Why I Check Return on Invested Capital Before Return on Equity

Take two companies that run an identical business. Each earns $16 of after-tax operating profit a year on $200 of capital. One reports a return on equity of 12.3%, the other 20.8%. Same factory, same customers, same profit. Yet one looks nearly twice as good. The gap comes entirely from how the capital was financed, and it is the reason I look at return on invested capital, ROIC, before I look at return on equity.

My thesis is simple. ROE tells you how well the shareholders’ slice of the money is working, and management can move that slice with a buyback or a loan. ROIC asks how well all the money is working, whoever supplied it. When the two agree, either is fine. When they diverge, the distance between them is information about borrowing. That is the part I want to read first.

A worked example with round numbers

Company A has $100 of equity and $100 of debt. Its operations produce $16 of after-tax operating profit, so ROIC is 16 over 200, which is 8%. The debt costs 5% before tax; with a 25% tax rate that is $3.75 of after-tax interest. Net income is $12.25, and ROE is 12.25 over 100, or 12.3%.

Company B is the same business. Management borrows $50 more and buys back $50 of shares, so equity is $50 and debt is $150. Operating profit is unchanged at $16 and ROIC is still 8%. Interest rises to $7.50 before tax, $5.63 after tax, net income falls to $10.37, and ROE is 10.37 over 50, which is 20.8%.

Net income went down. ROE went up by more than eight points. Nothing about the business improved, and one dollar of profit was replaced by a smaller dollar on a smaller equity base. If you sorted a screen by ROE alone, B would beat A every time. B is also riskier.

Item	Company A	Company B
Equity	$100	$50
Debt	$100	$150
After-tax operating profit	$16	$16
ROIC	8.0%	8.0%
Net income after 5% debt cost	$12.25	$10.37
ROE	12.3%	20.8%

Same business, same operating profit, less equity. Only the wrapper changed.

What ROIC measures, and what it leaves out

The definition I use is after-tax operating profit divided by the capital tied up in the business, meaning equity plus debt, sometimes net of cash. It ignores how the funding is split, which is the point. A manufacturer’s ROIC is the return on its plants, inventory and receivables, before anyone asks who paid for them.

The catch is that providers do not agree on the inputs. Some subtract cash. Some include leases as debt. Some use average capital while others use year-end. Two websites can show different ROIC for one company. That does not make the measure useless; it means I compare a company to itself over several years and to peers computed on the same basis, never to a number I saw elsewhere.

I also treat one year with suspicion. A single strong year in a cyclical business can flatter it, and a year of heavy investment can depress it before the payoff arrives. I would want at least three or four years of the same company before judging the trend.

The hurdle I compare it against

A return only means something against the cost of the money. I use a rough 8% to 10% band for a mature, ordinary company. It is my own rounding, and the right number moves with interest rates and with how risky the business is. If ROIC sits above the band year after year, each new dollar invested is probably creating value. If it sits below, growth destroys value, however good the revenue chart looks.

In the example above, an ROIC of 8% is at the bottom of my band. Company B’s 20.8% ROE would tempt many people to buy it. Once I see that the business earns 8% and is funded with more debt, I ask what happens if interest rates rise or profit dips by a quarter. In that case profit of $12 would leave B with about $6.4 after interest, and ROE would fall below 13%. The amplification works in both directions.

That asymmetry is why I read a large gap between ROE and ROIC as a warning light. The gap does not prove anything is wrong. It says the shareholders’ return depends on something other than operations, and I need to find out what.

Where the two agree, and where they do not

A company with modest debt and a normal equity base will show ROE and ROIC close together, and the choice between them stops mattering. The same is true for asset-light businesses that carry little debt and hold plenty of cash.

The gap grows in three situations. The first is heavy buybacks, which shrink equity. Some large, cash-rich firms have bought back so much stock that book equity is small or even negative. ROE then becomes meaningless. The second is a business that has taken on debt. The third is accounting: large write-downs or acquisition charges can push equity around for reasons that have nothing to do with what the business earns. In every case I go to the filings, and the 10-K tells me which situation applies.

I would apply this same discipline to a company like UnitedHealth, where a medical cost ratio spike can move earnings quickly but the capital base moves slowly, and to a capital-hungry name like Oracle, where a big build-out lowers ROIC for years before revenue catches up. Those are two different reasons for a low ratio, and only one of them is a warning.

A bank is where ROIC breaks down

Here I have to admit a limit. For a bank or a lender, debt is not a way of financing a factory. Deposits and borrowings are the raw material the business sells on, and interest paid is an operating cost. Dividing operating profit by equity plus debt makes little sense there, so analysts use ROE, or return on tangible equity, for those companies. I do not force ROIC onto them.

So what do I use instead?

Capital One is a fair test case. At around $202.43 the market value is about $124.2 billion, and price to book is 1.1, against a five-year average of 1.0. Dividing market value by price to book gives book equity of roughly $113 billion, an estimate and not a reported figure. Trailing earnings per share of $18.84 times about 613 million shares comes to around $11.6 billion of earnings, so trailing ROE is about 10%.

Reported net income for 2025 was $2.5 billion, down from $4.8 billion the year before, while revenue rose 37% to $53.4 billion. On the same equity estimate that is a return near 2%, well below the trailing figure. Equity and share count both changed over that period, so my division is crude, and I would check the 10-Q before relying on it. But the lesson is clear: for a lender a swing in earnings and a swing in ROE are the same story, and no financing trick or debt ratio explains it.

Look also at the multiples: 10.7 times trailing earnings and 11.4 times forward, with the forward EPS estimate of $17.74 below the trailing $18.84. The market is paying about a book multiple of 1.1 for a lender whose ROE is around 10% on my estimate. That is roughly what you would expect for a business that earns about its cost of equity. The price of $202.43 sits 21.1% below its 52-week high of $257. If ROE recovered to the mid-teens, price to book would probably need to be higher, and if it fell to mid-single digits the market would pay less. That is a testable link, unlike a slogan about quality. I compare banks against each other in my table of four big US banks.

How I would run a screen

Start with ROIC over four or five years, sorted highest to lowest. Drop anything below the hurdle unless it is early in an investment cycle. Then, on the survivors, add ROE and look at the gap. A small gap needs no explanation. A big gap gets a note naming the cause: debt, buybacks or accounting.

Then look at what the return is made of. A high ROIC from a thin margin and fast turnover, like a grocer, behaves differently from one built on a fat margin and slow turnover, like software or a branded consumer product. The first is exposed to cost inflation. The second is exposed to a competitor copying the product. Knowing which one you own tells you what to watch.

Last, check whether the return persists. Ask whether the company earns its ROIC on the next dollar it invests, or only on the old capital already in place. A business that earns 20% on past investments and 6% on new ones will see its ratio fall each year, and the ratio will look fine until it does not.

What would prove this approach wrong

I can name the conditions under which leading with ROIC would mislead me. If a company’s debt is nearly free and stable, for example long-dated and fixed at a low rate, then ROE captures something real: the shareholder truly earns more because the cheap money is working. In that case the gap between ROE and ROIC is a genuine advantage, not a distortion. Utilities and some infrastructure owners fit here.

The second failure is definitional. If I compare ROIC across two providers that treat leases differently, I am comparing accounting choices. It is an easy mistake, and it would push a lease-heavy retailer or airline into the wrong bucket.

So my rule is modest. Use ROIC to see the business, use ROE to see what the owner gets, and read the gap as a question to answer. When the gap is more than about ten points and I cannot explain it from the filings, I do not buy until I can.

Analysis and opinion only, not investment advice. Figures for Capital One come from its filings on SEC EDGAR and its investor site, with market data checked on September 22, 2026; the two-company example is hypothetical and the equity and share-count numbers are my own approximations.