Profitability and efficiency ratios
The answer
Return on equity tells you what a company earned on the shareholders' money, including whatever it borrowed. Return on invested capital tells you what the business earned on all the money put into it, whoever supplied it.
The second one is the number that describes the business. Almost all the difficulty is in building the denominator honestly.
Why this costs you money
Return on equity, return on capital employed and the 3-part DuPont breakdown are covered in "Profitability and debt ratios". This article is about the versions that survive contact with a real annual report.
Here is where the money goes.
You compare 2 companies in the same industry. Company A shows a return on equity of 28%. Company B shows 14%. You conclude A is twice as good and pay twice the multiple.
Then you rebuild both numbers from the reports.
- A's equity is small because it has bought back shares for 6 years. The denominator shrank; the business did not improve.
- A has a large amount of capital sitting in projects under construction, which earn nothing yet, and this is excluded from the ratio your data provider used.
- B carries goodwill from an acquisition made 8 years ago. Including that goodwill, B's return on the money actually spent is 9%. Excluding it, the operating business earns 21%.
- A leases its factories and B owns them. Before 2019 that difference alone changed the reported capital base by a large amount.
After the rebuild the ranking may reverse, or the gap may vanish. None of the 4 adjustments requires judgement about the future. All 4 come from the notes.
The general problem: profitability ratios are quoted by data providers who use 1 definition for every company, and the definition that is right for a lease-heavy retailer is wrong for an acquisitive software group. If you did not build the denominator, you do not know what the ratio means.
How it works
Return on invested capital, defined so it can be calculated
Numerator: NOPAT, net operating profit after tax. Take operating profit, which is profit before interest and tax, and multiply it by 1 minus the effective tax rate.
NOPAT = operating profit × (1 − effective tax rate)
The effective tax rate is total tax expense divided by profit before tax, from article 5. Using operating profit rather than net profit removes the effect of how the company is financed.
Denominator: invested capital. There are 2 routes and they should agree.
From the funding side: total equity + total debt + lease liabilities − cash and cash equivalents not needed to run the business.
From the asset side: net property, plant and equipment + right-of-use assets + intangibles and goodwill + net working capital.
ROIC = NOPAT ÷ average invested capital
Use the average of the opening and closing balance, not the closing balance. A company that raised capital in March would otherwise show a ratio distorted by money it held for 3 weeks.
Why this beats return on equity. Return on equity mixes 2 different things: how good the business is, and how much it borrowed. Two companies with the same operating performance will show different returns on equity purely because of debt. ROIC removes the financing decision and describes the business.
The 5 denominator decisions, and what each one is measuring
1. Cash. Operating cash needed to run the business belongs in invested capital. Surplus cash sitting in deposits does not, because it is not part of the operation. Leaving surplus cash in the denominator makes a cash-rich company look worse than it is. Removing all cash makes it look better. Decide, state which you did, and be consistent.
2. Goodwill and acquired intangibles. This is the most consequential choice.
- Including goodwill measures the return on all money spent, including money spent buying other companies. It answers: has management's capital allocation worked?
- Excluding goodwill measures the return the operating business earns on its own assets. It answers: is this a good business?
Both are useful and they answer different questions. Calculate both. If the gap is large, the company is a good business that has overpaid for acquisitions, and that is a specific and important finding.
3. Capital work in progress. Money already spent on assets that are not yet producing anything. Including it lowers the ratio, correctly, because the money is committed. Excluding it flatters a company in the middle of a large expansion. For Indian capital-heavy companies this line can be a large share of the asset base, and the Schedule III ageing schedule now shows how long it has been sitting there.
4. Right-of-use assets and lease liabilities. After Ind AS 116 and ASC 842 took effect around 2019, leased assets appear on the balance sheet. So a retailer's invested capital rose sharply in that year and its reported return on capital fell, with no change to the business. Any return series that crosses that year is not a series. It is 2 series.
5. Revaluation. If assets have been revalued upwards, the denominator rises and the ratio falls. The revaluation reserve within equity tells you the amount.
The 5-step DuPont breakdown
The 3-part version splits return on equity into margin, asset turnover and leverage. The 5-part version splits it further, and the extra 2 pieces are where the information usually is.
ROE = tax burden × interest burden × operating margin × asset turnover × equity multiplier
- Tax burden = net profit ÷ pre-tax profit. Below 1. Rises when the effective tax rate falls.
- Interest burden = pre-tax profit ÷ operating profit. Below 1 for a company with debt. Falls as interest costs rise.
- Operating margin = operating profit ÷ revenue. The business.
- Asset turnover = revenue ÷ total assets. The efficiency.
- Equity multiplier = total assets ÷ equity. The leverage.
Calculate all 5 for 5 years and look at which one moved. A rising return on equity driven by the tax burden term is a government decision. Driven by the equity multiplier, it is borrowing. Driven by operating margin or asset turnover, it is the business. Only the last 2 deserve a higher multiple.
Return on incremental invested capital
This is the most useful and least used measure in this article.
ROIIC = (NOPAT this year − NOPAT 3 years ago) ÷ (invested capital this year − invested capital 3 years ago)
It answers: what did the new money earn? A company can have an excellent historical ROIC from assets built 20 years ago while every rupee or dollar it invests today earns very little. The average conceals it. The incremental figure does not.
Use 3 to 5 year gaps, never 1 year, because new capacity takes time to produce profit. If incremental returns are well below average returns for several periods, the company is growing by destroying value, and its growth deserves a lower multiple, not a higher one.
Efficiency, which is the other half of returns
Returns are margin multiplied by turnover. Two companies can earn the same return on capital in opposite ways: a jeweller with high margin and slow turnover, a distributor with thin margin and fast turnover. Neither is better. They are different businesses and they break differently.
The efficiency measures worth calculating:
- Fixed asset turnover = revenue ÷ net property, plant and equipment. Falling fixed asset turnover after a large capital expenditure programme means the new capacity is not being used.
- Working capital intensity = net working capital ÷ revenue. Rising intensity means growth is consuming more cash per unit of sales.
- The cash conversion cycle, built in article 6.
- Capacity utilisation, disclosed by many Indian manufacturers in the management discussion, and the cleanest efficiency measure there is when available.
What good looks like, by type of business
There is no universal threshold, and quoting one is how people misuse these ratios. What exists is a normal shape per business model.
- Branded consumer goods: high margins, high asset turnover, often negative working capital, so returns on capital are very high and the constraint is reinvestment opportunity.
- Capital-heavy manufacturing and commodities: returns swing with the cycle. Judge across a full cycle, not a year, and compare with the cost of capital rather than with a fixed number.
- Software and services: few physical assets, so return on capital can be enormous and close to meaningless. Use return on incremental capital and cash conversion instead.
- Regulated utilities and infrastructure: returns are set by a regulator. Above-normal returns are usually temporary or are being earned somewhere the regulator does not reach.
- Banks and lenders: ROIC does not apply, because debt is raw material. Use return on assets and return on equity together with capital adequacy.
What it tells you, and what it does not
A high and stable ROIC tells you the company has something competitors have not copied. That is the link to article 3, and it is the strongest quantitative evidence a moat claim can have.
Incremental returns tell you what growth is worth. Growth at returns above the cost of capital creates value. Growth at returns below it destroys value while increasing revenue.
It does not tell you the price is right. A company earning 40% on capital can be a poor investment at 70 times profit.
It is distorted by accounting, always. Companies that expense research and development show smaller asset bases and higher returns than identical companies that capitalise it. Article 5 covers the difference.
It says nothing about survival. A company can earn high returns and still fail if it cannot refinance. That is article 9.
The decision rule
Calculate return on invested capital 2 ways: including goodwill and excluding it. Then calculate the incremental return over 3 years.
If average returns are high but incremental returns are low, treat additional growth as a cost rather than a benefit — unless the new capital is in assets that are genuinely not yet operating, in which case exclude capital work in progress and recalculate before concluding anything.
Try this now
Five minutes, 1 holding, its latest annual report. You are going to rebuild 1 ratio and see how much the definition matters.
- From the consolidated statement of profit and loss, write down operating profit (profit before finance cost and tax), profit before tax and total tax expense. Calculate the effective tax rate, then NOPAT.
- From the consolidated balance sheet, write down total equity, total borrowings (current plus non-current), lease liabilities and cash and cash equivalents plus other bank balances and current investments.
- Calculate invested capital 2 ways: first as equity + borrowings + lease liabilities, and second as the same total minus cash. Divide NOPAT by each. You now have 2 versions of the same ratio.
- Now find 2 numbers in the notes: goodwill, in the intangible assets note, and capital work in progress, in the property, plant and equipment note. Subtract each from your denominator in turn and recalculate.
- Write the 4 resulting percentages on 1 line.
What you should see. Four different numbers for the same company in the same year, and for many companies the highest is more than 1.5 times the lowest.
That spread is the point of the exercise. Nobody is lying. Each version answers a different question. The version excluding goodwill and capital work in progress tells you what the working business earns. The version including everything tells you what management has earned on all the money it has been given.
If the 2 are close, the company has grown organically and its reported profitability is what it looks like. If the version including goodwill is far lower, the company has bought growth and paid more for it than it was worth. That is a finding you can act on, and no screener will show it to you.
Three real cases
1. Boeing, 2013 to 2019 (United States) — return on equity with almost no equity Boeing repurchased a very large quantity of its own shares over these years, which reduced shareholders' equity substantially. Reported return on equity was extremely high, and at points the equity balance itself turned negative. A negative denominator makes return on equity meaningless rather than excellent. Then the 737 MAX was grounded worldwide from 13 March 2019 after 2 fatal crashes, deliveries stopped, and the company needed to raise very large amounts of debt. The lesson is mechanical: a ratio whose denominator has been shrunk by buybacks measures the buyback, not the business, and it also removes the buffer that absorbs a shock.
2. Tata Steel and Corus, from April 2007 (India and United Kingdom) — the gap between returns with and without goodwill Tata Steel completed the acquisition of the Anglo-Dutch steel producer Corus in April 2007 for approximately $12.9 billion, a very large price relative to the acquirer. The European business subsequently faced weak demand and overcapacity, and the group recorded impairments against those operations in later years. For a decade afterwards, the Indian operations earned strong returns on capital while the group's return on total invested capital was far lower, because the denominator included the money spent on the acquisition. Both numbers were true. Calculating only 1 of them would have told you either that this was an excellent business or that it was a poor one, and neither answer alone was complete.
3. Sears Holdings, to 15 October 2018 (United States) — efficiency falling while the ratios were managed Sears Holdings was formed from the 2005 combination of Kmart and Sears. Over the following decade it closed stores, sold or separated assets including its Lands' End business in 2014 and the Craftsman brand in 2017, and reduced its store base substantially. Revenue fell year after year. Sales per store and inventory turnover deteriorated. Asset sales produced gains that flattered reported results in individual years while the operating business was shrinking. The company filed for Chapter 11 bankruptcy protection on 15 October 2018. The measure that described this correctly throughout was not a profitability ratio at all. It was the efficiency pair: revenue per square foot and inventory turnover, both falling for years.
The question that resolves it
A novice sees a high return on capital and asks: is this a good business?
An expert sees the same number and asks: what is in the denominator, and what did the last 3 years of new investment earn?
The first question is often answered yes, correctly, about a business that will still be a poor investment, because the past capital earns well and the new capital does not.
What would make this wrong
If high returns on capital reliably produced high investment returns, buying the highest-ROIC companies would beat the market. It does not, consistently, because the ratio is public and priced.
Honest limits.
Accounting choices drive a large part of the difference between companies. Asset-light structures, expensing versus capitalising, and lease treatment can move reported returns more than actual performance does.
High returns attract competition. That is what the entire moat discussion in article 3 exists to test. A high return with no barrier is a forecast of decline.
The measure fails for financial companies entirely, and for early-stage companies where the asset base is being built and there is no meaningful profit yet.
Averages hide segments. A group's overall return can be excellent while a division destroys capital every year. The segment note is where you check.
In India
Schedule III makes the rebuild easier. Capital work in progress, intangible assets under development, and the split of borrowings between current and non-current are all separately presented, so the adjustments in this article can be done from the face of the statements.
Capital work in progress is often large and slow. Indian capital-heavy companies frequently run multi-year projects. From financial years beginning on or after 1 April 2021, Schedule III requires an ageing schedule for capital work in progress and for projects whose completion is overdue. A project ageing beyond 3 years is capital earning nothing, and it belongs in your denominator with that fact attached.
Return on capital employed is the more commonly quoted Indian measure, usually calculated as earnings before interest and tax divided by capital employed. It is close to ROIC without the tax adjustment. Check which definition of capital employed your source used, because Indian data providers differ.
Other income distorts returns. Many Indian companies hold substantial surplus cash and treasury investments. The income appears below operating profit, and the cash appears in the denominator unless you remove it. Removing both, consistently, is the only way to compare across companies.
Promoter-controlled groups can move profitability between entities. Royalty paid to a parent, purchases from a related supplier and management fees all shift margin. A high return in a listed entity that buys from a privately held related party is a number that depends on a related-party price. Article 7's method for reading the related-party note applies directly here.
In the United States
Buybacks dominate the equity denominator. Large US companies repurchase shares continuously, so equity can shrink for years. Return on equity becomes progressively less informative and can break completely when equity turns negative. Use ROIC, which is unaffected, and always look at total profit rather than per-share figures when judging the business.
Stock-based compensation affects both parts of the ratio. It is an expense in the numerator, and the shares it creates are equity in the denominator. Adjusted profitability figures that exclude it are measuring something that does not happen.
Segment reporting allows division-level returns. US companies typically disclose segment assets as well as segment results, which lets you calculate return on capital by division. That is often the fastest way to see that a company is 1 excellent business and 2 mediocre ones.
Research and development is expensed, so research-heavy US companies show small asset bases and very high returns on capital. Comparing such a company with an Indian or European peer that capitalises development costs requires adjusting one of them.
Pension assets and liabilities sit inside the capital base at older industrial companies, and they move with interest rates. A change in the discount rate can move reported equity, and therefore return on equity, without anything happening in the business.
Where they differ, and what that tells you
In the United States the denominator is most often distorted by buybacks. In India it is most often distorted by capital work in progress and by surplus cash.
That is not a small observation. It tells you which adjustment to make first in each market.
For a US company, start by ignoring return on equity entirely and building ROIC, because a decade of repurchases has usually made equity a number about financing history rather than about the business. Then check whether earnings-per-share growth exceeded total profit growth, which tells you how much of the improvement was arithmetic.
For an Indian company, start by pulling capital work in progress and surplus cash out of the denominator, then putting capital work in progress back with its ageing attached. Indian companies build capacity in large steps, and a company part-way through a project will look unprofitable on a raw calculation and excellent 3 years later. Neither reading is right without the ageing schedule.
A second difference matters for comparison. Because US GAAP expenses research and development while Ind AS permits capitalising development costs, an American research-heavy company and an Indian one will never be directly comparable on any return measure. Adjust 1 of them or compare each with its own domestic peers. Comparing across the boundary without adjusting produces a conclusion about accounting standards and none about the companies.
Carry this
- Build ROIC yourself: NOPAT over average invested capital. The denominator is the whole argument.
- Calculate returns both including and excluding goodwill. The gap tells you what acquisitions cost.
- What the last 3 years of new capital earned matters more than what the old capital earns.