Valuation ratios

Two companies trade on the same price-to-earnings multiple of twenty. On any screen sorted by P/E, they are equally valued, equally "expensive." But one of them grows its earnings at five percent a year and the other at twenty-five. In a few years the fast-grower's earnings will have multiplied while the slow-grower's have barely moved — so paying twenty times for one is dear and paying the same twenty times for the other can be a bargain. The multiple was identical; the value could hardly be more different. illustrative

A valuation ratio is the most compressed number in all of analysis, because it folds two things together: the price the market is paying, and — buried inside that price — everything the market expects. A P/E of twenty is not just "twenty rupees of price per rupee of earnings"; it is a statement that the market expects those earnings to grow, and to be real, and to last, enough to justify twenty. So a multiple is never simply high or low; it is high or low relative to what must come true to deserve it. This module is about the family of valuation multiples — P/E, EV/EBITDA, price-to-book, price-to-sales, free-cash-flow yield — what each one assumes, when each is the right tool and when it goes dark, and how to read every one of them as a statement of expectations rather than a verdict on cheapness.

Every valuation multiple divides what you pay by some measure of what you get — but there are several measures of "what you get," and each choice carries an assumption that can be right or wrong for the business in front of you. Price-to-earnings divides the share price by earnings per share, and assumes those earnings are real, representative, and a fair base for the future — which is why it breaks for a cyclical at its peak, a loss-maker, or a firm whose earnings are flattered by one-offs. EV/EBITDA and EV/EBIT divide enterprise value — the market value of the equity plus net debt — by a pre-interest profit, so they price the whole business independently of how it is financed, which makes them the fair way to compare two firms with very different debt. Price-to-book divides the price by net asset value, and is the natural anchor for a business whose value is its balance sheet — a bank, an insurer — and near-useless for an asset-light firm with almost no book. Price-to-sales divides by revenue, a crude measure used mainly where there are no earnings yet, as with an early-stage platform. And free-cash-flow yield turns the whole thing around — free cash flow divided by market value — to ask the most honest question of all: what cash return is this business actually throwing off for its price?

The reason there are so many, and the reason this module exists, is that each multiple answers the valuation question through a different lens, and picking the wrong lens produces confident nonsense — an EV/EBITDA on a bank, a P/E on a loss-maker, a price-to-book on a software firm. And underneath all of them sits the single idea that makes valuation make sense: a multiple is a claim about the future dressed as a fact about the present. A high multiple is not "expensive" in the abstract; it is the market pricing in growth and durability that then have to be delivered. Reading a multiple well means reading the expectation inside it — which is exactly what the later part on why prices move builds into a full skill.

Pick the numerator, pick the denominator, and know what the pairing assumes.

Price versus enterprise value. The first fork is what you are valuing. A price-based multiple (P/E, P/B, P/S) values the equity — the owners' slice. An enterprise-value multiple (EV/EBITDA, EV/EBIT) values the whole firm, because enterprise value is market cap plus net debt, and it is paired with a profit measured before interest. The practical rule: when you are comparing two businesses with very different debt levels, use the enterprise-value multiples, because P/E is contaminated by financing — a heavily-indebted firm's earnings are struck after heavy interest, so its P/E reflects its balance sheet as much as its business. When leverage is similar, P/E is simpler and fine.

Match the denominator to the business. Earnings (P/E) work for a stable, profitable firm whose earnings mean something. EBITDA-based multiples suit capital-heavy businesses where you want to compare operating cash generation before differing depreciation and financing — but remember EBITDA flatters an asset-heavy firm by ignoring the depreciation that is a real economic cost, so EV/EBIT (which charges depreciation) is often the truer measure for a business that genuinely consumes its assets. Book value (P/B) is the anchor where the balance sheet is the business — read it always against ROE, since a bank earning 18% on equity deserves a higher P/B than one earning 8%. Sales (P/S) is a last resort for firms with no earnings yet, and dangerous precisely because sales say nothing about whether those sales will ever be profitable. And free-cash-flow yield is the one that cannot be gamed by accounting choices, because it is built on cash — the most honest of the family for a mature, cash-generative business.

Read the expectation, not the level. A multiple's level is meaningless without the growth, quality and durability it is pricing. The PEG idea makes this explicit for P/E — a P/E of 30 on 30% growth (PEG of 1) is very different from a P/E of 30 on 8% growth (PEG near 4) — but the principle runs through every multiple. A high EV/EBITDA prices in years of growing cash flow; a low price-to-book prices in a return on equity the market does not believe will hold. So the question is never "is this multiple high or low?" It is "what does this multiple require the business to deliver, and is that likely?" A high multiple is a high bar the company has now promised, on your behalf, to clear.

Which multiple to reach for is set almost entirely by the sector, and the same multiple that anchors one business is the wrong tool for the next. Here are four businesses and the multiple each is properly valued on — with the point that importing the wrong one produces a confident, useless number.

The thread is that a valuation multiple is only a question worth asking if it is the right multiple for the business. For a bank, enterprise value is a category error, so price-to-book against ROE is the only sensible anchor. For a capital-heavy operator, enterprise-value multiples are essential because they neutralise the differing debt, but EBITDA flatters and EV/EBIT is the truer read. For a stable consumer firm, P/E and free-cash-flow yield fit cleanly, and the whole question becomes whether the growth justifies the premium. For an early-stage loss-maker, there is no honest earnings multiple at all, and price-to-sales is a placeholder pricing a profitability that may never come. In every case the discipline is the one this whole part has taught: know the business first, pick the multiple its economics actually fit, and read the expectation the multiple is pricing rather than calling the number "cheap" or "dear" on its own. This module hands the next part — why prices move — the raw material it needs: the multiple is where the market's expectation is written down.

A composite branded consumer company trades at a P/E of 55, and two readers look at the same number. One says "far too expensive — 55 times earnings is absurd" and passes. The other says "quality compounder, worth every rupee" and buys. Neither has actually read the multiple; both have reacted to its level. Read the expectation inside it instead. illustrative

A P/E of 55 is a statement about the future, so unpack what future it requires. Turn it around: to earn a reasonable return from here, the earnings have to grow fast enough, for long enough, to bring that multiple down to something ordinary over your holding period. Do the rough arithmetic — if the business can compound earnings at, say, 20% a year for a decade while sustaining its high return on capital and its cash conversion, then today's 55 is demanding but not absurd, because the earnings will have multiplied several times over. If, more soberly, the realistic growth is 12%, then 55 is pricing in a decade of delivery the business is unlikely to manage, and the multiple is genuinely stretched. The number 55 did not answer the question; the growth-and-durability it requires did, and that is a judgement about the brand's moat, its runway, and its pricing power — not about whether 55 sounds high.

Now cross-check with a different lens, because a single multiple can mislead. The free-cash-flow yield: at 55 times earnings with high cash conversion, the FCF yield is only around 1.5–2%, which tells you plainly that you are being paid almost nothing in current cash and are buying entirely on future growth — a useful reality check on the "worth every rupee" enthusiasm. And the PEG: 55 divided by 20% growth is a PEG near 2.7, on the expensive side even for a quality name. Put together, the multiple is not "absurd" and not "worth every rupee" — it is a specific, demanding bet that this business compounds at a high rate for a long time, and whether to make that bet depends on how confident you are in the runway, not on how the number 55 feels. The habit to build is to convert every multiple into the expectation it encodes, then judge the expectation.

A valuation multiple cannot tell you whether the expectation inside it is correct — only what that expectation is. A P/E of 40 tells you the market is pricing in strong growth; it cannot tell you whether that growth will actually arrive. Deciding whether the expectation is too optimistic or too pessimistic is the work of building your own view of the business's future and comparing it with the market's — a judgement the multiple frames but does not make. The number reveals the bet; whether the bet is good is a separate, harder question, and it is the whole subject of the part on why prices move.

Nor can a multiple see the quality of the earnings or assets it divides by. A P/E rests on earnings that may be flattered by aggressive accounting or one-offs; a price-to-book rests on a book value that may carry impaired assets or hidden liabilities; an EV/EBITDA rests on an EBITDA that ignores real depreciation and can be inflated. A multiple computed on a manipulated denominator is precise and worthless, and the multiple itself gives no warning. Whether the earnings and assets underneath are honest is a question for the cash-quality checks and the forensics part, not for the valuation ratio.

And multiples are only comparable within a genuinely comparable set. The same P/E means different things for a cyclical and a stable earner; the same price-to-book means different things for a bank earning 18% on equity and one earning 8%; the same EV/EBITDA means different things for a firm with light capex and one that must constantly reinvest just to stand still. Ripping a multiple out of its sector and its growth-and-return context to rank it against unlike companies — exactly what a screen does — produces a false ordering. The multiple supplies the market's expectation; it cannot supply the comparison that makes the expectation meaningful. That you must bring.

How it comes up. Management rarely discusses its own multiple directly, but the expectation inside the multiple is exactly what a sharp analyst tests — by pressing on the growth and durability the valuation is pricing. The question sounds like this: "The stock is on 45 times earnings, which implies years of 20%-plus growth. What's the realistic runway, and what return on capital can you sustain as you scale?" The analyst is checking whether the business can deliver what the multiple already assumes.

A good answer, verbatim-style.

"I won't comment on the share price, but on the fundamentals: our addressable market is still under-penetrated, and we see a long runway for double-digit volume growth, with pricing on top. We'd expect return on capital to stay in the high twenties even as we invest, because the model is asset-light. Whether that justifies any particular multiple is for the market — but the growth and the returns are what we're managing for."

That answer engages the substance the multiple is pricing — the runway, the growth, the sustainability of returns — without pretending to justify a number, which is exactly the honest posture.

An evasive answer, verbatim-style.

"We think the market rightly recognises the quality of our franchise, and we're confident the stock will continue to reward long-term shareholders. Our premium valuation reflects our premium business, and we see no reason that should change."

Notice the move. It endorses the multiple as self-justifying — "premium valuation reflects premium business" — instead of addressing the growth and returns that would actually have to materialise. A management talking up its own multiple rather than the delivery beneath it is a small red flag in itself.

The follow-up nobody asks, and what its absence means. "At the current multiple, the market is implying roughly X% earnings growth for a decade — do you think that's achievable, and what breaks it?" That converts the valuation into an explicit growth expectation and asks management to own it or push back. If the room lets "premium business, premium valuation" stand without translating the multiple into a required growth rate, either the implied growth is higher than management privately believes achievable, or nobody is checking the expectation the price already embeds. A high multiple that no one converts into a delivery requirement is a bet whose terms have gone unread — which is the setup the next part exists to expose.

1. Reading a multiple as cheap or dear on its level. A multiple is a claim on the future; the same P/E is expensive on slow growth and cheap on fast growth. Read the growth and durability it prices, not the number alone.

2. Using P/E to compare differently-levered firms. Earnings are struck after interest, so a heavily-indebted firm's P/E reflects its financing. When leverage differs, the enterprise-value multiples (EV/EBITDA, EV/EBIT) are the fair comparison.

3. Trusting EBITDA multiples on asset-heavy businesses. EBITDA ignores depreciation, which is a real cost for a firm that consumes its assets. EV/EBIT, which charges depreciation, is often the truer read for capital-heavy operators.

4. Applying the wrong multiple to the sector. EV/EBITDA on a bank, P/E on a loss-maker, price-to-book on an asset-light firm — each is a category error. Match the multiple to the business the sector modules describe.

5. Reading a cyclical's trailing P/E straight. At the peak a low P/E sits on record earnings and signals expensive; at the trough a high P/E sits on depressed earnings and can signal cheap. The multiple inverts against the cycle; use book value or mid-cycle earnings.

6. Trusting a multiple without checking its denominator. A P/E on flattered earnings, a P/B on impaired assets, an EV/EBITDA on inflated EBITDA — each is precise and worthless. Cross-check with free-cash-flow yield, which is built on harder-to-fake cash.