DuPont: taking ROE apart
Two companies can earn the same return on equity in completely different ways. The DuPont formula splits ROE into margin, turnover and debt, so you can see which one is doing the work.
Three ratios that multiply into one
The formula takes its name from the DuPont chemical company, where a finance man called Donaldson Brown worked it out more than a hundred years ago. He later took it with him to General Motors. The trick is plain algebra: write ROE as a chain of fractions whose middle terms cancel.
Net margin is how much of each sale ends up as profit. Asset turnover is how hard the assets work: $2 of sales for every $1 of assets is a turnover of 2. The equity multiplier shows how much of the balance sheet someone other than the shareholders paid for. A multiplier of 3 means shareholders put in one dollar in every three.
Same ROE, opposite businesses
Two illustrative companies, each earning an 18% return on equity.
Halvard earns its return on margin. It sells something customers pay well for, and it barely borrows. Corrin makes almost nothing on each sale, turns its stock over fast and funds two-thirds of its assets with other people's money, mostly supplier credit and debt. Both models work.
They don't take trouble the same way, though. Knock one percentage point off each net margin. Halvard's ROE slips from 18% to about 17.3%. Corrin's halves, to 9%. A grocer lives on thin margins, so small slips in pricing or costs hit its ROE hard, and the high multiplier makes them hit harder.
The five-step version
The three-step formula lumps interest and tax into net margin. The five-step version pulls them out, which shows debt working on both sides: it raises the equity multiplier and takes interest out of profit.
Debt lifts Corrin's ROE through the 3.0× multiplier and takes a fifth of its operating profit through interest. While the business earns more on its assets than the debt costs, the trade adds to ROE. When that flips, it subtracts.
Finding out why ROE moved
DuPont earns its keep when ROE changes. Say an illustrative company, Ormond Hardware, reports ROE up from 15% to 18% and management credits better execution. The breakdown tells a different story.
Every point of the improvement came from the equity multiplier, most likely from buybacks funded with debt. Had net margin gone from 10% to 12% instead, ROE would also have reached 18%, and that would have been a far better story. Same headline. Very different quality.
When the parts mislead
- Liabilities aren't all debt. The equity multiplier counts every liability. A software company that collects a year's subscription in advance holds that cash as deferred revenue, which is a liability, so its multiplier can look high with no borrowing at all. A grocer's unpaid supplier bills work the same way.
- Tiny or negative equity. After years of buybacks, equity can shrink toward zero or below it. The multiplier then explodes or turns negative, and the breakdown stops meaning anything.
- Goodwill sits in total assets. After a big acquisition, asset turnover drops even if every shop and factory is running exactly as before.
- Year-end balances. Using balance-sheet figures from a single date makes the ratios jump when a company raises or repays a lot of capital during the year. An average of the opening and closing balances smooths that out.
- One-off items. A tax credit or a gain on an asset sale lifts net margin for a year. The five-step version helps you catch it, since the tax burden or the gap between EBIT and pre-tax profit will look odd.
The DuPont card on Dividend Line
Every company page has a DuPont & Quality Audit section near the bottom of the Financial Flows tab. It shows the three-step breakdown from the latest annual accounts (net income ÷ revenue, revenue ÷ total assets, total assets ÷ equity) using year-end balances, with a five-year trend beside it.
Under the numbers come a few short readings. The card names the engine, meaning the part that sits furthest above a rough typical value (10% for net margin, 0.5× for turnover, 2× for the multiplier), and when the engine is leverage it says so. It estimates which part moved ROE most since the previous year, checks how much of net income turned into operating cash flow, and puts ROE in a rough band: 25% and up, 15–25%, 8–15%, or below 8%.
Two cautions from this lesson apply there. The equity multiplier counts all liabilities, so read it next to the actual debt. And when equity is zero or negative, the card doesn't attempt the breakdown and reports insufficient data instead.
Takeaways
Sources for this lesson
- Stephen Penman, Financial Statement Analysis and Security Valuation (5th edition, 2013)
- H. Thomas Johnson and Robert Kaplan, Relevance Lost: The Rise and Fall of Management Accounting (1987)