Things to Consider About Per-Share Earnings

Imagine two students, and you have to guess which one is the stronger in maths. You're shown just one number for each: their score on a single test last week. Aarav scored 95. Ishaan scored 70. Easy - Aarav's better, right?

Now someone hands you the whole year's report card. It turns out Aarav scored 95 on that one test because it was on the exact chapter he'd revised the night before, and his other tests that year were 55, 60, 48, and 62. Ishaan's 70, meanwhile, was his lowest of the year - his others were 78, 82, 75, and 80. Suddenly the picture flips completely. Ishaan is the steady, strong student. Aarav had one lucky day. If you'd trusted the single test, you'd have judged both of them exactly backwards.

Companies have a number just like a single test score. It's called earnings per share, or EPS, and this whole chapter is about why you must never, ever trust just one year of it - for the very same reason you'd never judge a student by one lucky test.

Let's define it plainly, because the name sounds scarier than it is. A company earns a profit each year - the money left after all its costs. That profit belongs to all the owners, and ownership is split into little slices called shares. Earnings per share is simply that yearly profit divided up across all the shares - the slice of profit sitting behind each single share. If a company earns ₹100 crore of profit in a year and has 10 crore shares, then each share "earned" ₹10, so its EPS is ₹10. That's it. EPS is one share's slice of one year's profit.

It's a genuinely useful number. But - and here's the whole chapter - it is one test score, and one test score can be lucky, can be dressed up, and can lie about the student. Trust it alone and you'll judge companies exactly as wrongly as you'd have judged Aarav and Ishaan.

Here's the uncomfortable truth that makes this chapter matter: EPS is not a fact carved in stone. It's more like a photograph - and photographs can be posed, lit flatteringly, and cropped to hide the mess. The company itself does a lot of the choosing about how its profit gets counted, and that means a single year's EPS can be quietly made to look better than the real business deserves.

There are two big ways this happens. The first is one-off luck - something that puts money in the company's pocket this year but will never happen again. Suppose a company sells off an old building it owned and makes a big one-time gain. That gain gets added into this year's profit, so this year's EPS shoots up - but it tells you nothing about how well the actual business runs, and next year there's no building to sell. It's Aarav's lucky-chapter test: a real number, but not a repeatable one.

The second way is accounting choices. There are legal, allowed decisions a company makes about when to count certain costs or profits, and those choices can nudge EPS up or down. Nobody has to break a single rule. They just pick the flattering options, the way you'd angle a photo to hide a stain on the wall. The result is a shiny EPS that looks like strength but is partly just good posing.

So why does this matter so much to you, the saver? Because a huge number of people decide what a share is worth by staring at that single EPS number and nothing else. When the number is dressed up, they pay too much for the share, believing the business is stronger than it is. Then the one-off luck doesn't repeat, the flattering choices run out, and the number sags back to the truth - and the share sags with it. They didn't get unlucky. They trusted one test score and called it a report card.

And notice the deeper danger: the years EPS looks best are often exactly the years you should be most careful, because a suspiciously great single number is precisely when one-off luck or flattering choices are most likely hiding inside it. The moment a company's profit jumps in a way that seems too good, the right response isn't excitement - it's a raised eyebrow and a question: is this the real student, or just one lucky test?

Let's open up a single dressed-up EPS and see the pieces inside, so you can spot them yourself. Picture a company whose real, everyday business earns a steady profit worth about ₹8 per share - that's what it makes from actually doing its job, year in and year out. That ₹8 is the honest, repeatable part. Call it the engine.

Now, in one particular year, two extra things happen. The company sells an old factory it no longer needs and books a one-time gain worth about ₹5 per share. And it makes a couple of flattering accounting choices that push another ₹1 per share into this year. Add it all up, and the number it proudly reports is ₹8 + ₹5 + ₹1 = ₹14 per share.

Look at what the picture reveals. The ₹14 is real in the narrow sense - the company really did have that much profit this year. But only the ₹8 engine is the actual, repeatable business. The ₹5 from the factory sale will never come again; you can only sell that factory once. The ₹1 from flattering choices is just posing. So next year, with no factory to sell and the choices spent, EPS quietly sinks back toward ₹8 - and anyone who paid a high price believing "this company earns ₹14 a share!" suddenly owns a share that only ever really earned ₹8.

The skill, then, is learning to mentally strip the costume off a single big number and ask, "How much of this is the engine, and how much is one-off luck and flattering choices?" You won't always be able to split it exactly. But even asking the question protects you, because it stops you from swallowing the shiny ₹14 whole. A single big EPS is often part engine and part costume - before you trust it, ask how much of it will still be there next year when the one-off luck is gone.

Let's put rupees on it and watch someone get caught. illustrative

Two neighbours, Deepak and Haridya, are both looking at the same company. Its latest report proudly shows EPS of ₹14, up from a boring ₹8 the year before. The share is selling for ₹210.

Deepak sees the ₹14 and the rising share price and gets excited. He does a quick sum in his head: "₹210 for a share that earns ₹14 - that's only about 15 times its earnings, that's not expensive at all for a company growing this fast!" He puts in ₹42,000, buying 200 shares, feeling like he found a bargain rocket.

Haridya does one extra thing: she reads why the profit jumped from ₹8 to ₹14. Buried in the report, she finds it - the company sold a factory this year for a one-time gain, and that gain is most of the jump. The everyday business barely grew at all; it still earns around ₹8–9 per share from actually doing its job. So Haridya redoes Deepak's sum with the engine number: "₹210 for a share that really earns about ₹8 - that's more like 26 times earnings. That's not a bargain; that's a bit expensive." She passes.

Now play it forward one year. The factory can't be sold twice. With no one-off gain, the next report shows EPS back down around ₹8, right where the engine always was. Deepak, who bought believing the company earned ₹14, watches the "earnings" seem to collapse - and the share price, which had been resting on that ₹14 illusion, sags with it. He didn't buy a growing rocket. He bought one lucky test score and paid rocket prices for it. Haridya, reading the same report, saw the report card behind the single test and simply stepped aside. The difference between them wasn't cleverness or a tip. It was that one of them asked where the number came from and the other just swallowed it. A number like "₹14 a share" means nothing until you know the story behind it - a one-off factory sale and a real, growing business look identical on the surface and are worlds apart underneath.

So if one year can lie, what do you trust instead? The same thing you'd trust for the student: the whole report card. You look at several years of EPS together, and you average them, so that one lucky year and one unlucky year both melt into a truer picture. illustrative

Take a company whose EPS over five years reads: ₹6, ₹9, ₹15, ₹5, ₹8. Look how jumpy that is. If you'd met this company in the ₹15 year, you'd think it was a star. If you'd met it in the ₹5 year, you'd think it was struggling. Both single snapshots mislead. But add all five and divide by five, and you get an average of about ₹8.6 per share - and that number, the whole-report-card number, is a far honester picture of what this business really earns in a typical year.

Two things make this cure so powerful. First, averaging over years cancels out the noise. The lucky factory-sale year and the unlucky bad year pull in opposite directions and mostly cancel, leaving the steady engine showing through. Second - and this is the same idea from the last chapter - you check the profit against real cash. If a company keeps reporting nice EPS but the actual cash never seems to arrive, that's a loud warning that the earnings are more costume than engine, because real earnings eventually show up as real money in the bank.

There's a right span to average over, too. One or two years isn't enough, because good times and bad times come in waves - most businesses have fat years when everything sells and lean years when it doesn't. If your five years happen to be all fat, or all lean, even the average lies a little. The honest span is a full up-and-down cycle - enough years to catch at least one good stretch and one rough stretch - so the boom and the slump both get counted. For our jumpy company, the ₹6, ₹9, ₹15, ₹5, ₹8 already contains a high year and a low year, which is exactly why the ₹8.6 average feels trustworthy: it isn't the company on its best day or its worst day, it's the company across the whole weather. Judge what a company can really earn by averaging its profit across several years - ideally a full boom-and-slump cycle - never by one year alone, because any single year is puffed up or beaten down by the weather and by one-off items.

Notice something freeing here: you are not trying to nail the exact right number. Whether this company's true earning power is ₹8.4 or ₹8.9 per share doesn't really matter. What matters is knowing it's roughly ₹8–9 and definitely not ₹15 - because that rough-but-right judgement is enough to stop you overpaying, while a precise-looking ₹15 would have walked you straight off a cliff. Chasing false precision on one flashy year is worse than useless; a sensible range built from several years is what actually protects you. It's far safer to say "this earns roughly ₹8–9 a share, averaged over years" than to trust a precise, flattering "₹14" from one lucky year - a rough truth beats an exact illusion.

So far we've stared hard at the profit - the top of the EPS fraction. But remember what those three letters mean: earnings per share. There's a bottom to that fraction too - the number of shares - and here's the sneaky part almost everyone forgets: the bottom can grow, and when it does, your slice shrinks even if the company's total profit goes up.

Think of a pizza. The pizza is the company's whole profit; your share is one slice. If the pizza gets a little bigger next year but the company also cuts it into more slices, each slice can end up smaller than before. You're told "the pizza grew!" and it's true - but the slice on your plate got thinner. Total profit up, your piece down. That's called dilution, and it happens whenever a company creates new shares. illustrative

Take a company earning ₹100 crore of profit, split across 10 crore shares - so EPS is a tidy ₹10. Now next year the business genuinely does a bit better and earns ₹110 crore. Total profit up 10% - sounds great. But during the year the company also handed out a pile of new shares: some to its bosses as bonuses, some to raise fresh money, some from old promises that finally converted into shares. The share count swells to 12.5 crore. Now redo the slice: ₹110 crore ÷ 12.5 crore shares = ₹8.80 per share. The company earned more money in total and yet each owner's slice fell from ₹10 to ₹8.80. If you only watched the top of the fraction, you'd cheer the rising profit while your own piece quietly got thinner.

This is why a careful reader also asks a second question about share count: how many shares could there be? Companies often have promises floating around - stock options for staff, and "convertibles" (bonds or preferred shares that can turn into ordinary shares later). None of those are ordinary shares today, but every one of them is a slice waiting to be cut. So you count the shares that could exist once all those promises come due - the fully-diluted count - and measure your slice against that bigger number, not the flattering smaller one. It's the same spirit as the whole chapter: refuse the flattering version, insist on the honest one. What you own is measured per share, so watch the bottom of the fraction too - new shares from options, convertibles and fresh issuance can lift total profit while thinning each owner's slice, so always judge per share and count every share that could yet exist.

The slip is almost always the same: grabbing the single most recent, most flattering EPS and treating it as the whole truth - judging the student by one test because it happened to be the one on the table.

It sounds like "Earnings jumped to ₹14, the company's clearly booming!" - but you haven't asked whether that ₹14 is engine or one-off luck. It sounds like "The price is only 15 times earnings, that's cheap!" - but if the "earnings" are a dressed-up single year, your cheap price is measured against an illusion, and it isn't cheap at all. It sounds like "The report says profit is up, so the business must be stronger." - but profit can be up because of a sold building or a flattering choice, while the real engine sits exactly where it was. And the quiet one: "The exact number is right there in black and white, so it must be solid." - the precision of the printed figure is what fools you into skipping the question of what's inside it.