EV/EBITDA: What It Measures, When to Trust It, When to Walk Away

Why this multiple exists

The P/E ratio breaks the moment you try to compare companies with different capital structures. Take two manufacturers with identical plants, identical products, identical operating margins. One funded expansion with equity; the other borrowed heavily. The leveraged company now runs $200M in annual interest expense. That comes straight out of net income, making it look twice as expensive on P/E — not because its business is worse, but because of a financing choice. EV/EBITDA was built to strip that distinction out entirely.

The same problem appears across tax jurisdictions. A company domiciled in Ireland faces a 12.5% statutory rate; a comparable U.S. company faces 21% federally plus state taxes. Compare them on P/E and you're measuring tax strategy as much as business quality. Add differing depreciation schedules — one company uses a 20-year useful life on equipment another depreciates over 10 years — and the income statement reflects accounting choices as much as operating performance.

EV/EBITDA collapses all of that to a single question: what does the total business cost, and what does it earn before those noise factors enter the picture? Private equity firms made the multiple standard practice precisely because they operate with debt-heavy structures. When you're planning to refinance a target after acquiring it, the existing capital structure is irrelevant. What matters is the operating cash generation the business can service debt against. EV/EBITDA answers that question — at the cost of some honesty about real cash costs, which we'll get to.

What 'EV' actually includes

Enterprise value is the total cost of acquiring a business outright — not just buying the stock, but absorbing every claim against it. The formula: market capitalization (share price × shares outstanding), plus total debt (short-term and long-term), plus preferred equity, minus cash and short-term investments.

Debt gets added because when you buy a company, you inherit its obligations. Think of buying a house with a mortgage: if you pay $500,000 for a property carrying a $300,000 mortgage, the total consideration is $800,000 — you're now on the hook for that debt. A company with a $3B market cap and $2B in net debt has a $5B enterprise value. A buyer has to make both equity holders and bond holders whole.

Cash gets subtracted because it offsets the purchase price. A buyer can theoretically extract the cash immediately after closing. A company sitting on $1B in cash and trading at a $4B market cap has a $3B enterprise value, before adding debt. The cash belongs to whoever owns the company, so it lowers what you're actually paying for the operating business underneath.

The practical implication: two companies with identical stock prices but different debt loads are not priced the same. A $20 stock at a company with $5B in debt is a more expensive acquisition than a $20 stock at a company with no debt. EV captures that. Market cap alone does not.

What 'EBITDA' actually measures — and what it deliberately ignores

EBITDA is not a GAAP metric. No accounting standard requires it to be reported. But every major financial data provider calculates it the same way: start from operating income (EBIT), add back depreciation and amortization from the cash flow statement. Or work down from net income: add interest expense, income taxes, depreciation, and amortization. What you get approximates cash generated by operations before the cost of financing the business, paying taxes, or writing down long-lived assets.

Each stripped item has a rationale. Interest is a capital structure artifact — two operationally identical businesses financed differently will show different net income purely because of interest expense. Taking it out lets you compare the underlying business without financing noise. Taxes are a jurisdiction and corporate structure artifact — effective rates vary enormously based on where the entity is incorporated, what credits it has accumulated, and what intercompany arrangements it uses. D&A is an accounting artifact — the same physical asset can be depreciated over dramatically different useful-life assumptions, and amortization of intangibles often reflects purchase accounting from past acquisitions rather than current operations.

Strip all three and you get an approximation of operating cash generation that holds up for cross-company comparisons within sectors where capital structure varies but business models are similar. The formula works cleanly there. Where it stops working is when the businesses being compared differ significantly in how much physical capital they must maintain to keep operating. That's the flaw — and it's structural, not accidental.

The flaw: depreciation is not always fictional

Depreciation is described, correctly, as a non-cash charge. No wire transfer happens when accountants record it. But in capital-intensive businesses, it represents something real: the gradual consumption of physical assets that will eventually need to be replaced. That replacement costs money. When EBITDA strips out D&A, it strips out an accounting entry that, in asset-heavy industries, closely approximates a future cash outflow.

Consider two companies, each reporting $500M in annual EBITDA and each trading at 10x — implying a $5B enterprise value. Company A is a software platform: $30M in annual D&A (servers, office lease improvements), $20M in capex. EBITDA and free cash flow are nearly identical. The 10x multiple is roughly 10x on real cash earnings. Company B is a steel manufacturer: $180M in annual D&A across blast furnaces, rolling mills, and continuous casters — and $220M in actual capex to maintain and replace that equipment. EBITDA is $500M. Free cash flow, after capex and before accounting for it, is closer to $280M. The "same" 10x multiple is actually 18x on real cash. Same number. Very different businesses.

Airlines make this extreme. A major U.S. carrier might trade at 6–8x EV/EBITDA, which sounds almost cheap in most contexts. But aircraft are extraordinarily expensive to maintain and replace. Engine overhauls alone run millions per unit. Heavy maintenance checks on a narrow-body aircraft — the full structural inspection, not the routine between-flight checks — can run $3–5M per airframe. Fleet renewal across hundreds of aircraft requires billions annually. Spread that against the EBITDA and you understand why airline stocks have historically destroyed shareholder capital despite reporting positive EBITDA every year. The multiple was never telling the truth about the underlying economics.

Pipelines illustrate this clearly without the operating drama of airlines. A midstream company running $4B in annual EBITDA, with thousands of miles of steel in the ground, might carry $1.5B in annual depreciation. EV/EBITDA treats that $1.5B as a non-event. But when pipelines need relining, compressor stations need rebuilding, and valves and instrumentation need replacement across 10,000 miles of infrastructure, that cash will come from somewhere. It won't come from the tooth fairy.

Warren Buffett made precisely this point in Berkshire's 1999 letter: "Does management think the tooth fairy pays for capital expenditures?" BNSF, Berkshire's railroad, runs roughly $6–7B in EBITDA annually and also runs roughly $3.5B in annual capex to maintain 32,500 miles of track, bridges, tunnels, and rolling stock. The true earnings power, measured against what actually stays in the business after keeping the infrastructure intact, is materially lower than the EBITDA number implies. The market partially adjusts for this — railroad multiples run at 8–10x, well below the S&P median — but the adjustment is intuitive and approximate. The only way to make it precise is to do the math yourself.

When the multiple is the right tool

EV/EBITDA earns its place in specific situations. Knowing those conditions is as important as knowing the structural flaw.

• M&A transaction comps. Investment banks building fairness opinions and pitch books use EV/EBITDA almost universally for precedent transaction analysis. The buyer in an acquisition is going to refinance the target's debt anyway — the existing capital structure doesn't matter. What matters is what EBITDA they're acquiring and what multiple recent comparable deals paid. For this purpose, the multiple is exactly right.
• Cross-border sector comparisons. Statutory corporate tax rates range from under 10% to over 30% across developed markets. A German industrial at 30x P/E and an equivalent U.S. industrial at 22x P/E might be priced identically on an operating basis — the spread is tax rate, not business quality. EV/EBITDA neutralizes most of that and lets you compare underlying economics directly.
• Sectors with uniform capex intensity. If every company in the comparison set has similar capex-to-revenue ratios — U.S. wireless carriers, large-cap quick-service restaurant chains, specialty chemicals — EV/EBITDA differences across the group more likely reflect real valuation differences than capex adjustment issues. The condition matters: verify the capex profiles are actually similar before assuming they are.
• Infrastructure build-out phases. A data center REIT or cell tower company in aggressive expansion will show suppressed free cash flow because capex massively exceeds D&A on the assets already running. EBITDA captures the economics of the operating asset base without penalizing the company for building future capacity. In this narrow context, EV/EBITDA is a more useful lens than EV/FCF because FCF is being temporarily distorted by deliberate growth investment.
• Quick screens for outliers. When ranking 30 companies in a sector by relative cheapness before doing deeper work on the cheapest ones, EV/EBITDA is fast and directionally useful. The danger is stopping there. Use it to identify which companies deserve closer examination — not to close the investment argument.

How to check whether a low multiple is real

The translation takes three steps and requires two inputs you can pull from any financial statement: D&A as a percentage of revenue (from the cash flow statement or income statement footnotes) and capital expenditures as a percentage of revenue (from the investing section of the cash flow statement).

Step one: start with the EV/EBITDA multiple as given. Step two: compute EV/EBIT. Because EBIT = EBITDA minus D&A, the multiple expands by a factor of EBITDA margin divided by EBIT margin. A business with a 20% EBITDA margin and 8% D&A runs a 12% EBIT margin — the EV/EBIT multiple is 20÷12 = 1.67× larger than EV/EBITDA. A 9x EV/EBITDA becomes roughly 15x EV/EBIT. Step three: substitute actual capex for D&A using the same structure. FCF margin equals EBITDA margin minus capex as a percent of revenue (a simplified pre-tax estimate). That ratio between EBITDA margin and FCF margin is how much larger EV/FCF is than EV/EBITDA.

The gap between EV/EBITDA and EV/FCF tells you how much work the multiple is doing to make the company look cheap. A tight gap — 2–3x — means capex closely tracks depreciation and EBITDA is a reasonable proxy for earnings. A wide gap — 10–15x — means the business eats a large fraction of its EBITDA in capital spending before it can return anything to investors. The calculator below makes this concrete. Enter a real company's numbers and see where the multiple actually lands once capital costs enter the picture.

Try it: From EV/EBITDA to EV/FCF

Enter any company's EV/EBITDA multiple and its cost structure below. The translator shows what that multiple actually implies once D&A and capex enter the picture. Toggle between the SaaS and manufacturer presets to see an identical starting multiple land in very different places after the adjustment — that's the lesson in one interaction.