DuPont Analysis Explained: 3-Step and 5-Step ROE

DuPont analysis is the standard way professional analysts decompose return on equity (ROE) into its underlying drivers. The point is simple: two companies can both post a 25% ROE for radically different reasons – one because it runs fat margins, one because it turns its assets fast, one because it is highly levered. If you cannot tell which, you do not really understand why a business earns what it earns.

TL;DR: ROE = Net Profit Margin x Asset Turnover x Equity Multiplier. That is the 3-step DuPont identity. The 5-step version further breaks the margin term into tax burden, interest burden, and EBIT margin so you can separate operating performance from financing and tax effects. Use it to compare companies across industries, to spot whether high ROE is durable or just leverage, and as a sanity check on every DCF assumption you make.

Why decompose ROE at all?

Return on equity by itself is a useful headline number – it tells you how many cents of net income a company earned for every dollar of book equity. But it conflates three very different things:

  • How profitable each dollar of sales is (the margin story)
  • How hard the asset base is working (the turnover story)
  • How much of that asset base is funded by debt vs equity (the leverage story)

A software firm and a discount retailer can both earn 25% on equity. The software firm gets there with 30% net margins, 0.6x asset turnover, and 1.4x leverage. The retailer gets there with 3% net margins, 3.0x asset turnover, and 2.8x leverage. Same headline. Completely different businesses. DuPont analysis is what makes that visible.

The framework comes out of the DuPont Corporation in the 1910s and 1920s – the original version, developed by Donaldson Brown, was actually built to analyze return on investment for industrial operations, not ROE. Modern textbooks generalized it.

The 3-step DuPont identity

ROE = (Net Income / Sales) x (Sales / Assets) x (Assets / Equity)

Or in words:

ROE = Net Profit Margin x Asset Turnover x Equity Multiplier

Algebraically the Sales terms cancel and the Assets terms cancel, leaving Net Income / Equity – which is just ROE. The three ratios are the answer to a different question each:

  • Net Profit Margin (NPM) = Net Income / Sales. How much of every revenue dollar survives as profit after all costs, interest, and taxes.
  • Asset Turnover (AT) = Sales / Total Assets. How much revenue the company generates per dollar of assets. A measure of how productively the balance sheet is being used.
  • Equity Multiplier (EM) = Total Assets / Total Equity. Leverage. If a firm finances $100 of assets with $40 of equity and $60 of debt, the equity multiplier is 2.5x.
The 3-step DuPont identity A waterfall-style diagram showing how ROE equals net profit margin times asset turnover times equity multiplier, with each term labeled by the business question it answers. ROE = Net Profit Margin x Asset Turnover x Equity Multiplier

Net Profit Margin Net Income / Sales Operating skill: how much profit per dollar of sales x

x

Asset Turnover Sales / Assets Capital efficiency: how much revenue per dollar of assets

x

Equity Multiplier Assets / Equity Leverage: how much asset base per dollar of equity

=

Two firms with the same ROE can have very different mixes of these three terms. A retailer leans on turnover. A software firm leans on margin. A bank leans on leverage. DuPont analysis is how you tell them apart.

The 3-step DuPont identity decomposes ROE into a margin term, a turnover term, and a leverage term.

A worked example: same ROE, different business

To make the framework concrete, take three stylized but realistic firms – a software company, a discount retailer, and a US commercial bank – and let each one earn a 24% ROE. (Numbers are rounded for clarity; the point is the mix, not the absolute precision.)

Firm Net Profit Margin Asset Turnover Equity Multiplier ROE
Software firm 30% 0.60x 1.33x 24%
Discount retailer 3% 2.86x 2.80x 24%
US commercial bank 25% 0.10x 9.60x 24%
Stylized example. NPM x AT x EM = ROE; all three firms reach 24% but through wildly different mixes. Bank leverage is illustrative of typical US large-bank equity multipliers; see Federal Reserve Z.1 Financial Accounts for sector aggregates.

Three observations the decomposition forces on you:

  • The retailer’s edge is operational, not financial. If asset turnover slips from 2.86x to 2.0x because store productivity falls, ROE drops to 16.8% even with leverage unchanged.
  • The software firm has almost no leverage to lose. Its 24% comes from margin. If pricing power slips and NPM goes from 30% to 20%, ROE drops to 16%.
  • The bank’s 24% is leverage all the way. Strip the equity multiplier back from 9.6x to a software-firm-like 1.33x and the same business earns roughly 3.3% on equity. That is what the bank capital rules are actually constraining – and it is why CET1 and Tier 1 ratios matter so much for bank ROE durability.

The 5-step DuPont (extended)

The 3-step is the workhorse. The 5-step is what you reach for when you want to separate operating performance from financing decisions and tax effects – useful when comparing two companies in the same industry but with different capital structures.

ROE = Tax Burden x Interest Burden x EBIT Margin x Asset Turnover x Equity Multiplier

Where each new term is:

  • Tax Burden = Net Income / Pre-tax Income. Roughly 1 – effective tax rate. A US firm paying a 21% effective rate has a tax burden of 0.79.
  • Interest Burden = Pre-tax Income / EBIT. How much operating profit survives interest expense. An unlevered firm sits at 1.00; a highly indebted one might be 0.60 or lower.
  • EBIT Margin = EBIT / Sales. Pure operating profitability before interest and taxes.

The trick is that Tax Burden x Interest Burden x EBIT Margin = Net Profit Margin. The 5-step does not change ROE – it just opens the margin term so you can attribute changes correctly. If net margin falls 200 bps, you want to know whether it was a tax-rate change, a debt issuance, or genuine operating compression. Same total, very different read-throughs for the next quarter.

How the math behaves in real US sectors

Aswath Damodaran’s NYU Stern data set is the public benchmark for these ratios at the industry level. The numbers below illustrate how the 3-step DuPont mix differs across major US sectors. The exact values shift annually as he refreshes the data; the pattern – high-margin/low-turnover for software, low-margin/high-turnover for retail, high-leverage for financials – is structural.

US sector (illustrative) Net Margin Asset Turnover Equity Multiplier Implied ROE
Systems software ~22% ~0.55x ~1.7x ~21%
Semiconductors ~25% ~0.65x ~1.5x ~24%
Pharmaceuticals ~15% ~0.45x ~2.5x ~17%
Retail (general) ~3% ~2.0x ~2.5x ~15%
Restaurants ~7% ~1.1x ~3.5x ~27%
Utilities ~10% ~0.3x ~3.5x ~11%
Banks (money center) ~25% ~0.05x ~10x ~12%
Airlines ~5% ~0.8x ~6x ~24%
Illustrative US-sector mix. For current, authoritative figures see Aswath Damodaran – Return on Equity by Sector (NYU Stern) and Damodaran – Margins by Sector, refreshed January annually.
DuPont mix by sector Bar chart comparing net profit margin, asset turnover, and equity multiplier across four illustrative US sectors: software, retail, banks, and utilities. Same ROE, different DuPont mix – four illustrative US sectors Margin x Turnover x Leverage, normalized for display

Software High margin

Retail High turnover

Banks High leverage

Utilities Mid margin + leverage

Net Profit Margin Asset Turnover Equity Multiplier

Bar heights normalized for visual comparison only. See sector data link below the chart.

Source: pattern illustrative; underlying sector ratios from Damodaran NYU Stern – Return on Equity by Sector.

Common mistakes when reading DuPont

1. Mistaking leverage for skill

Leverage juices ROE without changing operating performance. A firm that issues debt and uses the proceeds to buy back stock can mechanically raise ROE without earning a single extra dollar of operating profit. That can be smart capital allocation; it can also be a recipe for a balance-sheet accident the next time the cycle turns. Always look at all three terms, not just the ROE total. This is also why ROIC – which strips out the leverage effect entirely – is the better pure measure of operating skill.

2. Comparing across very different capital structures with the 3-step

If you are comparing a debt-heavy utility to a debt-free SaaS firm, the 5-step DuPont is much more honest. The 3-step lumps the interest burden into net profit margin, which makes the SaaS firm look “more profitable per dollar of sales” partly because it pays no interest. The 5-step exposes that distinction.

3. Using year-end vs average balance-sheet items inconsistently

The denominators in DuPont (assets, equity) are balance-sheet items that shift through the year; the numerator (sales, net income) is a flow over the period. Most academic and Bloomberg-style implementations use the average of beginning and ending balance-sheet values. Mixing year-end and average across companies you compare will silently bias your conclusions.

4. Treating goodwill-heavy serial acquirers like organic compounders

If a company has paid up for many acquisitions, equity is inflated by accumulated goodwill, suppressing ROE relative to a competitor that grew organically. The economic returns might be identical; the optics are not. Either strip goodwill from the equity base for like-for-like comparisons, or use ROIC and cash-on-cash returns as a check.

5. Banks are not industrials

For financial firms, “sales” is interest plus non-interest income, “assets” is dominated by loans and securities, and equity multipliers regularly run 10x or higher. The 5-step DuPont is still informative but the interpretation is different – asset turnover means almost nothing on its own; net interest margin, fee income mix, and CET1 leverage are the operating variables that matter. For more on that, see bank capital ratios.

How DuPont connects to the rest of the toolkit

  • ROIC vs ROE. ROIC measures unlevered return on operating capital. ROE measures levered return on equity. The difference between them is mostly the equity multiplier – which is exactly the third DuPont term. ROIC is the cleaner cross-industry comparator; ROE plus DuPont is the cleaner read on equity-holder economics.
  • Sustainable growth. A textbook identity: sustainable growth rate = ROE x retention ratio. If you decompose ROE via DuPont, you are also decomposing the engine of the growth rate the company can self-fund.
  • DCF. Every DCF makes an implicit assumption about future margin, turnover, and leverage. A DCF that requires permanent margin expansion of 800 bps while turnover stays flat is one DuPont term doing all the work – which is exactly the kind of fragile assumption a careful analyst flags.
  • P/B and the Gordon model. The justified price-to-book ratio in the Gordon constant-growth model is (ROE – g) / (r – g). High ROE expands justified P/B; DuPont tells you whether that ROE is durable.

The short version

DuPont analysis is the simplest way to stop being fooled by ROE. Two firms with the same ROE can have completely different economic stories; the 3-step identity makes the story explicit by splitting ROE into margin, turnover, and leverage, and the 5-step extends the split to separate operating performance from financing and tax effects. Use it every time you look at a profitability ratio you cannot explain, and use it as a sanity check on whether your forecasts secretly require one term to do all the work.

If you only memorize one decomposition in finance, this is the one.

Sources & further reading

Disclosure: This article was produced with AI assistance and reviewed before publication. It is for informational purposes only and is not investment advice.

Leave a Comment