PE and PB, and the denominators nobody examines

What a price ratio actually asks

A PE ratio is the share price divided by the company's earnings per share. A PB ratio is the same price divided by its book value per share. Both answer exactly one question: what the market is paying today for one rupee of that reported figure.

Notice what the two have in common. The numerator is identical — the last traded price — and it is the only part of either ratio that is a fact. A price is what somebody actually handed over. It is not an estimate, it carries no policy choice, and it is remade every second the market is open.

The denominator is a different kind of object entirely. Earnings and book value are both produced once a quarter by people applying rules to judgements, and both could have come out differently without a single rupee moving. A price ratio is a fact divided by an opinion, and the article that stops at “PE is price divided by earnings” has taught you the easy half.

What the ratio buys you is comparability. Companies come in every size, so raw price tells you nothing — a ₹40 share is not cheaper than a ₹4,000 one, it is a different slicing of a different business, which is the point market capitalisation exists to make. Dividing by a per-share figure removes the size and leaves something you can line up.

What it costs you is everything the two inputs left out, which turns out to be most of what decides whether a business is worth owning. That is the trade-off, and the rest of this article is a tour of it.

The E is a set of choices, not a measurement

Profit is not counted the way cash is counted. Cash either arrived or it did not. Profit is the answer to “how much better off is this business over this period”, and answering it requires spreading costs over time, guessing at what will be collected, and deciding when a sale has happened.

Take depreciation, the plainest case. Say a company — an illustration, not a real one — builds a plant for ₹120 crore. Charged over an assumed ten-year life that is ₹12 crore a year against profit. Assume fifteen years instead and it is ₹8 crore. Four crore of profit appears out of an assumption, and no cash moved in either direction. On a share count of 10 crore, that is 40 paise of earnings per share, straight into the denominator of every PE anyone computes.

The same is true of a decision that looks purely technical: whether money spent building something gets charged to this year's profit or parked on the balance sheet and charged over several. An illustrative company spending ₹20 crore developing software can, if the criteria are met, carry it as an asset and charge perhaps ₹4 crore a year. Charge the whole ₹20 crore this year instead and reported profit is ₹16 crore lower and year-end book value is ₹16 crore lower too. Same cash out of the door, two different PEs and two different PBs.

Then there are the estimates that never settle: how much of what customers owe will actually be collected, what a long contract has earned so far, what a warranty or a legal claim will eventually cost. Each is a judgement made by people with a view on how this quarter should look.

And one-off items sit in the same line as the recurring ones. Sell a piece of land and the gain lands in profit; divide the price by that year's earnings and you have priced a transaction that happens once as though it happens every year. A ratio cannot tell a repeatable rupee from a one-time rupee — it just divides.

There is also a point where the ratio stops working at all. A loss-making company has a negative PE, which is not a cheaper share but an arithmetic artefact — a larger loss produces a more negative number, and nothing follows from that. Long before the sign flips, the scale comes apart: an illustrative company worth ₹500 crore earning ₹50 lakh carries a PE of 1,000, and at ₹5 lakh of earnings it is 10,000. A denominator near zero does not make the ratio large, it makes it meaningless, and PB behaves the same way once accumulated losses have taken book equity below zero.

None of this is an accusation. Accrual accounting is the only sensible way to describe a business that buys assets in one year and uses them for fifteen. The point is narrower: the E is an output of policy, so any comparison of PEs is partly a comparison of policies. Where each of those choices shows up in the statement itself is the subject of reading a profit and loss account.

Trailing, forward, and the quarter that rolls off

“The PE” is not one number. There are at least three in common circulation, and a page quoting one rarely says which.

Trailing uses earnings the company has actually reported, usually the last four quarters. It has the virtue of having happened. It also looks backwards by up to a year, and it carries every one-off in that window.

Forward divides today's price by somebody's estimate of next year's earnings. That is a different species of number. An estimate is a forecast, forecasts differ by whoever made them, and a forward PE is therefore as much a statement about the estimator as about the company. Comparing one company's forward PE with another's trailing PE is not a comparison at all.

The third is the trailing figure's least appreciated property, and it is worth seeing with numbers. Take an illustrative company whose last four quarters produced earnings per share of ₹8, ₹9, ₹10 and ₹11 — ₹38 in total. At a price of ₹760 the trailing PE is exactly 20.

Results are published. The new quarter comes in at ₹15, and the ₹8 from a year ago drops out of the window. The trailing total is now ₹45 and the PE is 16.9. The ratio fell more than three points without the price moving at all, because the denominator rolled over. Anyone watching the ratio rather than its inputs has just seen a share become “cheaper” on a day when nobody traded it differently.

Here is the mistake this produces, and it is common enough to be worth naming. You look up the same company on two websites, get two different PEs, and conclude one of them is broken. Usually neither is. One may be using earnings that include the subsidiaries — consolidated — while the other uses the parent company on its own; one may divide by the shares actually in issue while the other uses the diluted count, which also includes shares that options and convertibles could yet create; one may take the last four reported quarters and the other the last full financial year. A PE is a property of the calculation, not of the company, and the only way to compare two of them is to know that both were computed the same way.

The B is a residual, and the modern parts are missing from it

Book value is what is left when you subtract everything the company owes from everything it owns. It is the corporate version of net worth, and per share it is the accounting claim each share has on the business if the balance sheet were true and the whole thing were wound up at those values.

Start with that word, residual. Every asset on the sheet carries an estimate and so does every liability, and subtracting one from the other does not cancel the estimates — it collects them. Book value inherits the error of both sides, which is a strange foundation for a ratio people treat as the conservative one. What sits on each side, and how firm each line is, is set out in reading a balance sheet.

The larger issue is that assets are generally carried at what was paid for them, less what has been written off since. Land bought forty years ago sits at the forty-year-old price. Where the rules allow a revaluation and a company takes it, the same asset jumps; where they do not, or it does not, the same asset does not. Two businesses holding identical property can therefore report quite different books.

And then there is the part that never arrives on the sheet at all. A company that has spent two decades advertising has charged every rupee of it to profit as it went, so the brand that spending built is worth nothing on the balance sheet. The same goes for a trained workforce, a distribution relationship, and most software written in-house. The balance sheet records what was bought, not what was built.

Which produces a result worth sitting with. If one company builds a brand and a second buys the company that built it, the buyer carries the purchase price on its balance sheet as goodwill and the builder carries nothing. Identical assets in economic terms, wildly different books, wildly different PB ratios. The ratio is measuring how the asset was acquired.

Two more mechanics that move PB without anything happening to the business. A goodwill write-down cuts book value overnight and no cash moves, so the ratio rises. And a buyback shrinks equity: take an illustrative company with ₹100 crore of book equity across 10 crore shares, so ₹10 a share, trading at ₹40 — a PB of 4.0. It buys back 1 crore shares at ₹40, spending ₹40 crore. Equity is now ₹60 crore across 9 crore shares, or ₹6.67 a share, and at an unchanged ₹40 the PB is 6.0. Returning cash to shareholders raised the ratio by half.

The contrast worth holding on to comes from an entirely different corner of investing. A mutual fund's NAV is that fund's book value per unit, and it is computed from market prices every business day, which is why a fund does not trade at a premium or discount to it. A company's book is computed once a quarter from historical costs. Same idea, opposite relationship to reality — and that gap is the whole reason PB is interesting and the whole reason it misleads.

The two ratios, side by side

Setting the mechanics out in one place makes the shape of the problem visible: the numerator is shared, and every difference between the ratios is a difference between two accounting artefacts.

Read the last row twice. PB tells you what you are paying per rupee of accounting assets and says nothing about the return those assets generate; PE tells you what you are paying per rupee of accounting profit and says nothing about the assets it took to produce it. Each is precisely blind to the thing the other measures.

That blindness runs in opposite directions for the two kinds of business, which is the part worth carrying away. A company that owns a great deal reports a large book and therefore a small PB, whether or not any of those assets earns a rupee. A company whose value was built rather than bought reports almost no book and therefore a large PB, for a reason that has nothing to do with what anyone is paying.

The identity that connects them

The two ratios are not independent, and the link between them is one line of algebra that most explanations leave out.

Price divided by earnings, multiplied by earnings divided by book, gives price divided by book — the earnings cancel. The middle term, earnings divided by book equity, is return on equity. So PB equals PE multiplied by return on equity, always, as a matter of arithmetic rather than of finance.

Watch what that does to two illustrative companies. Company A earns ₹10 a share on ₹50 a share of book equity, a return on equity of 20%. Company B earns the same ₹10 a share on ₹200 of book, a return on equity of 5%. Price both at ₹200 and both carry a PE of exactly 20. But A's PB is 4.0 and B's is 1.0.

Same PE, four times the PB, and nothing separating them except how much capital each needed to produce its ₹10. Run it the other way and it is starker: put both on the same PB and their PEs must differ by the same factor of four.

So a low PB alongside a high PE is not two findings. It is one finding — a low return on equity — reported twice, and reading it as two independent signals double-counts a single fact. That is the arithmetic reason the two ratios so often seem to disagree about the same company.

One honest caveat, because the identity is exact only if you are careful. Return on equity is conventionally computed on average equity over the year, while PB uses the book value at a point in time. Feed a published return-on-equity figure into the identity and the answer will be close but not exact, and the gap is entirely down to which book figure each one used. Two ratios can only be reconciled on the same denominators, which is the article's argument in miniature.

Why a PE comparison across industries barely means anything

The ratio is price per rupee of accounting profit. Comparing PEs across industries assumes a rupee of accounting profit means the same thing in each. It does not, for at least four structural reasons.

Capital intensity changes what profit is struck after. Consider two illustrative companies each reporting ₹100 crore of profit: a toll road and a services business. The road's figure is after a large depreciation charge on an asset that may keep producing cash long after the accountants have finished writing it off. The services business owns almost nothing and has expensed nearly everything it spent on its own future. The two ₹100 crore figures have different relationships to cash and to the years ahead.

Capital structure sits above the earnings line. Interest is deducted before profit, so a heavily borrowed company reports lower earnings on the same operations, which raises its PE. Its book equity is also smaller, which raises its PB. Both ratios move with how the business was funded, and neither tells you that is what happened — which is why the funding mix gets its own measure in debt to equity.

Cyclicality inverts the ratio at exactly the wrong moment, and the arithmetic deserves to be seen. Take an illustrative commodity producer with revenue of ₹1,000 crore against largely fixed costs of ₹800 crore, so profit is ₹200 crore. At a market value of ₹4,000 crore its PE is 20. Now the commodity price rises 20%: revenue is ₹1,200 crore, costs are unchanged, and profit is ₹400 crore — it doubled on a one-fifth move. Suppose the price of the share rises 50%, to a market value of ₹6,000 crore. The PE is now 15. The ratio fell while the price rose by half, because operating leverage moved the denominator faster than the market moved the numerator.

That is the mechanism people are pointing at when they say a cyclical business looks cheapest at the top of its cycle. It is not a claim about any company, and it is not a rule about when to do anything. It is what happens to a fraction when the denominator is the more volatile of the two terms.

Growth and reinvestment are absent from the ratio entirely. Two businesses earning the same rupee today, one of which must spend heavily to earn it again next year and one of which need not, are not being described differently by a PE. The ratio has no slot for that information.

The same problem scales up to an index, which is where it quietly reaches ordinary investors. An index PE can be computed by adding every constituent's market value and dividing by the sum of their earnings, or by averaging the constituents' individual ratios, or by taking a median — and the three give different answers. A single large loss-maker drags the aggregate earnings down and pushes the aggregate ratio up, so whether it is included changes the headline. If you hold an index fund, the index PE quoted at you is a method as much as a measurement.

The number this article will not give you

So which PE is cheap? This article is not going to say, and the reason is not caution.

A ratio is a compression. It takes a business with a history, a balance sheet, a competitive position and an accounting policy, and returns one number. Turning that number back into a judgement requires everything the compression discarded — how repeatable the earnings are, what must be reinvested to repeat them, how much of the asset base belongs to lenders, whether the accounting is cautious or stretched, and what else the money could be doing. A threshold pretends the discarded information was not needed, which is the one thing we know is false.

There is a second reason, and it is the more interesting one. The ratio is not something the market has failed to notice. It is computed from two public numbers: a price the whole market set, and an earnings figure the company published to everyone at once. Every participant can see it.

So a low ratio is not evidence that a share is mispriced. It is a summary of what buyers and sellers already concluded looking at the same accounts you are looking at — that the earnings may not persist, or are not of good quality, or come with a risk that demands a larger discount. The ratio being low is the market's answer, not a question the market forgot to ask. Whether that answer is right is a separate enquiry, and it is not one a threshold performs.

One reframe does make the ratio easier to reason about, and it costs nothing. Turn it upside down: a PE of 25 is an earnings yield of 4%, and a PE of 12.5 is 8%. That puts the number on the same scale as the yield on a bond, which is how many people find it more intuitive. Then state the difference plainly, because the comparison hides it: a bond's coupon is a contractual obligation with a maturity date, while an earnings yield is this year's accounting profit with no promise attached to it — the same accounting profit this article has spent five sections qualifying.

FNOTrader is not a SEBI-registered investment adviser or research analyst, does not recommend shares, and nothing here is a view on any company or any level of any ratio.

Looking at the inputs instead of the ratio

Everything above points the same way: the ratio is the least informative object in the chain, because it is the end of a calculation whose interesting parts are its inputs. The practical move is to stop reading the output and go back a step — which earnings figure, over which period, on which share count, against which book.

FNOTrader's Stocks app screens across roughly 2,390 stocks and 17 NSE sector and size indices, and filters on the reported fundamentals rather than only on the ratios computed from them, so a screen can be built on the denominator you actually meant. The mechanics of writing one, and the ways a screen quietly answers a different question than you asked, are in the screener guide.

What no tool does is convert a ratio into a decision. It can put the same calculation across every company so that two numbers are at least comparable, and it can show you the inputs alongside the output. Reading them remains yours.

Common questions

Continue reading

More in Company Analysis · App: Stocks · Definitions: glossary · Free tools: calculators · All: every article