The return on equity formula
ROE = net income ÷ average shareholders' equity
Average equity = (equity at the start of the year + equity at the end) ÷ 2
Net income is the bottom line of the income statement, after interest and tax and after any preferred dividends, since those belong to a different class of owner. Shareholders' equity is the balance-sheet total of share capital, retained earnings and reserves. Profit is earned over a year while equity is a snapshot, so the average of the opening and closing balances is the honest denominator; year-end equity alone understates ROE for a growing company and overstates it after a buyback.
PackCo, the fictional packaging maker in the what is ROIC article, earned $202.5 million of net income last year on $1,150 million of equity, a return on equity of 17.6%. On its own the figure says the owners' book capital earned a healthy return. What it does not say is how, which is the next question.
The DuPont breakdown: three ways to earn the same ROE
ROE = (net income ÷ revenue) × (revenue ÷ total assets) × (total assets ÷ equity)
= net margin × asset turnover × equity multiplier
The identity, named for the company whose analysts popularised it, splits ROE into profitability, efficiency and leverage. For PackCo, with $2,000 million of revenue and $2,045 million of assets:
| Factor | PackCo | What it measures |
|---|---|---|
| Net margin | 10.1% | Profit per dollar of sales: pricing power and cost control |
| Asset turnover | 0.98× | Sales per dollar of assets: how hard the capital works |
| Equity multiplier | 1.78× | Assets per dollar of equity: how much is borrowed |
| ROE | 17.6% | 10.1% × 0.98 × 1.78 |
A luxury brand earns its ROE through margin; a discount retailer through turnover; a utility or a bank through the multiplier. Two companies with the same ROE can be very different businesses, and the breakdown says which lever is doing the work. It also says which lever is fragile: margin can be competed away, turnover tends to be structural, and leverage can be withdrawn by lenders at the worst moment.
What leverage does to ROE: a worked example
Suppose PackCo borrows $500 million at 5% and uses it to buy back shares. Nothing about the operations changes. Interest rises by $25 million, so net income falls from $202.5 million to $183.75 million. Equity falls from $1,150 million to $650 million. Assets are unchanged, because the cash borrowed went straight out again.
| Before | After the debt-funded buyback | |
|---|---|---|
| Net income | $202.5m | $183.75m |
| Shareholders' equity | $1,150m | $650m |
| Equity multiplier | 1.78× | 3.15× |
| ROE | 17.6% | 28.3% |
| ROIC | 14.5% | 14.5% |
The company earns less, carries more risk, and reports a return on equity eleven points higher. That is the whole case for reading ROE alongside return on invested capital, which measures operating profit against all the capital in the business, debt included, and does not move. The ROIC calculator computes it from the same inputs. Where ROE is high and ROIC is ordinary, the difference is borrowed.
What counts as a good ROE
The bar is the cost of equity, the return shareholders require for the risk, which for a large, stable company is usually somewhere around 8% to 10%. An ROE consistently above that creates value as the company retains and reinvests earnings; one below it means every retained dollar is worth less than a dollar paid out. As conventions: 10% to 15% is respectable, 15% to 20% sustained over a cycle is strong, and anything far above that deserves a check of the equity multiplier before it is admired. Consistency matters more than the level. Five years of 14% says more than one year of 30% followed by a loss, and the trend tells you whether the business is strengthening or being competed away.
Where ROE misleads
- Buybacks and write-downs shrink the denominator. A company that has bought back stock for a decade can have tiny or negative book equity and an ROE of hundreds of percent, or none at all. The business did not get better; the ledger changed.
- Goodwill inflates it. Acquisitions add goodwill to equity and lower ROE; a later impairment removes it and raises ROE mechanically.
- One-off profits. A disposal gain or a tax benefit lifts a single year's figure. Normalise net income first.
- Intangible-heavy businesses. Software and brand companies expense the investment that creates their value, so their book equity understates their capital and their ROE overstates their returns. Compare within an industry.
- Negative equity. The ratio is undefined or negative; look at ROIC and cash flow instead.
For banks and insurers ROE is the right measure, because debt is their raw material rather than their financing and ROIC has no clean meaning. For everyone else, the stock screener guide favours ROIC, with ROE as the cross-check that reveals leverage.
Using ROE in a screen
A reasonable rule set pairs the level with its quality: five-year average ROE above 12% to 15%, no year below the cost of equity, an equity multiplier below about 3× outside financials, and a ROIC within a few points of the ROE, so the return is earned rather than borrowed. Add a valuation rule such as free cash flow yield or a price-to-book ceiling, because a high-ROE company is usually priced as one; the price to book ratio article shows how ROE sets the book multiple a company deserves. The value stock screener article assembles a full set of rules, and VALYSTR's screener runs them on figures taken from each company's filings. Earnings per share growth, the other number ROE is often confused with, is covered in what is a good EPS.
Screen on returns that are earned, not borrowed
Combine a return-on-equity rule with ROIC, leverage and valuation limits in the screener, with the figures taken from each company's filings.
Frequently asked questions
What is a good return on equity?
One that sits consistently above the company's cost of equity, roughly 8% to 10% for a large stable business. Ten to fifteen percent is respectable, fifteen to twenty percent sustained is strong, and a figure far above that usually reflects leverage or a shrunken equity base rather than a better business. Compare within an industry and read it with return on invested capital.
What is the difference between ROE and ROIC?
ROE divides net income by shareholders' equity, so borrowing to buy back shares raises it even though the business is unchanged. ROIC divides after-tax operating profit by all the capital in the business, debt and equity alike, and is unaffected by financing. ROE is the right measure for banks and insurers; ROIC is cleaner for everyone else, with ROE as a check on leverage.
Why can ROE be misleading?
Because its denominator is book equity, an accounting residual that buybacks, write-downs and acquisitions move for reasons unrelated to the business, and because its numerator rises with leverage. A company with very little equity can show a huge ROE while earning ordinary returns on its capital. The DuPont breakdown shows whether margin, efficiency or borrowing is responsible.
Should I use average equity or year-end equity?
Average equity, the mean of the opening and closing balances, because profit is earned across the year while equity is measured at a moment. Year-end equity understates ROE for a company that raised capital during the year and overstates it for one that bought back shares. Use the same convention for every company you compare.
What does a negative ROE mean?
Either net income was negative, meaning a loss, or shareholders' equity was negative, meaning liabilities exceed assets on the books, which often follows years of buybacks or large write-downs. In the second case the ratio has no useful meaning; judge the company on return on invested capital, cash flow and the trend in operating profit instead.