are value traps even with a high F-Score. Research the industry context: is the company in a cyclical trough (good), a structurally declining market (bad), or carrying goodwill from an overpriced past acquisition (potentially bad)?

Step 3 — Size the position relative to score and margin of safety
A score of 9/9 with P/B of 0.8 is a different risk profile than a score of 8/9 with P/B of 1.9. The deeper the value discount (lower P/B) and the higher the F-Score, the more favorable the risk/reward. Some practitioners use F-Score as a binary gate (≥ 8 = eligible for full sizing) and P/B as a continuous position-size variable (lower P/B → larger position).

Step 4 — Monitor for F-Score degradation
The F-Score is recalculated annually when full-year financial data becomes available. A company that scored 9 last year may drop to 5 this year if the improvement was temporary. Regular rescreening (quarterly, after 10-K filing) catches deterioration early. The academic evidence suggests most F-Score alpha materializes in the 12 months following the score calculation — holding beyond 12–18 months without rescoring introduces risk that the financial improvement has stalled.

What to look for in the column data:

• F-Score col 3: 9 is strongest; 8 is valid; use as a ranking, not just a binary cutoff
• P/B col 4: Lower is deeper value; 0.5–1.0 is compelling; 1.5–2.0 is moderate value anchor
• D/E col 6: Confirms F5 signal direction; declining D/E is the balance sheet repair signal
• FCF Margin col 7: Confirms F2 (positive OCF) and validates earnings quality
• Rev G TTM col 8: Not a direct F-Score signal but confirms operational momentum behind F9 (asset turnover improvement)
• Total Ret 1Y col 10: Low or negative = market has not priced in the F-Score improvement yet; positive = market is beginning to recognize recovery (still valid if early stages)

The Piotroski F-Score is not equally powerful across all sectors and market cap tiers. Understanding where the academic evidence is strongest — and where it is weaker — helps you interpret screen results with appropriate context.

Where F-Score works best:

• Small and mid-cap industrials and manufacturers: These companies often see cyclical ROA swings, leverage buildups in expansions, and asset turnover improvements during recoveries. The nine signals map cleanly onto real operating cycles. Analyst coverage is limited enough that mispricing persists.
• Cyclical materials and basic resources: Commodity-driven companies have highly variable ROA, leverage, and gross margins. A high F-Score in this group identifies companies catching a commodity tailwind while also improving their financial structure — a powerful combination.
• Small-cap financials (banks and insurance): With large balance sheets and leverage inherent to the business model, F-Score signals on leverage and liquidity translate differently than for operating companies. Use with caution; the current ratio (F6) is often not applicable for banks. Some practitioners exclude financials from F-Score analysis entirely.

Where F-Score has reduced predictive power:

• Large-cap technology: These companies rarely trade at P/B ≤ 2 and have exhaustive analyst coverage. The value anchor (low P/B) is almost never met, so the screen rarely surfaces them in any case.
• Capital-light software businesses: With minimal tangible assets, book value is often understated, making P/B ratios unreliable value signals. Asset turnover (F9) is also a weak signal when assets are primarily intangible.
• REITs and MLPs: These high-distribution entities are structured to distribute cash flow rather than accumulating retained earnings, making ROA trends and leverage signals mechanically different from operating C-corporations. GAAP net income for REITs is deliberately depressed by depreciation that does not reflect economic reality.

Market cap tier analysis:

Piotroski's original paper identified the strongest alpha in the smallest size quartile of his value universe. This makes mechanistic sense: smaller companies have fewer analysts, lower institutional ownership, and less frequent information updates — all factors that let financial improvement go unpriced for longer. This screen's $500M minimum cap keeps you in the zone where coverage gaps still exist while filtering out the micro-cap liquidity problems that make academic backtests hard to replicate in practice.

The $500M–$3B range is arguably the sweet spot: large enough for reasonable liquidity and financial reporting quality; small enough that not every hedge fund has already modeled the nine signals and priced in the improvement immediately after earnings release. F-Score alpha has been most reliably replicated in this mid-range in live trading studies.