Multiples done right — and when each one lies

A discounted cash-flow model is honest but slow. So most investors reach, most of the time, for something faster: a multiple — a single ratio of price to some measure of the business, like price to earnings, used to compare one company against another at a glance. Multiples are everywhere because they are quick, and they are dangerous for exactly the same reason. A number you can read in a second is a number you can be fooled by in a second.

Here is the uncomfortable truth this module is built on: every multiple measures one slice of a business and is completely blind to the rest. Price-to-earnings sees the profit but not the debt. EV/EBITDA sees the operations but not the capex needed to sustain them. Price-to-book sees the balance sheet but not the returns it earns. None of them is "the" measure of value; each is a narrow window, and each has precise conditions under which it lies to you.

So the skill is not to memorise which multiple is "best" — there is no best. The skill is to know what each one actually measures, and to know, in advance, the exact situations where it will mislead you. Get that, and multiples become a fast, useful cross-check. Miss it, and they become the most reliable way for a bad business to look like a bargain.

It helps to see what a multiple really is. A price-to-earnings ratio is not a rival to discounted cash flow — it is a discounted cash flow compressed into one number, with every assumption hidden. When you say a company "deserves" a P/E of 20, you are making an implicit statement about its growth, its risk and how much cash it keeps — the very things a DCF makes explicit. The multiple just buries them where you cannot see them.

That is the danger and the usefulness in one. The danger: because the assumptions are hidden, you can apply a multiple without noticing the beliefs you have smuggled in — that this year's earnings are normal, that this company is like its peers, that its growth will resemble theirs. The usefulness: once you know what each multiple hides, you can use it as a quick sanity check on a slower DCF, and as a fast way to ask "what would have to be true for this price to make sense?"

Underneath every multiple sits Buffett's five-word discipline: price is what you pay, value is what you get. A multiple is a statement about price — the numerator is always the price, or something close to it. It only tells you about value if the denominator is an honest measure of what you get. Half the errors in this module come from a denominator that quietly is not.

Meet the four you will actually use, each with what it measures and the precise place it lies. illustrative

Price-to-earnings (P/E). Price per share divided by earnings per share. The P/E asks: how many rupees am I paying for one rupee of this year's profit? It is intuitive and universal — and it lies whenever the "E" is not normal. Cyclical peak earnings make a great year look permanent, so a low P/E flags "cheap" right at the top. One-off gains inflate the E and shrink the ratio. And because earnings sit after interest, a heavily indebted firm's P/E is not comparable to a debt-free one's.

EV/EBITDA. Enterprise value — the price of the whole business, equity plus debt minus cash — divided by EBITDA, earnings before interest, tax, depreciation and amortisation. The EV/EBITDA multiple values the entire operation before capital structure, so it lets you compare a debt-free firm and an indebted one on the same footing. Its lie is depreciation and capex: EBITDA pretends the cash a business must spend every year to keep its assets running does not exist. For a capital-heavy firm, a low EV/EBITDA can sit on top of almost no free cash flow — which is why Charlie Munger dismissed EBITDA as, in effect, earnings before the costs that are real.

Price-to-book (P/B). Price per share divided by book value per share — the accounting net worth of the company. The P/B compares price to what the balance sheet says the equity is worth. It is most meaningful for banks and asset-heavy firms, where book value roughly tracks the real economic capital. It lies badly for asset-light businesses — a software or brand-led company's real worth lives in intangibles the balance sheet barely records, so its book value understates it and its P/B looks absurdly high. And P/B means nothing without return on equity: a low P/B on a business earning poor returns on that book is cheap for a reason.

Price-to-sales (P/S). Price divided by revenue. The P/S is used when there are no earnings yet — a young, fast-growing, loss-making firm — because sales exist even when profit does not. That is also why it is the most treacherous of the four: sales say nothing about whether those sales make money or devour capital. Two firms on the same P/S can be worlds apart if one earns a 20% margin and the other loses money on every rupee.

The deepest split among these is whose claim they measure, and a picture makes it unforgettable.

Watch two companies with identical operations score differently, purely because of debt. This is the single most useful worked example in relative valuation. illustrative

Both firms have an enterprise value of ₹10,000 crore and EBITDA of ₹1,000 crore — so both trade at EV/EBITDA of 10×. Their operations are twins. Now walk each down to earnings.

Firm A — no debt. From EBITDA ₹1,000 cr, subtract depreciation of ₹200 cr → operating profit ₹800 cr. No interest. Tax at 25% takes ₹200 cr → net profit ₹600 cr. With no debt, equity value equals enterprise value, ₹10,000 cr. So:

P/E of Firm A = ₹10,000 cr ÷ ₹600 cr ≈ 16.7×

Firm B — ₹4,000 cr of net debt. Same EBITDA ₹1,000 cr, same depreciation ₹200 cr → operating profit ₹800 cr. But interest at 10% on ₹4,000 cr takes ₹400 cr → pre-tax profit ₹400 cr. Tax at 25% takes ₹100 cr → net profit ₹300 cr. And equity value is enterprise value minus the debt: ₹10,000 − ₹4,000 = ₹6,000 cr. So:

P/E of Firm B = ₹6,000 cr ÷ ₹300 cr = 20×

Nothing about the business differs — the identical EV/EBITDA says so. Yet judged on P/E, Firm B looks a fifth more expensive. If you had ranked the two on P/E alone you would have called the leveraged firm dearer and the plain one cheaper, when their operations are the same and the only difference is a financing choice. Reverse the trap and it bites the other way: leverage can make a firm's P/E look lower and "cheaper" while quietly loading it with risk. The multiple you reach for silently decides the verdict.

Even used carefully, multiples have hard limits that no amount of cleverness removes.

A multiple cannot tell you whether the denominator is honest. P/E is only as good as the "E," and earnings are the most manipulable line in the accounts — one-offs, aggressive revenue timing, capitalised costs. A tidy multiple built on a dishonest denominator is precise nonsense. The multiple cannot check its own inputs; you must.

A multiple cannot supply the peer group. "Cheap versus the sector" assumes the sector truly compares — same growth, same capital intensity, same risk. Choose the wrong peers and any stock can be made to look cheap or dear. The comparison is a judgement you bring to the multiple, not something the number provides.

A multiple flattens the future into the present. It takes one year's earnings or sales and treats it as representative of a whole future. For a stable business that is fine; for a cyclical, a turnaround, or a fast-grower it can be wildly wrong, because the single year in the denominator is exactly the thing that will not persist.

Because of all this, a multiple is best held as a rough cross-check, not a precise verdict — a fast way to ask a question, which then sends you back to the business and, often, to a slower DCF, for the answer.

The same handful of multiple mistakes catch investor after investor.

1. "Low P/E means cheap." The most expensive four words in valuation. A low P/E can mean peak cyclical earnings, a one-off gain, heavy debt, or a decline the market has already seen. Always ask why the multiple is low before calling it a bargain.

2. Trusting EV/EBITDA on a capital-hungry firm. EBITDA ignores the capex a business must spend every year just to stand still. For asset-heavy firms a low EV/EBITDA can sit atop almost no free cash — check the capex before you trust the multiple.

3. Using P/B on an asset-light business. Book value barely records brands, software and know-how, so an intangible-rich company's P/B looks sky-high and an asset-heavy one's looks low, telling you about the balance sheet's shape, not the price's fairness. And never read P/B without return on equity beside it.

4. Treating P/S as safety. Sales are the loosest possible denominator — they say nothing about margin or capital needs. A "low" P/S on a business that loses money on every sale is not cheap; it is a warning wearing a friendly number.

5. Comparing across capital structures with an equity multiple. Ranking a debt-free firm against a leveraged one on P/E, as the worked example showed, mixes operating quality with financing choices. To compare the businesses, move up the stack to EV/EBITDA; to judge the equity, read P/E with the debt in full view. A multiple needs the story of why it is what it is before it means anything at all.