Return on equity is one number, and one number cannot tell you how it was produced. An 18 per cent ROE might come from a business that charges a premium price and keeps a fat slice of every sale, or from one that earns almost nothing per sale but turns its assets over twice a year, or from one that is quite ordinary at both and has simply borrowed a great deal. Those are three entirely different businesses carrying three entirely different risks, and the ROE column on a screener shows them as identical. DuPont analysis is the arithmetic that pulls them apart — it splits ROE into the components that produced it, so you can see which lever the company is actually pulling.
Why does one ROE number tell you nothing about how it was produced?
Return on equity divides net profit by shareholders' funds. Both halves of that fraction can move for reasons that have nothing to do with each other, and the ratio collapses all of them into a single percentage.
The numerator can rise because the company raised prices, or because it cut costs, or because its tax rate fell, or because it borrowed less. The denominator can fall because the company bought back shares, or paid out most of its profit as dividend, or wrote off a chunk of its net worth. Any of those movements produces a higher ROE. Only some of them describe a better business.
DuPont analysis, named for the American company whose finance team formalised it in the 1920s, is the response. It is not a model and it makes no forecast — it is a rearrangement. You insert terms into the ROE fraction that cancel out algebraically, and what emerges is the same number expressed as a product of three or five components, each of which you can read separately and track over time.
The three-step decomposition, and why it is exact
Start with net profit over shareholders' funds. Multiply the fraction by revenue over revenue and by total assets over total assets — both of which equal one, so the value is unchanged. Rearrange the terms and you have three familiar ratios multiplied together.
The cancellation is worth seeing once, because it is what makes DuPont trustworthy. Revenue appears in a numerator and a denominator, so it cancels. Total assets appears in a numerator and a denominator, so it cancels. What survives is net profit over shareholders' funds, which is where you began. Nothing has been assumed and nothing approximated. That is why the components always multiply back to the reported ROE, and why a mismatch means you have used inconsistent data — an average balance in one term and a year-end balance in another, say — rather than a discovery about the company.
Net profit Net profit Revenue Total assets
ROE = ---------- = ---------- x ------------ x --------------
Equity Revenue Total assets Equity
ROE = Net profit margin x Asset turnover x Equity multiplier- Net profit margin = how many paise of every rupee of sales survive to the bottom line
- Asset turnover = how many rupees of sales the company generates per rupee of assets it owns
- Equity multiplier = total assets divided by shareholders' funds; how many rupees of assets are supported by each rupee of owners' money
- Equity here means shareholders' funds: share capital plus reserves and surplus
What do the three terms actually mean in plain words?
Each term answers a separate question about the business, and each is owned by a different part of the company. Margin belongs to pricing and cost control. Turnover belongs to operations and working capital. The multiplier belongs to the treasury desk and the board.
Read in that order they form a sentence: this is how much we keep from each sale, this is how many sales we squeeze out of the assets we own, and this is how much of those assets we did not pay for ourselves.
- Net profit margin — the profitability lever
- Net profit divided by revenue. How much of every hundred rupees of sales survives everything: raw material, salaries, depreciation, interest and tax. High margins usually signal pricing power, a differentiated product, or a cost structure competitors cannot match.
- Asset turnover — the efficiency lever
- Revenue divided by total assets. How many rupees of sales each rupee of assets produces in a year. A distributor or retailer runs high turnover on thin margins; a cement or power plant runs low turnover because the asset had to be built before the first sale.
- Equity multiplier — the leverage lever
- Total assets divided by shareholders' funds. If it reads 3.0, every rupee of owners' money is supporting three rupees of assets — the other two came from lenders and other liabilities. It is the only one of the three that magnifies losses as faithfully as it magnifies profits.
- Why margin times turnover matters on its own
- The first two terms multiplied together are return on assets. That product is the operating half of the business, before any funding decision. Watching it separately keeps the leverage effect from hiding inside a headline ROE.
Working the three-step through an illustrative company
Take Ganga Speciality Ltd (figures illustrative, not a real company): revenue of Rs 600 crore, net profit of Rs 72 crore, total assets of Rs 600 crore and shareholders' funds of Rs 400 crore.
Work the three terms in order and they multiply back to the reported ROE exactly. That closure is the check you should run every time before interpreting anything.
- 1
Net profit margin — 12.0%
Rs 72 crore of net profit on Rs 600 crore of revenue. Twelve paise of every rupee of sales reaches the shareholder after every cost, including interest and tax.
- 2
Asset turnover — 1.0 times
Rs 600 crore of revenue against Rs 600 crore of total assets. Each rupee of assets produces one rupee of sales in the year — a moderately capital-intensive business.
- 3
Equity multiplier — 1.5 times
Rs 600 crore of assets against Rs 400 crore of shareholders' funds. Two-thirds of the asset base is funded by owners, one-third by lenders and other liabilities. This is a conservatively financed balance sheet.
- 4
Multiply the three — ROE 18.0%
0.12 x 1.0 x 1.5 = 0.18. Cross-check it directly: Rs 72 crore of net profit divided by Rs 400 crore of shareholders' funds is also 18.0 per cent. The identity closes, so the components are usable.
Two companies, the same ROE, completely different risks
This is the insight the whole method exists to deliver. Set Ganga Speciality alongside Yamuna Distributors Ltd (also illustrative, not a real company), a business that buys and resells at low margin and finances a large part of its balance sheet with borrowings. Both companies own Rs 600 crore of assets. Both report an ROE of 18.0 per cent. A screener sorted on ROE places them next to each other.
They are not comparable. Ganga earns its return from margin, holds relatively little debt, and would still be profitable if sales fell meaningfully — it has twelve paise of cushion in every rupee. Yamuna earns the same headline from thin margin, fast turnover and an equity multiplier of 3.0. Three paise of cushion per rupee of sales means a small cost shock or a small price war can take net profit through zero, and the borrowings that produced two-thirds of its asset base still have to be serviced when it does.
So the two ROEs describe different animals. One is a business making money on what it sells. The other is a business making money on how much it can move and how much it can borrow. Neither structure is wrong — high-turnover, low-margin distribution is a legitimate and often excellent model — but the risk profile, the sensitivity to an interest rate cycle, and the behaviour in a downturn are nothing alike. The debt and leverage topic in this module deals with what that equity multiplier costs when conditions turn.
| Component | Ganga Speciality Ltd | Yamuna Distributors Ltd |
|---|---|---|
| Revenue (Rs cr) | 600 | 1,200 |
| Net profit (Rs cr) | 72 | 36 |
| Total assets (Rs cr) | 600 | 600 |
| Shareholders' funds (Rs cr) | 400 | 200 |
| Net profit margin | 12.0% | 3.0% |
| Asset turnover | 1.0x | 2.0x |
| Return on assets (margin x turnover) | 12.0% | 6.0% |
| Equity multiplier | 1.5x | 3.0x |
| ROE | 18.0% | 18.0% |
Watch out — Ganga earns a 12% return on its assets; Yamuna earns 6% and doubles it with borrowed money. The identical 18% ROE is the last thing you should look at on this table, not the first.
The five-step DuPont, and what the extra granularity buys you
The three-step version leaves one term doing too much work. Net profit margin bundles together three completely different things: how well the business operates, how much it pays its lenders, and what the government takes. A margin that fell because operations weakened and a margin that fell because a tax holiday expired look the same in the three-step model, and they mean opposite things.
The five-step decomposition splits net profit margin into three separate ratios — tax burden, interest burden and operating margin — by inserting profit before tax and EBIT into the chain. The terms cancel exactly as before, so the identity still closes to the reported ROE.
What you buy with the extra two terms is attribution. Now a change in ROE can be traced to one of five causes rather than three, and crucially, the operating performance of the business is isolated in a single term that no financing or tax decision can touch. That is the term you want to watch across a cycle.
Net profit PBT EBIT Revenue Total assets
ROE = ---------- x ------ x -------- x ------------ x --------------
PBT EBIT Revenue Total assets Equity
tax interest operating asset equity
burden burden margin turnover multiplier- Tax burden = net profit / profit before tax — the share of pre-tax profit the company keeps; 0.75 means an effective tax rate of 25%
- Interest burden = profit before tax / EBIT — the share of operating profit that survives the interest bill; 1.00 means no debt cost at all, 0.60 means lenders take 40 paise of every rupee of operating profit
- Operating margin = EBIT / revenue — the pure operating performance, untouched by financing or tax
- PBT = profit before tax; EBIT = earnings before interest and tax, i.e. operating profit
The same company through the five-step lens
Ganga Speciality's Rs 72 crore of net profit came from EBIT of Rs 120 crore, less Rs 24 crore of interest, less Rs 24 crore of tax on the resulting Rs 96 crore of profit before tax. Feed those into the five terms.
The five-step reading is a great deal more informative than 'net margin 12 per cent'. It says the business itself runs a 20 per cent operating margin, that lenders take a fifth of that operating profit, and that the government takes a quarter of what remains. If next year's ROE falls, you will know within a minute which of those three moved.
| Term | Calculation | Value |
|---|---|---|
| Tax burden | PAT 72 / PBT 96 | 0.75 |
| Interest burden | PBT 96 / EBIT 120 | 0.80 |
| Operating margin | EBIT 120 / Revenue 600 | 0.20 |
| Asset turnover | Revenue 600 / Assets 600 | 1.00 |
| Equity multiplier | Assets 600 / Equity 400 | 1.50 |
| Product of all five | 0.75 x 0.80 x 0.20 x 1.00 x 1.50 | 0.18 = ROE 18.0% |
Note — Tax burden and interest burden both read as fractions below 1.0, and a higher number is better for the shareholder — 0.85 interest burden means lenders take only 15 paise of each rupee of operating profit. Reading them as percentages of a cost, rather than of what survives, is the most common mistake with the five-step version.
How do you read a DuPont trend over five years?
A single year's decomposition is a description. Five years of decompositions is a diagnosis, because the components move at different speeds and for different reasons, and the pattern of those movements is the finding.
Lay the components out in a table with one row per year and read down each column, not across each row. Ask three questions in order. Is the operating margin holding, and if it is drifting, is the drift steady or lumpy? Is asset turnover rising or falling, and does a fall line up with a capacity expansion that has not yet started producing revenue? And is the equity multiplier stable, or climbing year after year?
The table below shows the pattern worth learning to recognise. ROE rises every single year — from 14.1 per cent to 17.3 per cent — which on a screener looks like steady improvement. Decomposed, the business is doing nothing of the sort. Margin is flat to slightly down. Asset turnover is falling every year, so the asset base is growing faster than the sales it produces. The entire increase in ROE is coming from an equity multiplier that has gone from 1.60 to 2.50.
| Year | Net margin | Asset turnover | Equity multiplier | ROE |
|---|---|---|---|---|
| Year 1 | 8.0% | 1.10x | 1.60x | 14.1% |
| Year 2 | 8.0% | 1.05x | 1.75x | 14.7% |
| Year 3 | 7.9% | 1.00x | 1.95x | 15.4% |
| Year 4 | 7.8% | 0.95x | 2.20x | 16.3% |
| Year 5 | 7.7% | 0.90x | 2.50x | 17.3% |
What is a rising equity multiplier alongside a flat margin telling you?
It is telling you the company has run out of operating improvement and is buying its ROE growth from lenders. That is not automatically a problem — it is a fact about where the return is coming from, and the appropriate response is to go and look at the debt rather than to reach a conclusion from this table.
There are benign explanations and there are worrying ones, and they are distinguishable. A benign version: the company raised debt to build a plant that has not yet started contributing revenue, so assets jumped before sales did. Turnover falls, the multiplier rises, and both should reverse within a couple of years as the plant ramps up. Check the capital work-in-progress line on the balance sheet and the capacity commentary in the management discussion — if a large project is genuinely under construction, the pattern has an end date.
A worrying version: the asset base is growing through receivables and inventory rather than plant. Sales are being bought with credit extended to customers who pay slowly, working capital is swelling, and borrowings are funding the gap. Turnover falls, the multiplier rises, and nothing reverses because there is no plant to commission. The working capital and cash conversion cycle topic covers how to tell those two apart from the balance sheet alone, and the accounting red flags topic covers what it looks like when it goes further.
The other question this pattern forces is what happens on the way down. Leverage is symmetrical. A multiplier of 2.5 turns a 6 per cent return on assets into a 15 per cent ROE, and it turns a bad year at the operating level into a considerably worse one at the shareholder level, while the interest bill carries on regardless.
Pro tip — When the equity multiplier is doing the work, stop reading DuPont and go to the cash flow statement and the debt maturity schedule. The decomposition has told you where to look; it cannot tell you whether the borrowing is safe.
Where DuPont stops working
The method is an identity, so it is always arithmetically true. That does not make it always informative, and there are four situations where it will mislead you if you treat the output as analysis rather than as a starting point.
It inherits every distortion in the inputs. If net profit contains a one-off gain from selling land, the margin term absorbs it and the decomposition happily attributes the ROE to profitability. Strip exceptional items and large other-income lines before you decompose, or the five-year trend will contain a step that describes a transaction rather than a business.
It breaks entirely for lenders. For a bank or an NBFC, an equity multiplier in double digits is normal and expected, because deposits and borrowings are the raw material rather than a financing choice. A standard DuPont on a bank produces a leverage term so large it swamps everything and tells you nothing. Banks are decomposed differently, against a regulated capital adequacy requirement, which is covered in the banks and NBFCs topic.
It cannot see off-balance-sheet obligations. Operating commitments, guarantees given to subsidiaries, and contingent liabilities sitting in the notes do not appear in total assets, so the equity multiplier understates real leverage for a company that uses them heavily.
And it explains nothing about the future. DuPont tells you which component produced last year's ROE. Whether the margin is defensible, whether the turnover can recover, whether the debt can be refinanced — none of that is in the arithmetic. It is a very good diagnostic and a very poor forecast.
- Exceptional items and one-off gains flow straight into the margin term and misattribute the return.
- Banks and NBFCs need a different decomposition entirely — their leverage is the business model, not a choice.
- Off-balance-sheet items and contingent liabilities never reach the equity multiplier.
- Year-end balance sheet figures mixed with full-year profit figures will stop the identity closing; use averages consistently in every term.
- A high ROE on a net worth eroded by past accumulated losses is arithmetic, not performance.
Running a DuPont yourself in ten minutes
Everything needed is in the consolidated profit and loss statement and balance sheet of any annual report filed with the exchanges. Use consolidated figures, use the comparative column for the opening balances, and keep one convention — averages everywhere or year-end everywhere — across all five terms.
Build the five-year decomposition
- Pull revenue, EBIT, profit before tax and net profit for five years from the consolidated P&L.
- Pull total assets and shareholders' funds for the same five years, and average opening and closing for each year.
- Compute all five terms per year and confirm the product equals the reported ROE before reading anything into them.
- Remove exceptional items and unusually large other income from net profit, then recompute, and note how much of the ROE was riding on them.
- Read down the operating margin column first — that is the business, isolated from tax and financing.
- Read the equity multiplier column last, and if it is climbing, cross-check the borrowings note and the capital work-in-progress line before drawing any conclusion.
- Repeat for two direct competitors, because a component is only high or low relative to peers running the same business model.
Key points
ROE = Net profit margin x Asset turnover x Equity multiplier = (Net profit / Revenue) x (Revenue / Total assets) x (Total assets / Shareholders' funds)
Example — Ganga Speciality Ltd and Yamuna Distributors Ltd (both illustrative, not real companies) each report an 18.0% ROE on the same Rs 600 cr of total assets. Ganga gets there with a 12% net margin, 1.0x asset turnover and a 1.5x equity multiplier. Yamuna gets there with a 3% net margin, 2.0x turnover and a 3.0x multiplier. Ganga earns 12% on its assets; Yamuna earns 6% and doubles it with borrowings. The same headline number describes a business with twelve paise of cushion per rupee of sales and one with three.
Pro tip — Compute margin x asset turnover as a single figure and track that separately from ROE. It is return on assets, it is the operating half of the business, and it is the half that leverage cannot flatter. When ROA is flat and ROE is rising, you already know the answer without opening the balance sheet.
Warning — The five-step version fails silently if you mix conventions — an averaged balance sheet figure in one term and a year-end figure in another will still produce five plausible-looking numbers that no longer multiply back to the reported ROE. Always close the identity first. And never run a standard DuPont on a bank or an NBFC: their leverage is the raw material of the business, so the equity multiplier term swamps the decomposition and the output is meaningless.
Frequently asked questions
How do you calculate DuPont analysis?
For the three-step version, compute net profit divided by revenue, revenue divided by total assets, and total assets divided by shareholders' funds, then multiply the three. For the five-step version, replace the first term with net profit over profit before tax, profit before tax over EBIT, and EBIT over revenue. In both cases the product must equal the reported ROE exactly, because the decomposition is an algebraic identity rather than an estimate.
Three-step vs five-step DuPont — which should you use?
Use the three-step when you want a quick read on whether a return came from margin, efficiency or leverage. Use the five-step when the net margin itself has moved and you need to know whether operations, the interest bill or the tax rate caused it. The five-step is strictly more informative, and its extra two terms cost about a minute of work once you have the profit and loss statement open.
What is a good equity multiplier?
It is a description of funding structure rather than a score, and it is only interpretable within a sector. A multiplier of 1.5 means two-thirds of the asset base is owner-funded; 3.0 means only a third is. Capital-intensive and financing-led businesses structurally run higher multipliers than asset-light ones. What matters more than the level is the direction over five years and whether the interest bill it creates is comfortably covered by operating profit.
Why do my DuPont components not multiply back to the reported ROE?
Almost always because the inputs are inconsistent. Mixing an averaged balance sheet figure into one term and a year-end figure into another breaks the cancellation, as does using standalone profit with consolidated assets, or a net profit figure that includes minority interest against an equity figure that excludes it. Fix the convention and the identity closes; it cannot fail for a genuine business reason.
Can DuPont analysis be used for banks?
Not in its standard form. A bank funds itself with deposits and borrowings by design, so its equity multiplier runs far higher than any manufacturing company and dominates the decomposition without telling you anything about performance. Banks are analysed with a modified framework built around return on assets, net interest margin and the regulated capital adequacy ratio instead.
What does a falling asset turnover mean in a DuPont trend?
That the asset base is growing faster than the revenue it produces. That can be temporary and benign — a plant under construction shows up in assets long before it shows up in sales — or it can mean receivables and inventory are swelling because sales are being bought with generous credit. The balance sheet distinguishes them: check whether the growth sits in capital work-in-progress or in current assets.
Does a higher DuPont interest burden ratio mean more debt?
No, the opposite. Interest burden is profit before tax divided by EBIT, so it measures what survives the interest bill rather than what the interest bill takes. A ratio near 1.00 means interest is barely touching operating profit; a ratio of 0.60 means lenders are absorbing 40 paise of every rupee of it. A falling interest burden ratio is the signal that debt cost is rising.