Profitability Ratios: Margins, ROE, ROIC

• Difficulty: Intermediate
• Published: 2025-12-28

The practical point: a 24% ROE can be "great" in 1 spreadsheet and "fragile" in 1 recession, so you treat profitability ratios as diagnostics with thresholds (e.g., >40% gross margin or >20% ROIC) rather than as "1-number trophies."

Why Profitability Ratios Matter

Profitability ratios matter because they predict returns in published datasets with specific spreads, including 0.31% per month (~3.7% annually) between high and low gross profitability portfolios in 1 large-sample study.¹ Profitability also matters because factor evidence shows ~3.1% annual alpha spread between high and low profitability portfolios when profitability is explicitly modeled.² Profitability matters even more because abnormal operating profitability mean-reverts at ~21% per year with a ~3-year half-life, so your "great year" needs a multi-year durability test.³

The Margin Stack (3 Layers, 1 Logic)

You interpret margins as 3 sequential filters on $1.00 of revenue: gross margin keeps the first $0.XX, operating margin keeps the second $0.XX after overhead, and net margin keeps the final $0.XX after everything (including taxes and financing).

Gross Margin (Pricing Power vs. Cost Structure)

Definition (1 line): gross margin is (Revenue - COGS) / Revenue, so 40% means $0.40 gross profit per $1.00 revenue.

Thresholds (3 bands):

• Strong: >40% suggests pricing power or structural cost advantage (1 moat hypothesis).
• Adequate: 25–40% is "competitive" in many differentiated industries (2 outcomes: win on brand or on process).
• Weak: <25% signals a commodity-like model that must win via volume/turnover or 2–3% net margin discipline.

Trend rule (1 trigger): a decline of >200 bps over 3 years is a "why now?" investigation even if the level is still >25%, because 200 bps is $0.02 per $1.00 revenue lost to either price or cost.

Historical anchor (2015–2023): Apple's consolidated gross margin improved from 40.1% (2015) to 44.1% (2023) as Services (with 71% gross margin) grew from 8.5% to 22% of revenue.⁴ That mix shift coincided with ROIC rising from 31% to 56% and a stock price gain of 847% versus 158% for the S&P 500 in the same 2015–2023 window.⁴

Operating Margin (Business Model After Overhead)

Definition (1 line): operating margin is Operating Income / Revenue, and you treat it as "repeatable" only after excluding 1-time items (e.g., restructurings).

Thresholds (3 bands):

• Strong: >20% indicates operational excellence and scale effects that show up as $0.20+ per $1.00 revenue after overhead.
• Adequate: 10–20% is "viable" but still exposed to a 100–300 bps shock in pricing or cost.
• Weak: <10% leaves a thin buffer where a 200 bps hit can cut operating profit by ~20–40% (e.g., from 10% to 8%).

Coverage rule (1 safety floor): operating profit should cover interest expense by ~4x for investment-grade-like resilience, because 4x gives you room for a 50% earnings drop and still 2x coverage.

Historical anchor (2016–2018): GE's industrial operating margin fell from 14.7% (2016) to 10.2% (2018) while ROIC collapsed from 10.8% to 2.1%, and the company still paid $8.2B of dividends annually during negative free cash flow.⁵ The market outcome over 2016–2018 was a market-cap decline from $280B to $63B (a 77% destruction) plus a dividend cut of 92% (from $0.96 to $0.04 per share).⁵

Net Margin (What Equity Actually Owns)

Definition (1 line): net margin is Net Income / Revenue, so 7% means $0.07 per $1.00 revenue reaches equity holders after taxes and financing.

Interpretation rule (1 caution): net margin is the most "complete" ratio, but it is also the most distorted by 1 period's tax rate or leverage, so you still reconcile it back to gross and operating layers.

Historical anchor (2010–2023): Costco sustained a 2.0–2.6% net margin range while running inventory turnover of 11.5x versus Walmart's 8.2x, and it still averaged 24.3% ROE despite net margins being roughly 60% below an industry average of 5.2%.⁶ The shareholder outcome over 2010–2023 was a 17.2% stock price CAGR, supported by a membership renewal rate >90% and membership fee income of $4.6B (2023).⁶

ROE (Return on Equity): Great Number, Dangerous Shortcut

Definition (1 line): ROE is Net Income / Average Equity, so 24% ROE means $0.24 earned per $1.00 equity.

The 3-Part ROE Reality (DuPont in 3 Numbers)

ROE decomposes into net margin x asset turnover x equity multiplier, so 24% can come from 12% x 0.8 x 2.5 or from 7% x 1.0 x 3.0, and those 2 ROEs do not carry the same risk over 1 downturn. The point is that 1 headline ROE can hide 1 leverage bet, and market participants have underreacted to these components in measured coefficients (e.g., 0.089 for turnover changes and 0.156 for margin changes, with t=4.21 and t=6.87 respectively).⁷

ROE Quality Thresholds (4 Tests)

Use 4 quantified ROE checks for non-financials:

1. Target range: 15–25% ROE is often "good," while >25% triggers a durability audit across 5+ years.
2. Leverage ceiling: equity multiplier <3.0x reduces the odds that "great ROE" is mostly debt.
3. Economic composition: >60% of ROE should come from net margin + asset turnover rather than from multiplier inflation.
4. Stability: a 5-year ROE coefficient of variation <0.25 signals franchise-like consistency rather than 1-cycle luck.

ROIC (Return on Invested Capital): The Value-Creation Ratio

Definition (1 line): ROIC is NOPAT / (Debt + Equity - Excess Cash), and 20% ROIC means $0.20 after-tax operating profit per $1.00 invested.

ROIC Thresholds That Map to Value

You treat ROIC as value-creating only relative to a cost of capital (WACC) with explicit spreads:

• Exceptional: >20% ROIC sustained for 5+ years suggests a wide moat hypothesis with multi-year evidence.
• Good: ROIC exceeds WACC by ~5 percentage points, because 5% spread is a meaningful "economic profit" wedge.
• Acceptable: ROIC exceeds WACC by 2–5 percentage points, which is positive but not moat-level.
• Value-destructive: ROIC below WACC is a capital-allocation red flag with 1 management credibility implication.

A practical reinvestment gate is numeric: approve growth capex only if incremental ROIC is projected to beat the hurdle rate by ~3 percentage points, because 3% is large enough to survive estimation error.

Evidence anchor (10+ years): companies sustaining >20% ROIC for 10+ years showed 14.2% median annual TSR versus 8.1% for companies with 8–12% ROIC, and valuation frameworks estimate a 1 pp ROIC improvement can have 2–3x the valuation impact of a 1 pp growth improvement in mature firms.⁸ A second durability anchor links "ROIC > WACC by 5+ points for 10 consecutive years" to identifiable advantages in 87% of studied cases.⁹

Worked Example (You Analyze 2 Companies in 6 Steps)

You compare 2 industrial equipment manufacturers with a 5-year horizon: Company A (premium) versus Company B (cost leader).

1. You calculate margin structure: Company A shows 42% gross, 18% operating, 12% net; Company B shows 28% gross, 11% operating, 7% net, so A's gross advantage is 14 percentage points.
2. You measure 5-year gross margin trend: A expands from 38% to 42% (+400 bps), while B contracts from 31% to 28% (-300 bps), so you treat B's trend as a 300 bps competitive warning.
3. You decompose ROE: A's ROE is 24% = 12% x 0.8 x 2.5; B's ROE is 21% = 7% x 1.0 x 3.0, so B's "near-ROE" relies on 3.0x leverage versus 2.5x.
4. You compute ROIC vs WACC: A's ROIC is 19% against 9% WACC (+10% spread), while B's ROIC is 11% against 9% WACC (+2% spread), so A creates 5x the economic spread per dollar (10% / 2% = 5x).
5. You test stability: A holds ROIC >15% for 8 consecutive years with CV 0.12, while B ranges 8–14% with CV 0.31, so A clears a 0.25 stability yardstick and B fails it.
6. You size the decision: you give A a full weight with an expected 12–14% annual return in the baseline (77–93% over 5 years), you upgrade to 16–18% annual if gross margin reaches 45% and ROIC reaches 22% (110–129% over 5 years), and you downgrade to 6–8% annual if gross margin falls to 38% and ROIC to 14% (34–47% over 5 years) while using 15% operating margin as a numeric "reduce" trigger.

Common Implementation Mistakes (You Pay in % When You Skip the Math)

1. You trust ROE without DuPont: when you buy "high ROE" names with equity multiplier >4x, you risk 4.2% annual underperformance versus filtering for margin-driven ROE, and leverage-dependent ROE has been 2.3x more volatile in recessions.⁷
2. You do 1-year margin snapshots: when you ignore 3+ consecutive years of gross margin compression, the subsequent underperformance has been 8.7% over the next 3 years, and 67% of those firms deteriorate further.³
3. You compare ROIC across industries raw: when you compare a software 35% average ROIC to an industrial 12% average ROIC without spreads, 73% of analysts in one survey mis-rank value creation, and you can prefer a 5% spread case (12% ROIC vs 7% WACC) over a 2% spread case (20% vs 18%) despite the lower absolute ROIC.⁸

Implementation Checklist (Tiered by ROI)

Tier 1 (30 minutes, highest ROI):

• Compute 5 ratios for 5 years: gross margin, operating margin, net margin, ROE, ROIC, then flag any >100 bps annual margin compression for 2+ years.
• Add 3 thresholds: gross >40%, operating >20%, ROE 15–25%, then annotate every exception with 1 driver (price, mix, cost, leverage).

Tier 2 (2 hours, high ROI):

• Build DuPont with 3 components and reject non-financials with equity multiplier >3.0x unless you have a sector reason plus 2 stress cases.
• Compute ROIC - WACC spread and require ~5 pp for "good," 2–5 pp for "acceptable," and <0 pp for "value-destructive."

Tier 3 (1 day, compounding ROI):

• Normalize 3 categories of distortions: 1-time items, lease capitalization for pre-IFRS 16/ASC 842 comparability, and "tangible" ROIC excluding goodwill for acquisition-heavy firms.
• Add a stability screen: 5-year ROE CV <0.25 and ROIC CV near 0.12–0.25, then document 2 moat mechanisms (switching costs, scale, cost advantage).

The durable lesson

The durable lesson: you are not buying 1 ratio, you are buying a 5–10 year spread, and the numbers that survive that horizon are the ones that (a) clear threshold bands like >40% gross margin or >20% ROIC, (b) keep ROE quality with leverage <3.0x, and (c) resist mean reversion that runs at ~21% per year in operating profitability.³

---

Footnotes

1. Novy-Marx, R. (2013). The other side of value: The gross profitability premium. Journal of Financial Economics, 108(1), 1–28. https://doi.org/10.1016/j.jfineco.2013.01.003 ↩

2. Fama, E. F., & French, K. R. (2015). A five-factor model explaining stock returns. Journal of Financial Economics, 116(1), 1–22. https://doi.org/10.1016/j.jfineco.2014.10.010 ↩

3. Nissim, D., & Penman, S. H. (2001). Ratio analysis and equity valuation: From research to practice. Review of Accounting Studies, 6(1), 109–154. https://doi.org/10.1023/A:1011338221623 ↩ ↩² ↩³

4. Apple Inc. Form 10-K filings (2015–2023), SEC EDGAR (services mix, margins, ROIC, and period returns as summarized in compiled research notes). ↩ ↩²

5. General Electric Form 10-K filings (2016–2018), SEC EDGAR; period outcomes summarized in compiled research notes (operating margin, ROIC, dividends, market cap, dividend per share). ↩ ↩²

6. Costco Wholesale Corporation Form 10-K filings (2010–2023), SEC EDGAR (net margin range, turnover, ROE, renewal rate, membership fees, and period returns as summarized in compiled research notes). ↩ ↩²

7. Soliman, M. T. (2008). The use of DuPont analysis by market participants. The Accounting Review, 83(3), 823–853. https://doi.org/10.2308/accr.2008.83.3.823 ↩ ↩²

8. Koller, T., Goedhart, M., & Wessels, D. (2020). Valuation: Measuring and Managing the Value of Companies (7th ed.). McKinsey & Company. https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/valuation-measuring-and-managing-the-value-of-companies ↩ ↩²

9. Greenwald, B., Kahn, J., Sonkin, P., & van Biema, M. (2001). Value Investing: From Graham to Buffett and Beyond. Wiley. https://www.wiley.com/en-us/Value+Investing ↩