Why does a stock get more expensive when its earnings fall?

Vinati Organics makes speciality chemicals. In the year to March 2020 (FY20) it earned about ₹32.5 of profit per share. The next year, FY21, its plants were shut for weeks in the COVID lockdown and demand for one of its key products softened. Earnings per share fell to about ₹26.2, a drop of roughly 19%.¹

A beginner would expect the stock to fall with the profit. It did the opposite. After the March 2020 crash it recovered and went on to new highs. So the market was paying a higher price relative to the earnings the business had just reported. How?

The answer is that the price of a stock is not the earnings. It is the earnings multiplied by something else. And that something else moved far more than the earnings did.

We can take any stock price and split it into two numbers: the earnings the company has already reported per share, and the multiple investors are paying for those earnings. Price equals EPS times P/E.

That multiple is the price-to-earnings ratio, the P/E. It is simply how many rupees investors pay today for one rupee of the company's trailing earnings. A P/E of 30 means investors are paying ₹30 for every ₹1 of those trailing earnings.

One caveat, and it matters. This is an identity, not a theory of how the market sets prices. The P/E is defined using the price, so we are not claiming EPS and P/E cause the price. We are only decomposing it. The usefulness is exactly that: it lets us separate how much of a move came from earnings changing and how much came from the valuation changing. Read it slowly, because everything else here is a consequence of it: if the price moves, one of the two on the right moved, and it is worth knowing which.

Because price is earnings times a multiple, a rising stock can be doing one of two very different things. Either the earnings grew, or the multiple people pay for those earnings grew. Usually it is some of both.

This gives you a framework you can point at any stock, ever, and it is exact, not a rule of thumb. The ratio of the new price to the old is the ratio of the new EPS to the old, times the ratio of the new P/E to the old: P1/P0 = (EPS1/EPS0) times (PE1/PE0). In plain words, the change in your stock price is earnings growth combined with multiple expansion or contraction. So if a company grows profit 50% and the market also pays a 50% richer multiple, you do not make 50%, you make 125%, because 1.5 times 1.5, minus 1, is 125%. And if profit grows 50% while the multiple halves, you make almost nothing. The business did well and you still went nowhere.

This is why two investors can look at the same wonderful company and one triples his money while the other loses. They did not disagree about the business. They paid different multiples going in, and the multiple is half the engine.

Now the crucial part. Why would the multiple rise exactly when earnings fall? Because the two numbers are looking in opposite directions in time.

Reported EPS is backward-looking. It is a photograph of a year that has already ended. The FY21 EPS is a fact about the past, and COVID is baked into it forever.

The price is forward-looking. Nobody buys a share for the profit the company made last year. They buy it for the stream of profits it will make in all the years ahead. The market can see a recovery coming months, sometimes years, before that recovery ever shows up in a reported number.

So in FY21 the market was looking past a temporarily damaged year. It saw a chemicals business with new capacity coming on and demand returning, and priced the future it expected, not the dented present it was handed. Earnings looked backward at the damage. Price looked forward past it. The gap between those two views is the expanding multiple.

Put the pieces together on the real company. FY21 was Vinati's weak year: plants shut in the lockdown, softer demand, EPS down about 19% to ₹26.2. That is the denominator of the P/E falling.

Meanwhile the price, after crashing with the whole market in March 2020, recovered and pushed to new highs. That is the numerator rising. By the identity, if the price is up and EPS is down, the P/E must have risen by even more than the price did. The P/E expanded from both ends at once.

And the recovery the market was pricing eventually arrived. Earnings snapped back the very next year: FY22 EPS returned to about ₹33.7, and FY23 rose further to about ₹40.8, comfortably past the pre-COVID level.² The investors paying up in FY21 were, it turned out, paying for a recovery that arrived.

A rising P/E is not one thing. It is two things that look identical on a screener and mean opposite things.

The first is price-led. The price rises faster than earnings, so the market is paying more for each rupee of earnings. This is a real re-rating: a change of opinion about the future. It might be right or wrong, but the market is now willing to pay more for the same rupee of earnings.

The second is earnings-led, and it is a trap for the unwary. Earnings fall faster than the price, so the P/E rises mechanically. Nobody re-rated anything. The denominator just shrank. A stock can look more and more 'expensive' on trailing P/E while the market is actually getting more worried about it, not less.

Vinati in FY21 had both working the same way: the price rose and earnings fell, so the multiple was lifted from the top and the bottom at once. But you must always ask which term is moving, because the same number, a higher P/E, can mean 'the market loves this' or merely 'last year was bad'.

Most people are taught a single sentence: a P/E rises when the market expects growth. It is not wrong, but it hides the most useful distinction in valuation. There are really three separate things at work here, and it pays to keep them in two boxes, not one.

Two of them are engines of your return. Earnings growth: the company genuinely earns more. And multiple expansion or contraction: the market chooses to pay more, or less, for each rupee of earnings, a real re-rating. Those two, multiplied together, are your entire return, exactly as the identity above showed.

The third is different, and it is the one that fools people. Call it denominator distortion: earnings fall, so the trailing P/E rises all by itself. Notice this one moves no share price at all. If EPS drops from ₹30 to ₹20 while the price sits still at ₹1,000, the P/E jumps from about 33 to 50 and the stock has not moved a rupee. It is not an engine of return; it is a way the reported multiple can mislead you.

That is the point most beginners never reach: a rising P/E does not necessarily mean investors became more optimistic. Sometimes the multiple rises simply because the denominator shrank. When you see a P/E move on any screener, the useful question is not 'is this expensive?' but 'which of the three just fired: real earnings, the market's valuation, or only the denominator?'

The stock price reflects the market's assessment of the company's future, while the P/E only tells you how large that price is relative to the earnings already reported. So a high P/E on a depressed earnings year can be the market saying that it does not believe the depressed year represents the company's true earnings power.

Those are the terms to keep: when the multiple rises it is a re-rating (or multiple expansion); when it falls it is a de-rating (or multiple contraction). And the return you finally pocket is earnings growth and the re-rating, multiplied together.

None of this says a falling earnings number is good news, or that a rising multiple is a signal to buy. It is not. A rising P/E on falling earnings is the market placing a bet that the future is brighter than the reported present. Vinati's bet paid off. Plenty do not.

When the recovery the price was counting on fails to arrive, the same machinery runs in reverse. Earnings stay weak, the market loses faith, and the multiple contracts. Now both terms fall together and the stock drops hard, because it was carrying a hopeful multiple on top of weak earnings. That is de-rating, and it is exactly as brutal as re-rating is kind.

So the useful skill is not spotting that a multiple expanded. It is judging whether the future the expanded multiple is paying for is actually likely to show up. That judgement is the whole game, and this framework only tells you where to point it, not what the answer is.

Everything so far has run in one direction: earnings fell and the multiple rose. Turn the whole thing over and you get the case that confuses beginners just as much. The earnings climb, year after year, and the multiple falls anyway. The business is visibly getting better, but the market is paying progressively less for each rupee of profit. Same identity, opposite term moving.

The clean Indian example is BLS International, the visa-outsourcing company. Look at the chart. The pale bars are trailing earnings per share and they march almost straight up, from under ₹2 to about ₹17 over the window. The purple P/E line does the reverse: it starts in the 40s, drifts down through its own median of 36.3, and ends near 14. Earnings did nothing but rise. The multiple did nothing but fall. The two fought each other the entire way.

The exact figures let us split the move directly, no CAGR arithmetic needed. Reported EPS went from about ₹1.22 in FY21 to ₹16.68 in FY26, so earnings grew about 13.7 times.³ Over the same window the share price grew only about 7 times (a compounding of roughly 48% a year). By the identity, price equals EPS times P/E, so the whole gap between those two must be the multiple. The chart bears that out: the P/E falls from roughly 30 times earnings to roughly 15 (chart-read, approximate), about a halving. And it ties out: 13.7 times a half is about 7. So the earnings did the heavy lifting, the shrinking multiple handed part of it back, and what was left was still an exceptional return, just well short of what the earnings growth alone would have suggested.

Why would the market pay less for each rupee even as earnings pile up? Because the price is a bet on the future, and the market's view of that future can change even while reported earnings keep improving. In BLS's case, concerns about government-contract concentration, the company's move into lower-margin digital and loan-distribution businesses, and more recently the October 2025 MEA debarment (set aside by the Delhi High Court that December) all gave investors reasons to reassess what those future earnings were worth. These are possible explanations for the de-rating, not proof that any one of them caused it. The multiple spans years; most of these worries do not, so treat them as candidates, not a verdict.

This is a de-rating, the same machinery as Vinati but running the other way, and it carries the opposite lesson. A falling P/E does not mean the business is failing, exactly as a rising P/E did not mean it was thriving. Sometimes the multiple falls simply because the market has decided the future is less bright, or less certain, than the earnings trend alone would imply. Whether it is right is the whole question. If BLS holds its margins, a compressed multiple on a compounding business is the kind of setup that can re-rate back up. If the mix keeps thinning the profit, the low multiple was the market seeing it early. The framework only tells you the multiple did the work; it does not tell you who is right.

Set the two cases side by side, because they are the same equation viewed from opposite ends. At Vinati, EPS fell while the P/E rose: the market looked past a temporarily weak year to a recovery it could already see. At BLS, EPS rose while the P/E fell: the business kept improving, but the market became willing to pay less for each rupee of that improving profit. In both, the business and the valuation were moving in opposite directions, and the stock did something a beginner would call impossible.

So the lesson is not only the one you started with, do not read a P/E in isolation. It is also its twin: do not read earnings growth in isolation either. Great earnings growth stacked under a contracting multiple can leave you with a mediocre return. Falling earnings under an expanding multiple, as Vinati showed, can leave you with a surprisingly strong one. The two levers are separate, and either can quietly cancel the other.

That is the whole case study in one line. Your return is earnings growth multiplied by the change in the valuation the market puts on those earnings. Watch both. A wonderful business bought at a multiple that then contracts for years is how you can be completely right about the company and still disappointed by the stock.

Here is the habit to carry away, and it works on any stock screener. A screener only knows today's number. It shows you a P/E, and sometimes that the P/E has risen, but it cannot tell you why it rose. It cannot see the company's own history. So it can never, on its own, tell a genuine re-rating apart from a denominator that simply shrank.

That is what the chart in this piece is for. Pull up any company's P/E against its own long-run median, with its earnings alongside, and the answer becomes visible in one look. Is the multiple sitting above its own history while earnings hold steady or rise? That may indicate a price-led re-rating. Is the P/E high only because the earnings bars dipped? That is the trap, the multiple inflating on a falling denominator while nothing hopeful has actually happened.

Two cautions before you trust what you see. First, mismatched clocks: comparing a one-year price move against multi-year earnings growth can make an ordinary year look like a re-rating, so line up the same period on both sides. Second, cyclicals: in a business whose profits swing with a commodity or a demand cycle, a high trailing P/E often just means earnings are near a low, not that the market has fallen in love. In those, a low P/E on peak earnings can be the more dangerous number.

None of this tells you what to buy. It tells you which question you are actually answering. Not 'is this P/E high?', but 'which engine moved it, and do I believe the future it is paying for?'