June 12, 2026·6 min read·Jack

Owner Earnings vs. Reported Earnings: What We Use

Quick summary

Reported earnings are an accounting output. Owner earnings are the cash an owner could actually pull out of a business each year while keeping it competitively intact. We use owner earnings, defined as net income plus non-cash charges minus the maintenance capex required to stand still. Reported net income can overstate that number (when a company is plowing cash into growth or carrying heavy real capital needs) or understate it (when non-cash charges like acquired-intangible amortization depress the GAAP line). The gap between the two is where most mistakes get made.

A company can report rising earnings and still be getting poorer. It can also report mediocre earnings and be quietly minting cash. The income statement alone won't tell you which is which, because reported net income is built on accounting conventions, not on the cash a business throws off. That gap is the single most useful thing a fundamental investor can learn to see.

Warren Buffett named the fix in his 1986 letter to Berkshire shareholders. He called it owner earnings, and he was blunt that it's a better guide to value than the GAAP figure even though it's messier to calculate. We agree, and it sits near the center of how we score a business.

What owner earnings actually means

Buffett's definition, lightly paraphrased: take reported net income, add back depreciation, amortization, and other non-cash charges, then subtract the average annual capital expenditure the business needs to maintain its competitive position and unit volume. Adjust for working capital swings while you're at it.

The logic is simple. Depreciation and amortization are real expenses on the income statement but they aren't cash going out the door this year, so you add them back. Capital spending IS cash going out the door but it doesn't hit the income statement directly, so you subtract it. What's left is roughly the cash an owner could take home without starving the business.

The reason this isn't just "free cash flow with extra steps" is the word maintenance. Free cash flow subtracts all capex. Owner earnings subtracts only the capex needed to stand still, and treats the rest as discretionary growth spending. Getting that split right is the whole game, and it's also the part no company hands you cleanly.

Maintenance capex vs. growth capex

This is the judgment call. A company reports one capex number. Your job is to figure out how much of it just keeps the lights on versus how much is building new capacity that doesn't exist yet.

A few ways we approximate it. Compare capex to depreciation over a full cycle: if a stable business spends roughly what it depreciates, maintenance capex is close to depreciation. Look at capex per unit of capacity in years the company wasn't expanding. Read the capex commentary in the 10-K and the calls, where management often tells you what's growth and what's upkeep if you're paying attention. None of these is exact. The point is to be approximately right rather than precisely wrong, because the difference between owner earnings and reported earnings is often larger than the difference between two analysts' price targets.

A live example where reported earnings overstate owner earnings: Oracle

Oracle just reported a record fiscal 2026. Revenue up 21%, non-GAAP EPS up 24%, a backlog that ballooned past $600B. On the income statement, a great year.

Now look through to the cash. Operating cash flow was about $32B. Capital spending was $55.7B, up 162%. Free cash flow for the year was negative $23.7B. The company spent every dollar of operating cash flow and then roughly $24B more, funded by issuing debt and equity.

You can argue, fairly, that most of that $55.7B is growth capex for the AI cloud buildout, not maintenance. Fine. But that's exactly why owner earnings matters here. Even if you generously assume only a slice is maintenance, the cash an owner could actually take out of Oracle this year is a tiny fraction of the reported profit, and arguably negative once you account for the capital the business is consuming to chase that growth. Reported EPS says "up 24%." Owner earnings says "we are pouring cash in, not pulling it out." Both are true. Only one tells you what you're buying.

A live example where reported earnings understate owner earnings: Campbell's

Flip it around. Campbell's reported fiscal Q3 2026 with GAAP EPS of $0.41 but adjusted EPS of $0.50. A chunk of that gap is amortization of the intangibles from its Rao's and Sovos acquisition, a non-cash charge that drags the GAAP line down every quarter without any cash leaving the building.

Add those non-cash charges back, and net of the modest capex a soup and sauce business actually needs (making pasta sauce is not capital intensive the way building GPU data centers is), the cash Campbell's can generate runs ahead of its reported GAAP earnings. That's why a stock printing $0.41 GAAP can still cover a dividend yielding over 7%. The reported number understates the owner earnings, and an investor anchoring only on GAAP EPS would think the payout looks more stretched than the cash actually says it is.

Same tool, opposite directions. Oracle's reported earnings flatter the cash reality. Campbell's reported earnings hide it. Owner earnings cuts through both.

Why we won't value a business on reported EPS alone

A P/E ratio is reported earnings in the denominator. If reported earnings are a poor proxy for the cash a business produces, then the P/E is a poor proxy for how expensive it is. Oracle at ~23x its own forward guidance looks reasonable until you remember those earnings sit on top of a balance sheet bleeding $24B a year. Campbell's at under 10x forward earnings looks cheap, and in this case the owner-earnings cross-check supports that rather than undermining it.

That's the discipline. Reported earnings start the conversation. Owner earnings finish it. When the two diverge sharply, the divergence itself is usually the most important fact about the company.

Where this fits in our framework

Owner earnings feed directly into two of our 167 points: the quality of earnings score (how closely reported profit tracks real cash) and the valuation score (we run the multiple on owner earnings, not GAAP EPS). A company whose reported earnings and owner earnings diverge wildly, in either direction, gets flagged for a closer read before it earns a quality rating.

Want the full 167-point breakdown, owner earnings included, on any stock? →

FAQ

Is owner earnings the same as free cash flow?

Close, but not identical. Free cash flow subtracts all capital spending. Owner earnings subtracts only maintenance capex, the amount needed to keep the business competitively intact, and treats growth capex separately. For a no-growth business the two converge. For a company in a heavy expansion phase, like Oracle right now, they can be miles apart.

How do I find maintenance capex if the company doesn't report it?

You estimate it. Compare capex to depreciation across a full cycle, look at capex in years the company wasn't expanding, and read management's capex commentary in the 10-K and on calls. It's a judgment, not a lookup. Being approximately right beats being precisely wrong.

Which number should I put in a P/E ratio?

For valuation we use owner earnings, not reported GAAP EPS, because reported EPS can be distorted by non-cash charges and ignores real capital needs. A multiple is only as honest as the earnings number underneath it.

Patient, AI-native investing — built for the long term.

Run any stock through the 167-point framework. Free to start, no card required.