EV/EBITDA vs P/E Ratio: Which One to Trust

Put two cement companies side by side. JK Cement trades at a P/E of 41; Shree Cement at 53 (market data, Jun 2026). Cheaper on earnings, by a fifth — the kind of gap that gets a stock onto a "value" screen. Except the gap is mostly an accounting mirage. JK carries about ₹2,800 crore of net debt; Shree is sitting on roughly ₹6,000 crore of net cash (company concall figures, FY26). Fold each one's balance sheet into the price you'd actually pay to own the whole business, and the discount evaporates: on EV/EBITDA they come out at 18.4 and 18.8 — a rounding error apart. That is the central flaw in the P/E ratio in one example. It sits below the interest line, so it quietly rewards a company for borrowing and punishes one for hoarding cash — financing decisions dressed up as valuation. EV/EBITDA was built to strip that out. Knowing when its neutrality matters, and the one place EBITDA itself becomes a trap, is what this field guide is for.

(These named companies are illustrations of how the two multiples behave, not a view on either stock.)

What do these two ratios actually measure?

They answer different questions, and the difference is the whole point.

The structural difference: P/E is an equity-holder's view; EV/EBITDA is a whole-business view. P/E asks "what am I paying for my share of the profits after debt is serviced?" EV/EBITDA asks "what am I paying for this business's operating engine, regardless of who financed it?" The second question is capital-structure neutral — it strips out the leverage that distorts the first.

Why does P/E get distorted by debt?

Go back to the two cement makers from the opening, because they make the mechanism concrete in a way no invented example can. Cement is the right place to look: it's brutally capital-intensive, the product is near-identical from plant to plant, and leverage varies wildly between otherwise similar peers. Here is what the two multiples say about the same trailing-twelve-month operating performance, using Inve's quarterly financials for EBITDA and the prevailing market price.

(Illustration of multiple behaviour, not a recommendation on either company.)

Look at the P/E first. JK Cement, the one carrying the debt, shows the lower P/E — 41.4 against 53.4 — and a naive screen flags it as the cheaper stock. But that "discount" is doing something sneaky: P/E divides the price of the equity by profit that has already had JK's interest bill subtracted, while saying nothing about the ₹2,800 crore of borrowings an owner inherits — the kind of leverage that, if cash flow ever falters, is the first ingredient of a debt trap. Shree, debt-free and sitting on a cash pile, gets no credit in its P/E for that cleaner balance sheet. Add each company's net debt (or subtract its net cash) to reach enterprise value, and the comparison flips onto a level field: 18.4 versus 18.8, effectively the same price for the same operating engine. The leverage that flattered JK's P/E is exactly the thing EV/EBITDA refuses to ignore — it's part of what you're buying.

Here's the homely version. Two flats in the same building, same size, same view, both "priced" at ₹50 lakh. One comes with a ₹20 lakh home loan you must take over; the other is freehold with ₹5 lakh of maintenance money left in an escrow account. Quoting only the ₹50 lakh sticker — the P/E — tells you nothing about which is the better deal. The all-in cost of owning each flat — the EV — is the number that matters, and it's ₹20 lakh apart. P/E quotes the sticker; EV/EBITDA quotes the all-in cost.

There's a second lesson hiding in the last row, and it's the one most P/E screeners miss. JK shows a 15.1% return on capital employed against Shree's 10.5%. So JK's lower multiple may not be purely a debt illusion — part of it might be deserved the other way, if JK is genuinely the more efficient deployer of capital. But be careful with that single number: this is one trailing ROCE reading, and ROCE in cement is deeply cyclical — it can also be flattered by an under-depreciated or recently revalued asset base. A return that looks high in one good year can be ordinary across a full cycle, so ROCE has to be read across several years, not a single trailing figure, before you let it "justify" anything. Treat it as a question to investigate, not an answer: a low multiple is sometimes a bargain and sometimes a verdict, and you can only tell which by reading the multiple against the return that earns it — over time.

One more fence around that whole table, because it is easy to misread. A single-period EV/EBITDA and one ROCE reading do not establish that either business is cheap, or that either is safe. Nothing here says JK is the better company to own, or Shree the worse — the point is only that the two multiples measure different things. Leverage that "looks reasonable" on today's numbers is the very first thing that hurts the day cement demand turns: borrowings stay fixed while EBITDA falls, and a debt load that seemed comfortable at the top of the cycle can become the whole problem at the bottom. Read the table as a teaching device built on one TTM snapshot, not as a view on either stock.

When should you trust EV/EBITDA over P/E?

EV/EBITDA is the better lens precisely where capital structure and accounting choices muddy the P/E.

When debt levels differ across the companies you're comparing. This is the headline case — the cement pair above is the whole argument in miniature. Telecom, infrastructure, real estate, capital goods, utilities, cement — sectors where leverage varies hugely between peers. EV/EBITDA neutralises the difference; P/E doesn't.

For capital-intensive businesses with heavy depreciation. Depreciation is a non-cash charge driven by accounting policy as much as economics. Two firms with identical operations but different depreciation schedules show different net profits and different P/Es. EBITDA, sitting above depreciation, sidesteps the distortion. This is also why the EV family itself splits by capital intensity: for non-capital-intensive businesses the cleaner choice is EV/EBIT (enterprise value to operating profit), while EV/EBITDA is specifically the capital-intensive multiple — because differences in depreciation method and historical asset cost distort EBIT, the very thing EBITDA strips out.

For cross-border or acquisition analysis. When you're effectively asking "what would it cost to buy this whole business?", enterprise value is the right numerator because an acquirer assumes the debt. This is why M&A and private-equity valuation runs on EV/EBITDA, not P/E.

When earnings are near zero but operations are healthy. A company in a heavy-investment phase can show negligible or negative net profit (useless P/E) while generating solid EBITDA. EV/EBITDA still functions.

When does EBITDA itself become a trap?

Here's where a field guide has to be honest, because EV/EBITDA's strength is also its weakness. EBITDA strips out interest, tax, depreciation, and amortisation — and in doing so it strips out real costs.

EBITDA ignores the cost of staying in business. Depreciation is a non-cash charge, but the capex it represents is brutally cash. A cement plant or a textile mill "earns" healthy EBITDA while constantly spending to replace wearing assets. For a genuinely capital-intensive firm, EBITDA flatters cash generation, sometimes grotesquely. Charlie Munger called EBITDA "bullshit earnings" for exactly this reason — it's the profit you'd have if equipment never wore out and debt never came due. It does and it does.

EBITDA is the line managements most love to guide on. Because it sits above the messy stuff — interest, one-offs, the capex bill — EBITDA and EBITDA margin are the metrics management commentary leans on hardest. And a margin target is the easiest kind of guidance to give and the easiest to walk away from.

Watch it happen. In its July 2020 earnings call, the textile maker Trident told investors it would "endeavour to achieve EBITDA margins around 18% to 20% on a sustainable basis" (Trident Q1 FY21 concall), and reiterated the line a quarter later (Trident Q2 FY21 concall). For that one year it more or less delivered — Trident's own FY22 annual report records an FY21 EBITDA margin of 18% (Trident FY22 annual report, 10-year highlights). Then the "sustainable" part went quiet. The annual margin was never compared back to the 18–20% target on later calls, and the number drifted: across Trident's most recent four quarters the EBITDA margin has run at about 12.8% (Inve data, TTM to Dec-25). Inve's Promise Tracker has the commitment logged from "new" in Q1 FY21 to "on-track" in Q2 FY21 to ghosted by FY22 — a target met for a year, quietly dropped, and never reconciled as it fell more than five points below the 18% floor management called sustainable.

(Trident is an illustration of how margin guidance behaves over time, not a recommendation.)

That word — sustainable — is the tell. An EV/EBITDA that looks cheap because you've pencilled in a "sustainable" 18% margin is only as good as that margin, and a margin a management stops repeating is a margin worth distrusting. Contrast it with a commitment that held: Dalmia Bharat told investors it would keep net debt below 2× EBITDA, and not only kept it — it reported 0.46× at the end of FY26 (Dalmia Bharat Q4 FY26 concall), comfortably inside its own line. Both are real managements; one guidance held and one faded. You cannot tell which from the multiple. You can only tell from the record.

So EV/EBITDA fixes the leverage distortion of P/E but introduces a cash-conversion blind spot of its own. The fix for the fix is to follow EBITDA down to free cash flow, and to check whether the EBITDA guidance management gave last year actually showed up — or quietly went silent like Trident's.

How should you actually use the two together?

Don't pick a winner — use them as a cross-check. The most informative reading comes from where they disagree. JK versus Shree is the worked example: the disagreement (cheap on P/E, level on EV/EBITDA) was the insight.

A practical sequence, ending on the owner's question:

1. Screen on EV/EBITDA when debt varies. It's the fairer first cut across a leveraged sector — cement, telecom, infra.
2. Look at the P/E–EV/EBITDA divergence. If a stock looks cheap on P/E but average on EV/EBITDA — exactly JK Cement's profile — leverage is flattering the P/E; proceed with the debt in full view. If it looks expensive on P/E but cheap on EV/EBITDA, heavy depreciation may be hiding real cash generation. And read each multiple against its companion variable — P/E against growth or ROE, EV/EBITDA against ROCE — because a low multiple can be deserved rather than a bargain: JK's lower P/E may be partly its higher ROCE (15.1% vs 10.5%) earning the right to it, not just its debt distorting it — but confirm that ROCE holds across a full cycle, not a single trailing year, before you treat it as deserved.
3. Drag EBITDA down to free cash flow. Subtract interest, tax, and maintenance capex. A wide EBITDA-to-FCF gap means the EBITDA multiple is flattering reality — common in capital-intensive sectors, and a classic quality-of-earnings red flag when reported profit keeps outrunning cash.
4. Audit the EBITDA-margin guidance. Has management hit the EBITDA targets it set, or guided high — "on a sustainable basis" — and let them lapse, the way Trident's 18–20% target faded to a sub-13% reality with no reconciliation? An attractive EV/EBITDA built on a margin target with a history of going silent is a softer number than it looks.
5. End on the owner's question: after the debt is serviced and the assets are maintained, how much cash actually reaches the owner — and is this management's record of delivering its margin commitments good enough to trust the forecast inside the multiple?

Steps 1–3 are arithmetic. Step 4 — tracking whether EBITDA and margin guidance is delivered or quietly dropped, quarter after quarter across a portfolio of ten or fifteen stocks — is the part nobody does by hand, and it's what Inve's Promise Tracker is built to hold. The Concall AI guidance tables show whether each margin commitment is being raised, maintained, or walked away from over time — the difference between Dalmia's kept net-debt line and Trident's faded margin target, laid out call by call.

Frequently asked questions

P/E and EV/EBITDA aren't rivals — they're two windows into the same building, and the view from each is incomplete. P/E shows you the equity holder's slice but lets leverage distort the picture, the way it made JK Cement look a fifth cheaper than Shree on earnings while the two were priced the same as businesses. EV/EBITDA neutralises that leverage but lets capex and interest hide off-screen. Trust the one that fits the situation, distrust either when it disagrees with free cash flow, and never forget that the EBITDA margin in the denominator is often a management commitment — and a commitment a management gives "on a sustainable basis" and then stops mentioning, the way Trident did, is worth exactly as much as its record.