Return on Equity (ROE)
Return on Equity measures how much net profit a company generates for every dollar of shareholders' equity it employs — it is the bottom-line return on the capital that belongs to owners. Defined as net income divided by shareholders' equity, ROE is the single most widely cited measure of how efficiently a management team compounds owner capital. Its core tension is that the same ratio can signal genuine business quality (a durable, high-return franchise) or financial engineering (a thin, debt-loaded equity base). The number alone does not tell you which — so ROE is only as useful as the decomposition behind it.
How it's calculated
The basic formula:
> ROE = Net Income / Shareholders' Equity
Net income is the after-tax "bottom line" attributable to common shareholders (often net income minus preferred dividends, divided by common equity). Shareholders' equity is total assets minus total liabilities — the book value of owners' capital. Analysts typically use average equity ((beginning + ending) / 2) to match a flow (a year of earnings) against a stock (balance-sheet equity at points in time), per AnalystPrep/CFA-curriculum convention.
DuPont decomposition is the essential second step. The 3-step DuPont identity factors ROE into three drivers (Wikipedia; Corporate Finance Institute; AnalystPrep):
> ROE = Net Profit Margin × Asset Turnover × Equity Multiplier > = (Net Income/Revenue) × (Revenue/Assets) × (Assets/Equity)
This isolates why ROE is what it is: operating profitability (margin), asset efficiency (turnover), and financial leverage (the equity multiplier, Assets/Equity). The 5-step DuPont splits it further into tax burden × interest burden × EBIT margin × asset turnover × leverage, separating operating performance from the effects of debt and taxes. The key insight: a company can raise ROE simply by adding leverage, with no improvement in the underlying business.
How it's used in practice
- Quality screening. A high, stable ROE achieved with little debt is the classic signature of a business with a competitive moat — pricing power or barriers to entry that let it reinvest at high rates. This is the lens most associated with Warren Buffett (see below).
- DuPont diagnosis. Comparing two firms with identical ROE, DuPont reveals whether one earns it through fat margins (e.g. luxury/software) and the other through high turnover (e.g. discount retail) or through leverage (e.g. banks). Tracking the components over time shows whether a rising ROE is real (margin/turnover) or borrowed (rising equity multiplier).
- Sustainable-growth estimation. ROE × (1 − payout ratio) = the sustainable growth rate — how fast a firm can grow earnings funding only from retained profits without new debt or equity.
- Cross-checking against ROA and ROIC. Return on Assets strips out leverage; a wide ROE-minus-ROA gap means leverage is doing the work. ROIC (sibling node) extends this to all invested capital (debt + equity) and is generally the cleaner moat metric because it is leverage-neutral.
- Benchmarking. ROE is meaningful only versus same-industry peers and the firm's own history — sector capital structures and asset intensities differ enormously.
Adoption, debate & evidence
ROE is near-universal — taught in every finance curriculum, reported by every data vendor, and central to fundamental "quality" investing. The widely repeated rule of thumb that 15–20%+ is "good" and 30%+ "exceptional" (WallStreetZen, Strike.money) is a heuristic, not a law; the right benchmark is the peer group. Buffett's 1987 shareholder letter cited a Fortune study finding only 25 of 1,000 companies averaged over 20% ROE with no single year below 15%, and that 24 of those 25 also beat the S&P 500 — an oft-quoted but anecdotal data point, not a controlled study.
The rigorous evidence sits in the academic "quality/profitability" factor literature, and here precision matters. Novy-Marx (2013) showed that gross profitability (revenue minus COGS, scaled by assets) predicts the cross-section of returns about as well as value metrics. ROE specifically is the profitability leg of the Hou-Xue-Zhang q-factor model (2015), where stocks are sorted on size × investment × ROE; the model's profitability factor carries a positive premium and the q-factor model is shown to subsume the Fama-French models in spanning tests (global-q.org; NBER w24709). Important nuance: the q-factor ROE leg is constructed from firms' most recently announced quarterly ROE, so it captures earnings momentum/innovations rather than a stable, slow-moving level (Hou-Xue-Zhang 2015) — a high static ROE is a weaker return predictor than a rising one. Net: the profitability premium is one of the better-documented anomalies, but the credit belongs to the broad factor, not to ROE as a standalone screen — and the academic ROE factor is a long-short, cross-sectional construct, not a buy-this-one-stock rule.
Strengths & limitations
Strengths. Simple, comparable, and directly tied to owner economics; the DuPont breakdown turns one number into a genuine diagnostic of a business model.
Limitations / failure modes:
- Leverage flatters it. The #1 misuse is reading high ROE as quality without checking debt. A buyback funded by debt shrinks equity and mechanically lifts ROE with zero operating improvement (dcf-model.com; Investopedia).
- Negative or near-zero equity breaks it. Mature firms that have bought back stock aggressively (e.g. some large consumer staples) can have negative book equity, making ROE meaningless or absurdly large — not a sign of distress, just a denominator artifact.
- Book-value distortions. Equity reflects historical accounting, not economic value. Write-offs, goodwill, intangibles expensed rather than capitalized, and accounting choices all warp the denominator. Asset-light firms (heavy on unrecorded intangibles) can show inflated ROE.
- Net income is gameable. One-time gains, accruals, and earnings management hit the numerator.
- Not comparable across industries. Banks, utilities, and software have structurally different ROEs.
The disciplined practice is to never trust ROE without the DuPont decomposition and a cross-read against ROA/ROIC and debt levels.
Sources
- AnalystPrep (CFA Level 1) — DuPont decomposition & average-equity convention: https://analystprep.com/cfa-level-1-exam/financial-reporting-and-analysis/dupont-analysis-return-equity/
- Wikipedia — DuPont analysis (3-step and 5-step identities): https://en.wikipedia.org/wiki/DuPont_analysis
- Corporate Finance Institute — DuPont analysis: https://corporatefinanceinstitute.com/resources/accounting/dupont-analysis/
- dcf-model.com — "Examining the Limitations of ROE" (leverage, buybacks, negative equity): https://dcf-model.com/blogs/blog/examining-limitations-roe
- WallStreetZen — "What Is a Good ROE?" (15–30% heuristics): https://www.wallstreetzen.com/blog/what-is-a-good-roe/
- Validea / Buffett ROE study (1987 letter, Fortune anecdote): https://blog.validea.com/building-a-quantitative-strategy-based-on-warren-buffetts-approach/
- Novy-Marx (2013), "The Other Side of Value: The Gross Profitability Premium," JFE — gross profitability ≈ book-to-market as a return predictor: https://mysimon.rochester.edu/novy-marx/research/OSoV.pdf
- Novy-Marx, "Quality Investing" — profitability subsumes most "quality" definitions: https://mysimon.rochester.edu/novy-marx/research/QDoVI.pdf
- Hou, Xue & Zhang q-factor model (ROE as profitability factor) — global-q.org & NBER w24709: https://www.nber.org/system/files/working_papers/w24709/w24709.pdf
Disputes flagged: the "15–20% = good" thresholds are folklore heuristics, not validated cutoffs; the Buffett Fortune figure is anecdotal; and the documented profitability premium belongs to the academic factor (and to changes in ROE more than its level), not to high-ROE single-stock screening.