- What P/E really means, built from first principles
- Why a high P/E can be justified and a low one dangerous
- The two numbers in the ratio — and why the bottom one moves more
A kirana shop is for sale at ₹50 lakh. It earns ₹5 lakh a year.
One question answers everything: how many years until you get your money back? Fifty divided by five. Ten years.
That's it. That's P/E. Price divided by earnings, and the answer is a number of years.
Now the shop next door is for sale at ₹1 crore and earns ₹5 lakh. Twenty years. Is it worse? Not necessarily — maybe it's on the main road and its earnings will double in three years, while the first shop's earnings have been flat for a decade.
That is the entire subtlety of P/E, contained in two shops.
- Share price
- what one share costs
- EPS
- net profit ÷ number of shares
P/E is not a measure of cheap or expensive. It is a measure of what the market expects. A high P/E says buyers expect earnings to grow; a low P/E says they don't. The only question that matters is whether they're right.
Why a high P/E can be perfectly rational
Shop A costs ₹50 lakh earning ₹5 lakh (P/E 10). Shop B costs ₹1 crore earning ₹5 lakh (P/E 20). Now add growth: A's earnings are flat, B's grow 25% a year.
- Shop A — earnings, years 1–5
- 5, 5, 5, 5, 5
- Shop B — earnings, years 1–5
- 5.0, 6.3, 7.8, 9.8, 12.2
- Shop A — total earned in 5 years
- ₹25 lakh
- Shop B — total earned in 5 years
- ₹41 lakh
The trick everybody misses: the bottom number moves too
P/E has two inputs, and almost every conversation about it treats the earnings as fixed and the price as the only thing that moves. In practice the earnings number is often the one doing the moving — and it can move violently.
The same company, one year apart, through a downturn.
- Year 1 — share price
- ₹400
- Year 1 — EPS
- ₹20
- Year 1 — P/E
- 20×
- Year 2 — share price
- ₹360down 10%
- Year 2 — EPS
- ₹8down 60%
- Year 2 — P/E
- 45×
In March 2020, at the depths of the COVID crash, the Nifty's trailing P/E was around 17. By February 2021 it had risen to roughly 42 — higher than the level reached before the 2008 crash.
Read as a cheapness meter, that says February 2021 was the most dangerously overvalued moment in Indian market history. But the index had not gone up threefold in eleven months.
What had happened was that the denominator collapsed. COVID crushed corporate earnings, and a shrunken E makes the ratio balloon even if P does nothing dramatic. As earnings normalised over the following year, the P/E fell sharply without the index falling at all.
Trailing, forward, and which to use
- Trailing P/EEarnings used: Last four reported quartersStrength and weakness: Real, audited, backward-looking. Distorted by one-offs and shocks.
- TTM P/EEarnings used: Trailing twelve months, rolled forward each quarterStrength and weakness: The practical default — current, and still based on actual results.
- Forward P/EEarnings used: Next year's estimated earningsStrength and weakness: Forward-looking, and only as good as the estimate — which is often optimistic.
The comparison that actually works
A P/E on its own means nothing. It only becomes information when compared — and there are exactly three comparisons worth making.
- 1Against the company's own history. Is today's 28× at the top or the bottom of its own ten-year range? This is the single most useful comparison, because it holds the business constant and isolates the change in expectation.
- 2Against its direct competitors. Same industry, same economics, same accounting. If one trades at 18× and a near-identical peer at 34×, one of two things is true: the market sees something you don't, or there's an opportunity. Find out which.
- 3Against the market. A stock at 45× when the index is at 22× is being priced as far better than average. Is it?
One company, its P/E range over ten years.
- 10-year minimum P/E
- 14×
- 10-year median P/E
- 24×
- 10-year maximum P/E
- 41×
- Today
- 37×88th percentile
An IT services company at 26× and a steel producer at 7× are not comparable, and the steel producer is not cheap. Different industries have structurally different multiples because they have structurally different growth, capital intensity, cyclicality and predictability.
A low P/E in a cyclical business often signals the top of the cycle, not a bargain — the next lesson but one is entirely about this, because it is where the most confident beginners lose the most money.
Loss-making companies — a negative P/E is not a number, it's an absence.
Companies with a one-off year — a land sale in the earnings makes the E meaningless.
Banks and NBFCs — usually valued on price-to-book alongside P/E, because their assets are financial.
Deeply cyclical businesses — the E swings so widely that the ratio inverts its meaning at the extremes.
The next lesson covers what to use instead in each case.
- ☐Is this trailing, TTM or forward? Is my comparison on the same basis?
- ☐Does the E include one-off items? Strip them and recompute.
- ☐Where does today's P/E sit in the company's own ten-year range?
- ☐How does it compare with two or three direct peers?
- ☐What growth rate would justify this multiple — and is that rate believable?
- ☐If the P/E moved sharply, which number moved: price or earnings?
See it live · Every fundamental report here plots the stock's P/E against its own multi-year band and shows exactly which percentile it's trading in.
See P/E bands →In one screen
- 1P/E = years of current earnings to repay the price. It measures expectations, not cheapness.
- 2A high P/E paired with real growth can be better value than a low P/E with none.
- 3The earnings number moves too — a rising P/E may mean collapsing profits, not a rising price.
- 4Compare against the company's own history first, then direct peers, then the market.
Check yourself
Nothing is scored or stored — the explanation after each answer is the point.
A stock's P/E rose from 18× to 44× while the share price fell 15%. What happened?
Company A trades at 12× earnings with flat profits for six years. Company B trades at 30× with profits compounding at 25%. Which is better value?