Return on Invested Capital

What ROIC actually measures

ROIC asks one question: for every dollar of capital this business has deployed — from shareholders and lenders both — how much operating profit does it earn after taxes? That framing is different from the other return ratios most investors reach for first. ROE only counts equity. ROA uses total assets. ROIC uses the capital that actually funds the business: long-term debt plus equity, minus cash sitting idle on the balance sheet. The denominator is what makes ROIC harder to manipulate.

Consider a company that borrows heavily to buy back its own shares. Return on equity climbs because the equity base shrinks — but the business isn't more productive. It's just more leveraged. ROIC stays flat or falls, because the new debt flows straight into the invested capital base. That divergence is the tell. When ROE and ROIC move in opposite directions, someone is engineering a ratio rather than building a business.

ROA fails differently. It divides by total assets, which includes accounts payable, accrued expenses, and other liabilities the company didn't have to fund with equity or debt. A business can post a low ROA simply because it carries a lot of supplier credit — not because its operations are weak. ROIC strips those financing-free liabilities from the denominator and gets to the capital that actually had a cost.

When Amazon (AMZN) built out its logistics network through the mid-2010s, ROIC fell sharply for several years as invested capital swelled ahead of profits. Observers who only watched earnings missed why the spending made sense: Amazon was deploying capital into assets that would eventually earn returns well above its cost of capital. Investors who understood ROIC understood why the company could keep spending without destroying long-term value — and why competitors without the same return potential on each marginal dollar couldn't afford to match it.

The formula — and where to pull the numbers

ROIC = NOPAT ÷ Invested Capital. Two terms. Neither shows up as a labeled line in a standard 10-K. You build them from pieces that do.

NOPAT — Net Operating Profit After Tax — starts with operating income, which most companies label "Operating income" or "Income from operations" on the Consolidated Statements of Operations. Then apply a tax shield: NOPAT = Operating Income × (1 − effective tax rate). The effective tax rate is income tax expense divided by pre-tax income, both of which appear on the same statement. You're stripping out interest expense — the cost of debt — so the result reflects how the core operations performed regardless of how the company is capitalized.

Worked example: In its fiscal 2023 annual report, Fastenal (FAST) reported operating income of $1.24 billion. Income tax expense was $303 million against pre-tax income of $1.24 billion, giving an effective tax rate of roughly 24.4%. NOPAT = $1.24B × (1 − 0.244) = approximately $937 million.

Invested Capital takes more assembly. The cleanest approach works from the balance sheet:

• Start with total assets (bottom of the assets section)
• Subtract non-interest-bearing current liabilities: accounts payable, accrued expenses, accrued compensation, deferred revenue — anything in the current liabilities section that the business didn't have to pay interest on
• Subtract excess cash — anything above roughly 1–2% of revenue. Operating cash is part of the business; a pile of Treasury bills earning 5% is not a capital allocation decision, it's a parking decision

What remains is the capital that shareholders and lenders actually funded. For Fastenal's fiscal 2023 balance sheet: total assets of roughly $4.3 billion, minus non-interest-bearing current liabilities of approximately $520 million, minus minimal excess cash (Fastenal runs lean and generates cash faster than it accumulates) ≈ $3.75 billion in invested capital.

ROIC = $937M ÷ $3.75B ≈ 25%. For a distribution business — an industry where most competitors struggle to clear 15% — that number raises a real question: what structural advantage allows Fastenal to earn so much more on the same category of assets? (The answer lives in its branch-level density model and high-touch service for plant and facility managers, which creates switching costs that commodity distributors don't have.) But ROIC alone doesn't answer that question. It just tells you the question is worth asking.

ROIC without WACC is a number with no context

A 12% ROIC sounds reasonable. Whether it actually is depends entirely on what it cost the company to raise that capital. The weighted average cost of capital — WACC — blends the after-tax cost of debt with the cost of equity, weighted by the proportion of each on the balance sheet. For most publicly traded companies, WACC runs between 7% and 12%. A company with a 12% ROIC and a 12% WACC is running in place. It earns exactly what its capital providers require. No surplus, no compounding, no value creation — just a business paying rent on its own capital.

The spread is ROIC minus WACC. Every percentage point of positive spread represents value the business creates with each additional dollar it deploys. Every point below zero is value it destroys — quietly, persistently, in a way that doesn't show up as a loss on the income statement.

Take two illustrative industrials sitting next to each other in a screener: Company A posts 14% ROIC; Company B posts 9%. Company A looks better on the ratio alone. But if Company A's WACC is 13% and Company B's is 7%, the picture inverts. Company A is creating 1 percentage point of economic surplus per dollar invested; Company B is creating 2. The spread, not the raw ROIC, is the signal.

This is why See's Candies mattered so much to Berkshire Hathaway's early compounding. See's earned roughly 25% ROIC through the 1970s and 1980s in an era when its cost of capital was approximately 10–12%. That 13–15 point spread, sustained year after year against a captive brand with real pricing power, was the compounding machine. Buffett didn't need to project a DCF. He needed to see that the spread was wide, durable, and tied to something competitors couldn't easily replicate. The ROIC number surfaced the quality; the WACC comparison confirmed it was actually creating value.

What counts as good — and why industry matters

The honest answer is that "good" is relative to the business model and the cost of capital. There is no universal threshold that applies across sectors. Capital-light businesses — software platforms, payment networks, asset-light services — routinely post ROIC of 30% to 60% or higher because their invested capital base is small relative to operating profits. Microsoft's ROIC has exceeded 40% in recent years. Visa's has run above 30% consistently. These companies don't own factories or pipelines; their competitive position lives in code, network effects, and long-term contracts. The capital base stays thin even as profits scale.

Capital-intensive businesses operate under a different math. Regulated utilities like Duke Energy (DUK) earn 8–10% ROIC by design — regulators set allowed returns tied to the rate base, which means ROIC is essentially capped. Midstream pipeline companies earn 10–14%. Large integrated oil companies oscillate between 8% and 15% through the commodity price cycle. Airlines, in aggregate, have historically earned below their cost of capital — which is why the industry has destroyed more shareholder wealth than it's created over its entire history, despite generating enormous revenue.

A few sectors where ROIC surprises people:

• Homebuilders: Most run 15–25% ROIC. NVR Inc. (NVR) has consistently posted 30–40%+ by using land option contracts instead of owning land outright — a structural choice that keeps invested capital permanently thin relative to peers who carry land on the balance sheet.
• Specialty distribution: Businesses like AutoZone (AZO) and Fastenal routinely hit 20–30% because their inventory turns and pricing discipline compress invested capital even as revenue grows.
• Commodity chemicals: Single digits in typical years, with ROIC tracking commodity prices more than operational quality. High ROIC in a commodity upcycle tells you almost nothing about the business.
• Asset-light consumer brands: Companies that license rather than manufacture — think Yeti (YETI) or certain apparel brands — can post 20–35% ROIC because they outsource the capital-intensive production while keeping the margin.

The question to ask is never "is 18% good?" but: is 18% above this sector's median, and is it above this company's cost of capital? A 12% ROIC is excellent for a regulated utility and mediocre for a software company. Comparing across sectors misleads. The only number that matters universally is the spread.

The reinvestment trap: high ROIC means nothing without room to grow

A company earns 30% ROIC. Now ask: on how much capital can it earn that rate, and for how long?

If the answer is $50 million per year — because the market niche is small, competition is intensifying, or the addressable demand is capped — then the 30% ROIC compounds on a bounded base. The business is valuable but its ceiling is visible. If the answer is $5 billion per year — because the company operates in a large, growing market with durable pricing power — then the same 30% ROIC becomes one of the rarest things in investing: a wide reinvestment runway at a high return rate. That's where compounding becomes exponential rather than linear.

This is why growth and ROIC can't be evaluated separately. The rough math is: value creation ≈ Capital deployed × (ROIC − WACC). A company that redeploys $3 billion per year at 20% ROIC with a 9% WACC creates $330 million of economic value annually. A company that redeploys $100 million at 30% ROIC with the same cost of capital creates $21 million. The second company has the better ratio and the worse compounding engine.

Microsoft's ability to pour massive cash flows into Azure through the late 2010s and early 2020s is the clearest recent example. The cloud business earned high returns and had a large and expanding total addressable market. Microsoft wasn't just earning high ROIC — it was earning high ROIC on a growing capital base, and the market eventually priced in the compounding. Between fiscal 2016 and fiscal 2023, Microsoft's ROIC roughly doubled, which is not a coincidence: it reflects what happens when a business with durable competitive advantages finds a large reinvestment opportunity at returns well above its cost of capital.

The mirror case is a business that earns 25% ROIC but distributes nearly all of its earnings as dividends or buybacks. That can be a rational decision — the company simply lacks reinvestment opportunities above its cost of capital, and returning cash beats deploying it poorly. See's Candies was exactly this: high ROIC, limited reinvestment runway, so Buffett took the cash and redeployed it into businesses with more room. That's the framework. High ROIC with limited reinvestment is a mature compounder. High ROIC with large reinvestment is a growth compounder. They deserve different multiples and different holding periods.

How to use ROIC when you're actually looking at a stock

The practical workflow starts with a five-year trend, not a single-year number. Pull ROIC from each of the last five 10-K filings — or use a screener that calculates it consistently — and look at the direction and stability before you look at the level. A company that has posted 18% ROIC for five straight years is a fundamentally different business than one that posted 24% in year one, 19% in year two, and 14% last year. The first company has a demonstrated ability to sustain returns. The second is in the early stages of compression, which you should understand before the market prices it in.

Then compare to sector peers over the same period. If the whole sector's ROIC compressed by 5 percentage points over three years, and your company compressed by 3, it may actually be gaining share. If the sector held steady and your company compressed, that's a company-specific problem worth diagnosing.

Green flags to look for:

• ROIC consistently above WACC for five or more years, not just one good period driven by a favorable macro environment
• ROIC expanding even as revenue and capital deployment grow — this is the scale economies signal. Returns should decline at scale in commoditized businesses; when they don't, something structural is at work
• Management discusses return on invested capital explicitly in earnings calls and shareholder letters. Companies that manage to ROIC tend to allocate capital better than those that manage to EPS

Red flags that warrant deeper investigation:

• ROIC below WACC for three or more consecutive years with no recovery. This is a business that is systematically destroying value on each incremental dollar deployed. The stock can still go up — markets are not always rational — but the underlying economics are working against you
• ROIC that jumped sharply in one year because of a large goodwill write-down or asset impairment. The denominator shrank; the numerator didn't improve. Check whether NOPAT grew or whether the invested capital base just got smaller
• A widening gap between ROIC and ROE. If ROE is climbing while ROIC is flat or falling, the company is adding leverage. That's a financing change, not an operational improvement
• A sudden ROIC drop immediately after a large acquisition. This is normal and expected — goodwill expands the invested capital base before profits from the acquired business contribute meaningfully to NOPAT. The real question is whether ROIC recovers toward pre-acquisition levels within two or three years. If it doesn't, the acquirer overpaid or failed to integrate

One last thing: ROIC is not a substitute for reading the business. It's a diagnostic. A high and rising ROIC prompts you to ask why — and the answer to that question is where the real analytical work happens. Fastenal's ROIC is high because of branch density and service quality. NVR's is high because of land optionality. Visa's is high because of network effects and no credit risk. The ratio points you at the question. The 10-K answers it.