There is no single good P/E ratio: a multiple is only cheap or expensive relative to something — the company’s own history, how fast its earnings grow, what similar businesses trade on, and what a safe bond pays. A P/E of 12 can be the dearest a company has ever been, and a P/E of 35 can be cheap for a business doubling its profit. Here is how to make each comparison, which earnings sit under the ratio, and the cases where the number misleads.
The P/E ratio is the price of a share divided by its earnings per share. What it really prices is not this year’s profit but every year’s profit to come — so the “right” multiple depends on how long those earnings will last, how fast they will grow and how sure anyone can be of either. A utility with a regulated, slow-growing income and a software company growing fast are both rationally priced on very different multiples.
It helps to turn the ratio upside down. Earnings divided by price is the earnings yield: a P/E of 20 is an earnings yield of 5%, a P/E of 10 is 10%, a P/E of 40 is 2.5%. That is the share of the price the company earns in a year, and it makes the comparisons below easier to see. Four reference points carry almost every honest judgement of a P/E:
| Against | The question it answers | Watch out for |
|---|---|---|
| Its own history | Is the market paying more or less for this business than it usually has? | A business that has changed shape is being measured against one that no longer exists. |
| Its growth | How quickly will earnings catch up with the price? | The growth rate is a forecast, and the answer moves a lot with it. |
| Its sector | Is it dear or cheap next to businesses that earn money the same way? | A whole sector can be expensive at once. |
| Interest rates | Is the earnings yield worth the risk against a safe bond? | A bond coupon is fixed; earnings are neither fixed nor guaranteed. |
The cleanest comparison is the company with itself. The same business, measured the same way, has traded on a range of multiples over the years, and today’s P/E sits somewhere in that range. Where exactly is a valuation percentile: the share of days in a window on which the multiple was lower than it is now.
An illustration: a company has traded between 14× and 26× earnings over ten years, with a median of 19×, and is on 15× today. It is near the bottom of its own range — cheaper than on most days of that decade. That is a statement of fact about one business, and it holds whether the business is a bank or a software company, which is exactly what a bare “15 is low” cannot say.
Two things to keep attached to it. The window: the same company can be at the bottom of its ten-year range and the middle of its one-year range, and both are true. And the shape of the business: if the company has sold its best division or taken on far more debt, its old multiples describe a company that no longer exists, and a low reading against them may simply be the new normal.
A high P/E is a bet that earnings will grow into the price. Take a company on 30× whose earnings per share grow 20% a year. After five years EPS is about 2.5 times larger (1.25 ≈ 2.49), so if the share price had not moved at all, the P/E would have fallen to about 12×. The “expensive” stock was cheap for anyone who could see the growth coming. If the growth stalls at 5% instead, the same price still stands on about 23.5× earnings five years on (30 ÷ 1.055), and the multiple usually comes down to meet the growth — through the price.
The PEG ratio folds this into one number: P/E divided by the expected annual growth rate in percent. The company above has a PEG of 1.5 (30 ÷ 20). A company on 12× growing 3% a year has a PEG of 4 (12 ÷ 3) — a lower P/E, but a dearer price per unit of growth. PEG is a quick sense check, not a valuation: the growth rate in the denominator is a forecast, small changes in it swing the answer, and it ignores how long the growth can last and what it costs to achieve.
Comparing with peers works when the peers earn money the same way. Businesses with heavy capital needs, thin margins or earnings that swing with the economy have usually traded on lower multiples than businesses with recurring revenue and little capital to tie up, and that difference is structural rather than a bargain waiting to be noticed. So compare a bank with banks and a software company with software companies — and remember that a whole sector can be dear at once, in which case “cheaper than its peers” only means “less expensive than something expensive”.
Interest rates set the backdrop for every multiple. A share on 25× earnings has an earnings yield of 4%. If a government bond pays, say, 2%, a buyer gives up little safe income to own the share; if the same bond pays 5%, the buyer is accepting less current earnings than a risk-free coupon, and is relying entirely on growth to make up the difference. A higher safe rate raises the return investors demand from shares, which pulls down the multiple they will pay; a lower one does the opposite. The comparison is rough — a coupon is fixed while earnings can grow or vanish — but it explains why the same P/E can be reasonable in one decade and stretched in the next.
“The P/E” is several numbers, depending on which earnings sit underneath it. One illustrative company with a $60 share price:
| Earnings basis | EPS | P/E at $60 |
|---|---|---|
| Trailing, adjusted | $3.00 | 20× |
| Trailing, GAAP (after a $1 write-down) | $2.00 | 30× |
| Forward, analysts’ estimate | $4.00 | 15× |
Trailing P/E uses the last twelve months of reported earnings: a fact, but a backward-looking one. Forward P/E uses analysts’ estimate for the coming year: closer to what you are buying, but a forecast that can be wrong in either direction. Adjusted earnings set aside one-off items; GAAP earnings are exactly as filed, write-downs included. None of them is the right one in general. What matters is comparing like with like — a forward P/E against a forward P/E, an adjusted history against an adjusted history — and noticing when the gap between them is large, because that gap is itself information.
On every company page, the Key metrics band on the Summary tab prints P/E · TTM and P/E · FWD side by side, so the gap between trailing and forward earnings is visible at a glance.
The comparison with the company’s own history is what the Fair Value tab is built on. Its Earnings lens draws the P/E over the company’s own past and places today’s reading as a percentile of a window you choose — 1Y, 3Y, 5Y, 10Y or All — then names it on a five-step scale from Deep Value (below the 10th percentile) through Good Entry, Fair Value and Rich to Expensive (above the 90th). The verdict is never drawn without the window it was judged over. An Adj | GAAP switch chooses which earnings the lens uses; there is no forward basis there, because forecasts have no history to take a percentile of. A company with no positive earnings over the last twelve months has the lens set aside with that reason beside it rather than a meaningless figure, and six other lenses — cash flow, revenue, book value and more — are read the same way. You can also type the multiple you think is right and see the fair value it implies today.
The Fair Value guide walks through the tab step by step.
The Fair Value tab is part of Ultimate; company pages and their key metrics are free. Every new account gets Ultimate for seven days once the email address is confirmed, with no card.
Educational only. Not investment advice.
Lemma Analysis gives independent investors both lenses in one terminal — a stock screener and an intrinsic value tool for the fundamentals, alongside price history and charts for timing.