ROIC and Economic Moats: Measuring Business Quality

The formula

Michael Mauboussin and Dan Callahan define it this way: "Return on invested capital (ROIC) \= Net operating profit after tax (NOPAT) / Invested capital (IC)." They define NOPAT as earnings before interest, taxes and amortization (EBITA) minus the cash taxes attributable to EBITA, and invested capital as "the amount of net assets a company needs to run its business."

There are two routes to invested capital, and they must agree because the balance sheet balances:

Their example for Cisco uses average invested capital (the mean of the opening and closing balance), which is the usual way to match a year's profit with the capital in use during that year.

• Operating approach (assets side): net working capital plus net property, plant and equipment, goodwill and other operating assets. Mauboussin and Callahan recommend it when you have the choice, because it shows how efficiently capital is used.
• Financing approach (liabilities and equity side): debt plus equity, plus items such as preferred stock, deferred taxes and pension-related liabilities. It shows how much capital has been supplied and how it was financed.

Why the spread matters

The paper's "$1 test" asks whether a dollar invested creates more than a dollar of market value, and says that happens "only when a business earns a return on investment in excess of the opportunity cost of capital." Their worked example uses a 10% cost of capital: a factory costing $1,000 that earns $80 a year is worth $800 and fails the test; one earning $120 a year is worth $1,200 and passes.

Two caveats from the same paper: the spread "provides no indication of how much capital the company can deploy at a given positive spread", and if two companies have the same spread, "the company that grows faster will create more value." A high ROIC on a small base is not the same as a large value-creating business.

Worked examples: Apple, Costco and Delta

We calculated ROIC for the latest fiscal year of three companies chosen for how different they are. Method, applied identically (our simplified version, not the paper's full adjustments):

Sources are Apple's fiscal 2025 statements, Costco's fiscal 2025 (year to 31 August 2025) and Delta's calendar 2025 earnings releases ($ millions):

(All derived figures are our calculations. Costco's invested capital is the same on the operating approach once its $2,460 million of long-term operating lease liabilities is added, which we checked; Apple's matches exactly.)

• NOPAT \= operating income × (1 − effective tax rate), where the rate is the tax provision divided by pre-tax income. This uses book tax, not the cash taxes the paper prefers.
• Invested capital \= debt + equity − cash and marketable securities, averaged over the opening and closing balance sheets.
• Operating leases are left out of capital and lease costs stay in operating income, consistently across all three.

What a moat is

Morningstar defines its Economic Moat Rating as representing "a company's sustainable competitive advantage": the ability to defend against competition and earn high returns on capital over time. A wide moat is one expected to last more than 20 years, a narrow moat 10 years, and no moat is "either no advantage or one that we think will quickly dissipate." Morningstar names five sources:

These are Morningstar's categories and its judgments about duration. They are an analyst's rating, not a measured quantity.

1. Switching costs: obstacles that keep customers from moving to another product.
2. Network effect: value rises for existing and new users as more people use the product. Morningstar's examples are eBay and Facebook.
3. Intangible assets: patents, licences and brands.
4. Cost advantage: providing comparable or better service at lower cost.
5. Efficient scale: a market served efficiently by one or very few competitors.

How ROIC and moats connect

In our reading, ROIC is the evidence and the moat is the explanation. A high return that competitors can copy will not last, so a moat is a claim about why it will.

McKinsey's study of about 7,000 non-financial US companies from 1963 to 2004 (companies above $200 million in revenue in 2003 dollars, including those later acquired or dissolved) found two things relevant here:

The article also says innovators "may earn above-average returns initially, but competition eventually compels most businesses to pass the savings along to consumers." That is the fade a moat is meant to resist. The study is old and ends in 2004, and McKinsey does not give a formula for its ROIC, so use it for the pattern rather than the levels.

For a rough current benchmark, Damodaran publishes return on capital by industry. His January 2026 US dataset lists these after-tax, unadjusted figures: all firms (5,994) 10.08%, with a "normalized" 10-year figure of 7.72%; software (system and application) 54.28%; semiconductors 41.30%; grocery and food retail 13.75%; air transport 11.58%; general utilities 6.12%. The page lists its date as January 2026 but its footer says "last updated January 2025", and we did not read his definitions page, so his method differs from ours and the figures are not directly comparable with the table above. They do show the same ordering as McKinsey's: asset-light and protected businesses at the top, regulated and capital-heavy ones at the bottom.

• The median ROIC excluding goodwill "averaged nearly 10 percent" and "was relatively constant and in line with the long-term cost of capital." Across the whole market, returns sit close to the cost of capital.
• Industry rankings were stable over time. Pharmaceuticals and consumer packaged goods had the highest returns, and McKinsey attributes this to "sustainable barriers, such as patents and brands." Capital-intensive and highly competitive sectors, such as basic materials and retailing, had low returns. Utilities had a median of 7%.

Pitfalls

• One year proves little. Cyclical businesses like airlines swing with fuel and demand. Look at five to ten years.
• Compare with the cost of capital, not with zero. Our later guide to debt, equity and WACC covers how to estimate it.
• Check what is in the denominator. Goodwill, cash, leases and buyback-shrunken equity can each change the result by multiples, as the sensitivity checks above show.
• A high ROIC is not a moat. It may be a temporary cycle peak, a one-off tax benefit or an accounting effect.
• Growth matters as well as the spread. A moat that can reinvest large amounts at high returns is worth more than one that cannot.

A checklist

This article is general information, not financial advice. Company figures are as reported in the sources below; NOPAT, invested capital, ROIC and the sensitivity cases are our calculations.

1. Calculate NOPAT from operating income and a tax rate, and say which one you used.
2. Calculate invested capital both ways and check they agree.
3. Use average capital, and run the number with and without goodwill and cash.
4. Split ROIC into margin and turnover to see which route the company uses.
5. Look at 5 to 10 years and compare with peers and with an estimate of the cost of capital.
6. Ask which of the five moat sources, if any, explains the return, and what would erode it.