5-step DuPont analysis

The 3-step's net margin still bundles three different claimants: operations, lenders, and the tax office. The 5-step separates them. Start from the same identity and multiply by 1 twice more, inserting EBT (profit before tax) and EBIT:

ROE = (NI/EBT) × (EBT/EBIT) × (EBIT/Revenue) × (Revenue/Assets) × (Assets/Equity)
       tax        interest     operating         asset              financial
       burden     burden       margin            turnover           leverage

How do you read the five factors?

EBT cancels between factors one and two. EBIT cancels between two and three. The first three factors together collapse back into net margin, and that's how you check your work. Use the top-down EBIT (EBT + interest expense) from the margin stack so the identity closes with nothing left over. Here's how to read each new factor:

• Tax burden = NI ÷ EBT is the fraction of pre-tax profit that survives the state, and it equals 1 − effective tax rate. It's around 0.75 for a mainstream Indian corporate (the ~25.17% new-regime rate incl. surcharge and cess) and 0.74 to 0.77 for a typical US corporate (21% federal plus state), both as of mid-2026, verify against the tax note. A tax burden far above these, say 0.92, means an unusually low effective rate. A tax holiday could explain it. So could a one-off credit, or losses carried forward. Ask whether it repeats before you capitalize it into "quality." The ETR reconciliation note is the audit trail.
• Interest burden = EBT ÷ EBIT is the fraction of operating profit that survives the lenders. It's 1.0 for a debt-free company, or slightly above if interest income exceeds expense, and it slides toward zero as debt service grows. At 0.62, creditors intercept 38 paise of every operating rupee before shareholders see anything. A falling interest burden over several years is leverage tightening its grip, so cross-examine it with the coverage ratios.
• Operating margin = EBIT ÷ Revenue is the engine itself, now cleanly separated from financing and tax.

Real filings add two wrinkles. The first is other income. Top-down EBIT sweeps it into the "operating" factor, and for a cash-rich company like Infosys that flatters operating margin. Accept it knowingly, or strip other income from EBIT and EBT symmetrically. The second is minorities. Put NI attributable to parent over consolidated EBT and the tax-burden factor ends up carrying the NCI slice without telling you. For most companies that's trivial. With NCI-heavy consolidators you decompose on consolidated figures and apply the parent split at the end.

Worked example: the same two companies, five ways

The same two companies, five ways (illustrative, same scale as the 3-step DuPont; check against filings). New inputs: HUL, EBIT (top-down) 14,200, interest 400, EBT 13,800, tax 3,500. Walmart, EBIT 29.5, interest (net) 2.5, EBT 27.0, tax plus minority slice 7.5.

The five-step adds one decisive fact the three-step hid. Neither company's ROE owes anything material to financing games or tax luck. HUL's burdens (0.746, 0.972) are almost the neutral Indian corporate profile, so its ROE is pure engine. Walmart's interest burden of 0.915 is a modest, comfortable creditor toll on a heavily supplier-financed model. Its route is engine times velocity, mildly amplified. You'll meet companies, probably every week, whose ROE leans on a 0.95 tax burden and a rising leverage factor while operating margin shrinks. The five-step is how you catch that quality rotting under a stable headline.

Buffett's five ways to raise ROE

Buffett got here too. His 1977 Fortune essay "How Inflation Swindles the Equity Investor" lists five ways a company can raise its ROE: higher turnover, cheaper leverage, more leverage, lower income taxes, wider operating margins. That's the 5-step DuPont in prose, written to prove that sustained ROE expansion is hard. Each lever has a ceiling, and most of them aren't operating achievements. Read it this week. It's free, and it's in the resources list.