DuPont Analysis

One ratio, three stories

Return on Equity — net income divided by shareholders' equity — is the single number most often cited when asking "is this a good business?" A 25% ROE sounds uniformly excellent. A 10% ROE sounds uniformly mediocre.

Both conclusions are wrong, because the same ROE can come from completely different places. DuPont analysis — named after the 1920s DuPont Corporation's financial team that formalized the decomposition — is the tool that breaks ROE into its underlying components so you can see which lever is actually doing the work.

The three-step DuPont

Where NI = net income, S = sales (revenue), A = total assets, E = shareholders' equity.

Each term has a clear business meaning:

• Net margin — "how much of each dollar of revenue ends up as profit after every expense?" High = pricing power, brand, scale efficiency.
• Asset turnover — "how much revenue do we generate per dollar of assets we operate?" High = capital-light business; low = capital-intensive.
• Leverage — "how much of those assets do the shareholders actually own vs creditors?" High = debt-funded; low = equity-funded.

The identity is exact. Multiply the three and you get ROE every time. The interesting thing is that two companies with identical 25% ROE can reach it via wildly different combinations — and the path matters enormously for how safe and how sustainable that return is.

Play with it

Toggle through tech / grocery / bank. Notice:

• Tech (Apple-like): 30% margin × 0.67× turnover × 1.25× leverage = ~25% ROE. Margin-driven.
• Grocery (Walmart-like): 2% margin × 2.0× turnover × 2.5× leverage = ~10% ROE. Turnover + moderate leverage; thin profits but fast asset cycling.
• Large bank: 10% margin × 0.08× turnover × 12× leverage ≈ 10% ROE. Leverage-driven. Banks rent their balance sheet; high leverage is how they earn acceptable returns on the spread between what they pay depositors and what they earn on loans.

The ROEs end up in the same ballpark. The businesses underneath are nothing alike. A 25% ROE from margins is far more robust than a 25% ROE achieved through 8× leverage — the latter blows up in a tight credit environment.

Why the components matter

High-margin, low-turnover businesses

Software, pharma, luxury goods, high-end consumer brands. The margin story: you own something (IP, brand, technology) that lets you charge more than cost to deliver. These are the businesses Buffett loves — "moats." Apple can maintain 30% net margins across product generations because customers continue to pay premium prices.

Risks: margin compression. If a competitor erodes the pricing differential, the whole ROE equation collapses — turnover and leverage aren't there to compensate.

High-turnover, low-margin businesses

Grocery, distribution, wholesale. The turnover story: tiny profit per item, but you sell so much so fast that total return on capital is acceptable. Walmart, Costco, Amazon retail.

Risks: any disruption to volume or supply chain hits immediately — there's no margin cushion. Grocery is classically a 2% margin business; a 1% input-cost shock wipes out half the profitability.

High-leverage businesses

Banks, insurance, REITs, some private equity. The leverage story: take a small spread on borrowed money and apply it to a much larger asset base.

Risks: credit events. A bank that earns 10% ROE on 12× leverage goes from healthy to insolvent when 1/12 of its loan book goes bad — because that 8.3% loss eats 100% of equity. Leverage makes good times look spectacular and bad times terminal.

The five-step DuPont — for deeper diagnosis

Analysts often extend the three-step into a five-step decomposition that separates out tax and interest effects:

Where EBT = earnings before tax, EBIT = earnings before interest and tax.

This lets you answer finer-grained questions:

• Tax burden rising? — new jurisdiction, expired credits, higher statutory rates
• Interest burden rising? — more debt or higher rates eating into operating income
• Operating margin falling while net margin holds? — the company is masking operational weakness through one-time tax wins
• Turnover falling? — investment in capacity that hasn't generated revenue yet (either a good growth story or an efficiency problem)

When a company's ROE changes year-over-year, five-step DuPont tells you which lever moved. That's more actionable than the headline number.

The DuPont mindset for screens

When you pull up a company with a great-looking ROE:

1. Compute the three-step. What mix got them there?
2. Compare to industry average. 25% ROE for a software company is normal; 25% for a grocery chain is exceptional; 25% for a bank means dangerous leverage.
3. Trend the components. Margins compressing? Turnover falling? Leverage rising? Each has a different cause and a different fix.
4. Stress-test leverage. What happens to ROE if interest rates rise 200 bps? If a 5% loan loss hits the balance sheet? If margins compress 1%?

This is why serious analysts dismiss "good ROE" as a screen. It's only a starting question — DuPont gives you the follow-up questions to understand the quality.

A concrete diagnostic

Consider three companies with 18% ROE:

• Company X is a high-margin software business — stable pricing power drives ROE
• Company Y is a grocery chain — razor margins, fast turnover, modest leverage
• Company Z looks like an industrial that's leaning on leverage to hit the same number

If you had to pick one to own at 18% ROE through a recession, the correct answer is almost always X — the margin business. Y depends on volume that contracts in a recession; Z is depending on leverage that gets more expensive and riskier precisely when you want less of it.

Quick check

What you now know

• ROE = Net Margin × Asset Turnover × Leverage — the three-step DuPont identity
• The same ROE can come from margin-driven (Apple), turnover-driven (Walmart), or leverage-driven (banks) business models — and they're not equally safe
• Five-step DuPont splits further into tax burden, interest burden, operating margin, turnover, and leverage — for diagnosing which component moved when ROE changes year-over-year
• Leverage-driven ROE is fragile — a small asset-side shock wipes out magnified-equity exposure
• When an ROE improves, always ask which lever moved, and whether it's sustainable or one-time

Next: Position Sizing — applying Kaufman's fixed-fractional and Kelly frameworks to turn an edge into a dollar amount of risk per trade.