When P/E Lies and EV/EBITDA Tells the Truth (and Vice Versa)

Pull up any stock screener and you'll see two valuation columns sitting next to each other: P/E and EV/EBITDA. Most of the time they roughly agree. When they don't, one of them is lying — and figuring out which one is the whole job.

Here's the short version: P/E lies when capital structure or non-cash charges distort net income. EV/EBITDA lies when capex, working capital, or stock-based compensation are doing real economic damage that EBITDA pretends doesn't exist. The rest of this post is how to tell which trap you're in.

What each ratio is actually measuring

P/E (price divided by earnings per share) is an equity-holder's ratio. It sits on top of the capital stack — after interest, after taxes, after depreciation. It answers: "What am I paying per dollar of profit that belongs to me as a shareholder?"

EV/EBITDA is a whole-business ratio. Enterprise value adds debt and subtracts cash, so it values the entire operation regardless of how it was financed. EBITDA (earnings before interest, taxes, depreciation and amortization) strips out financing decisions and accounting allocations to approximate operating cash generation.

The two diverge whenever the things EV/EBITDA strips out — debt, taxes, D&A — are doing meaningful work in the real economics of the business.

When P/E is the one lying

Three common setups:

1. Heavy, non-economic D&A. A cable operator, a tower REIT, or a private-equity-style rollup will carry enormous amortization of acquired intangibles. That charge is real on the income statement but doesn't reflect cash going out the door. P/E looks punishingly high; EV/EBITDA shows a business that actually gushes cash. American Tower, Liberty-style holdcos, and most levered media businesses have lived here for a decade.

2. Different capital structures inside the same industry. Compare two hotel operators where one owns its real estate (huge depreciation, lots of debt) and the other is asset-light. Their P/Es will be incomparable. EV/EBITDA at least puts them on the same footing before you adjust further.

3. One-time tax or interest noise. A company with a big NOL carryforward (a tax shield from prior losses) will print a flattering P/E this year that won't repeat. A company that just refinanced expensive debt will see EPS jump for reasons that have nothing to do with the underlying business. EV/EBITDA ignores both.

Rule of thumb: if D&A is more than ~20% of EBITDA and a meaningful chunk of it is intangible amortization from M&A, distrust the P/E.

When EV/EBITDA is the one lying

This is the trap that catches more sophisticated investors, because EV/EBITDA feels like the grown-up ratio. It isn't always.

1. Capex is real and recurring. EBITDA treats depreciation as an accounting fiction. For an airline, a semiconductor fab, a midstream pipeline, or a steel mill, depreciation is a rough estimate of the cash you must spend to keep the lights on. If maintenance capex roughly equals D&A, then earnings (which subtract D&A) are closer to truth than EBITDA. A 6x EV/EBITDA on a capital-intensive cyclical can be more expensive than a 15x P/E on an asset-light compounder.

2. Stock-based comp is a real cost. Most tech companies report EBITDA that adds back SBC. If a software company is paying out significant revenue in stock to employees, that's dilution coming straight out of your ownership. P/E (GAAP) at least includes it. EV/EBITDA, as commonly quoted, pretends it doesn't exist.

3. Working capital is funding the growth. A retailer or a distributor can grow EBITDA while inventory and receivables eat all the cash. EBITDA doesn't see it; free cash flow does. P/E sees it indirectly through the eventual write-downs.

4. Leverage is doing the work. A highly levered business shows a great EV/EBITDA multiple because EV includes the debt — but if the debt load means a single bad quarter wipes equity holders out, the "cheap" multiple is compensation for risk, not opportunity.

Rule of thumb: if capex consistently runs at 80%+ of D&A, or if SBC is material relative to revenue, EV/EBITDA is flattering the business.

A simple two-step reconciliation

When the two ratios disagree, don't pick a side — reconcile them. Walk down the income statement:

1. Start with EBITDA. Subtract maintenance capex (use D&A as a proxy if capex is lumpy). Subtract cash interest. Subtract cash taxes. Subtract stock-based comp. What you have left is something close to free cash flow to equity.
2. Divide market cap by that number. That's your "honest P/E."

If the honest P/E looks like the reported P/E, trust P/E. If it looks like the EV/EBITDA-implied multiple, trust that. If it looks like neither — which happens often — you've just learned that both shortcuts were wrong and you need to value the business on cash flow.

What to watch next

• Pull the D&A breakdown from the 10-K for any name where P/E and EV/EBITDA disagree by more than ~30%. Look for intangible amortization separately.
• Compare capex to D&A over a five-year average, not one year. Cyclicals will mislead you on a single year.
• Check the SBC line in the cash flow statement before quoting any tech-company EBITDA multiple. Recalculate with SBC subtracted and see how the multiple moves.
• Build a one-line cash-earnings number for every stock you own. You don't need a model — just EBITDA minus capex minus cash interest minus cash taxes minus SBC. Divide market cap by that. It's the single most honest valuation lens you can carry around.