Owner Earnings: Buffett's Profit Metric Beyond Net Income

Berkshire Hathaway (BRKB) chairman Warren Buffett has long argued that net income is an accounting figure, not a business one. His preferred metric — owner earnings — strips away depreciation games and tells you what cash actually belongs to the shareholder.

What is owner earnings?

Owner earnings is Buffett's adjustment to GAAP net income that subtracts the cash a company actually has to spend to maintain its existing earnings power. The math reframes profit as "what's left for shareholders after the business pays its own bills, including the unsexy ones depreciation hides."

Buffett introduced the term in his 1986 BRKB annual letter, framing it as a more honest read of long-term cash generation than reported EPS. The reasoning: depreciation is a backward-looking accounting estimate, while real businesses need real reinvestment to keep operating. If those two numbers diverge, owner earnings catches what GAAP misses.

The formula in plain English:

Owner Earnings = Net Income
                + Depreciation & Amortization
                + Other non-cash charges
                − Maintenance capex
                − Necessary working capital additions

Notice what's missing: stock-based compensation. Buffett treats it as a real cost (it dilutes you), so the cleanest version of the formula treats SBC as cash spent. We covered the SBC trap in fundamental analysis — that's the deeper rabbit hole.

How do you calculate it for a real company?

Start with the cash flow statement, not the income statement. The income statement is where management tells you the story they want; the cash flow statement is where the truth lives.

Take MSFT as a worked example. In fiscal 2025, the company reported approximately:

• Net income: ~$88 billion
• Depreciation & amortization: ~$22 billion
• Total capex: ~$56 billion (heavily AI-driven)
• Working capital change: ~$1 billion absorbed

The hard part: estimating maintenance capex. Total capex was about $56 billion, but maybe ~$15-20 billion of that is genuinely "maintenance" (replacing existing servers, software systems, real estate). The rest is growth capex.

Buffett's owner earnings for MSFT, plugging conservative estimates: ~$88B + ~$22B − ~$18B (maintenance capex) − ~$1B (working capital) ≈ ~$91 billion

That is meaningfully higher than reported net income — because reported net income deducts ~$22B in depreciation that's roughly catching up to but not exceeding ~$18B of true maintenance need. The gap matters: it tells you MSFT generates more shareholder cash than its income statement suggests.

Owner earnings vs reported earnings: a real comparison

Note: figures approximate, based on TTM filings as of April 2026. Maintenance capex estimated as approximately 60-70% of total capex for tech, 80-90% for financials.

Look at JPMorgan (JPM): owner earnings is lower than net income. Banks aren't bad businesses, but their reported earnings overstate cash generation because regulatory capital requirements force reinvestment that depreciation doesn't fully capture. That's a useful red flag — and one a P/E ratio alone never reveals.

Wells Fargo (WFC) shows the same pattern. The takeaway: owner earnings flatters asset-light businesses (software, brands) and penalizes asset-heavy ones (banks, utilities, airlines). It's not a one-size-fits-all metric.

Where does owner earnings break?

It breaks when maintenance capex is hard to separate from growth capex. Amazon (AMZN) is the textbook problem case: roughly 80% of its 2026 capex is AI/data-center buildout. Is that maintenance for the cloud business it already runs? Or is it growth investment for new revenue lines? Reasonable analysts come up with owner-earnings numbers that differ by approximately 30-40% on AMZN, depending on that single judgment call.

The metric also breaks for cyclical businesses. A steel company in a recession reports low net income and low capex, which makes owner earnings look fine. The recovery year then "discovers" that maintenance capex was actually higher than the trough year suggested. Owner earnings averages out over a full cycle — but in any given quarter it can mislead.

Three rules for using it well:

1. Always cross-check against free cash flow. If owner earnings is materially different from FCF, your maintenance capex estimate is suspect.
2. Use it for asset-light, predictable businesses first. AAPL, Costco (COST), and the durable consumer franchises were Buffett's preferred hunting ground for a reason.
3. Don't pretend it's precise. The point isn't to compute owner earnings to the dollar — it's to spot the businesses where reported EPS overstates economic reality.

For the broader valuation toolkit, see our fundamental analysis library and the super investors learning path, which traces this metric back through Buffett's letters.

Why does Buffett still prefer it after 40 years?

Because it forces you to think like an owner. A business that needs to spend an extra dollar of capex for every dollar of growth has zero owner earnings — it just looks profitable. That distinction is invisible in a P/E screen but obvious in an owner-earnings screen.

The metric also resists narrative inflation. Tech management teams love to quote adjusted EBITDA, which adds back stock-based compensation. Owner earnings refuses that adjustment, which is why companies like Salesforce (CRM) historically look more expensive on owner earnings than on EBITDA. The market has slowly come around to that view in 2025-2026, with valuation compression in heavily-SBC-dependent names.

The real value isn't the calculation. It's the discipline. Owner earnings forces you to ask, every quarter: how much cash actually belongs to the shareholder, after the business pays for staying alive? That's a harder question than it sounds.

Ready to analyze these stocks yourself? Search any ticker on MainRatios to see valuations from 6 legendary investors - free.