The price-to-earnings ratio divides a company’s share price by its earnings per share. The result is how many dollars the market is currently paying for each dollar of annual profit. A stock at $60 earning $3 a share trades at 20 times earnings.
It is the most quoted valuation figure in investing and the most casually misread. The persistent error is treating it as a verdict, where high means expensive and low means cheap, when it is a statement about what the market expects to happen next, wrapped around a denominator that is least reliable at exactly the moments the ratio matters most.
What You Are Really Buying at 20 Times
P/E = share price ÷ earnings per share
Read literally it is a payback period: at 20×, you are paying twenty years of current profit for a claim on the business. Nobody expects earnings to stay flat for twenty years, which is the point. The multiple is the market’s summary of everything it believes about future growth, durability and risk, compressed into one figure.
A company at 40× is not being called expensive. It is being credited with growth that will make today’s earnings look small. A company at 8× is not being called cheap. It is being warned about.
Where the Market Sits Now
Context matters more than the absolute number. As of September 2, 2026, the S&P 500’s trailing P/E stood at 29.56, against a long-run mean of 16.23 and a median of 15.08 for a series running back to the nineteenth century (multpl, 2026).
The market trades at roughly 1.8 times its historical average multiple. That is a fact, not a forecast. Elevated multiples have persisted for years at a stretch, and the index composition has shifted toward higher-margin businesses that plausibly deserve higher multiples than the railroads and manufacturers that dominated it a century ago.
Trailing and Forward Are Different Claims
Trailing P/E uses the last four quarters of reported earnings. It is factual and backward-looking. Forward P/E uses analysts’ consensus for the next four quarters. It is relevant and hypothetical.
Because estimates generally assume growth, forward P/E is normally lower than trailing, which makes any stock look cheaper on a forward basis. That is not manipulation, just what the arithmetic does when the denominator is expected to rise. But it means a quoted P/E is only meaningful once you know which one it is.
It also helps to know how demanding those estimates are as a bar. During the 2026 second-quarter reporting season, 86% of S&P 500 companies reported earnings above estimates. Over the previous five years that figure averaged 78%, and over ten years 76% (FactSet, 2026). Consensus is a bar companies help set and then usually clear.
Treat it as a reference point rather than an independent forecast.
When comparing two companies’ P/E ratios, confirm both come from the same basis: both trailing or both forward, both GAAP or both adjusted. Mixing them is the most common way a comparison silently becomes meaningless, and summary pages rarely label which they show.
Why the Denominator Is the Weak Point
Price is unambiguous. Earnings are an accounting output, and everything that distorts earnings per share flows straight into the ratio.
One charge, two valuations
Apple’s fiscal 2024 shows this cleanly. Reported GAAP diluted EPS was $6.08, depressed by a one-time $10.2 billion tax charge from the EU’s highest court reinstating the Commission’s State Aid decision. Excluding it, Apple’s own adjusted figure was $6.75.
At any given share price those two denominators produce P/E ratios about 10% apart: same company, same day, same filing.
When earnings vanish, the ratio inverts
Scale that up and the ratio breaks completely. The index P/E of 123.73 in May 2009 did not mean stocks were the most expensive in history. It meant aggregate earnings had collapsed in the financial crisis while prices had begun recovering, leaving almost nothing in the denominator. At that reading, equities were close to their most attractive entry point in a generation.
A ratio that reads “dangerously expensive” at the bottom of a crash is not measuring what its users think.
Why a Low P/E Is Often a Warning
The market usually reprices a stock before the accounts catch up. Price falls first; reported earnings, which describe the past, fall later. In between, the ratio drops — which looks like a bargain appearing and is often deterioration becoming visible.
This is sharpest in cyclical industries. A steelmaker or homebuilder at the peak of its cycle posts record earnings, and its P/E looks lowest precisely when the next move is down. The same company at the trough, earning almost nothing, shows a high or undefined ratio just as recovery begins. For cyclicals, the ratio inverts the signal a naive reading takes from it.
Three questions turn a low P/E from a screen result into a decision. Is the sector cyclical, and where is it in that cycle? Are the trailing earnings inflated by anything non-recurring? And is the forward estimate meaningfully below the trailing figure, meaning analysts already expect the denominator to shrink?
Comparability: Sector, and Not Much Beyond It
Multiples differ structurally by industry for real reasons. High-margin, low-capital software businesses support higher multiples than capital-intensive utilities. Banks, whose earnings depend on leverage and provisioning judgements, are conventionally valued on book value alongside earnings.
Identical businesses, different share counts
Two identical businesses can also carry very different EPS purely through capital structure. Consider illustrative arithmetic, not real securities:
The EPS figures differ fivefold; the P/E ratios are identical. This is exactly what the ratio is for — it normalizes away the share count, so it permits comparisons that raw EPS does not.
CAPE: Smoothing Out the Cycle
The cyclically adjusted P/E, or Shiller P/E, attacks the denominator problem directly by dividing price not by one year of earnings but by the average of the past ten, adjusted for inflation. Averaging a full business cycle prevents a single distorted year from dominating.
On September 2, 2026, CAPE stood at 41.93, against a long-run mean of 17.40, a median of 16.11, and an all-time high of 44.19 recorded in December 1999 (multpl, 2026). The current reading sits within a few points of the dot-com peak.
CAPE has its own critics. Ten-year averaging carries forward earnings from an economy that may no longer resemble the present one, accounting standards have changed across the window, and it has signalled overvaluation through extended periods in which markets kept rising. It is long-horizon context, not a timing tool.
PEG: Putting Growth Into the Denominator
The most useful single adjustment to P/E is to divide it by the expected earnings growth rate. A company on 30× earnings growing profits 30% a year has a PEG of 1.0; one on 30× growing at 5% has a PEG of 6.0. The multiples are identical and the propositions are not remotely alike.
A PEG near or below 1.0 is conventionally treated as reasonable, though the convention is rougher than it sounds. The growth rate is a forecast, so PEG inherits all the uncertainty of the estimate and then compounds it by putting that estimate in the denominator. A modest error in the growth assumption produces a large error in the ratio. PEG works best as a sanity check on whether a high multiple is supported by any plausible growth path, rather than as a precise valuation.
Four Habits That Keep the Ratio Honest
In practice, the ratio earns its keep as a prompt rather than a screen output. Four habits make the difference:
- Chart the company’s own P/E over five to ten years. A business trading at the low end of its own historical range is a more meaningful observation than one trading below an arbitrary market average.
- Compare against three or four named peers in the same industry, on the same basis, rather than against the index.
- Look at trailing and forward together. When forward sits well below trailing, the market expects earnings to rise; when it sits above, it expects them to fall. The direction of that gap often carries more information than either number alone.
- Read what is in the denominator. Open the income statement, check for one-time items, and confirm whether you are looking at GAAP or adjusted figures.
None of this converts P/E into a decision on its own. It is one input among several, and the specific mistake worth avoiding is letting a single multiple substitute for reading the business.
The Blind Spots
- It ignores debt entirely. Two companies on the same multiple can carry completely different balance-sheet risk, which is why enterprise value multiples exist.
- It is undefined for loss-making companies. Any business with negative earnings has no ratio at all, which excludes much of the early-growth market.
- It says nothing about the growth rate. The PEG ratio divides P/E by expected earnings growth for this reason: 30× on 30% growth is a different proposition from 30× on 5%.
- It inherits every accounting choice. GAAP versus adjusted, one-time items, revenue recognition timing — every one of them lands in the denominator.
- It is not cash. Earnings can diverge from free cash flow for years, and cash is what funds dividends and buybacks.
The Bottom Line
P/E is a good first question and a poor final answer. It compresses growth expectations, risk and accounting policy into one figure, then presents it with a precision it does not have.
Use it to compare a company against its own history and against direct sector peers, always on a stated basis. Treat an unusually low multiple as something to explain rather than something to buy, particularly in cyclical industries. Check what the trailing earnings actually contain. And keep the market context in view: at 29.56 against a 16.23 mean, the index-level multiple is doing a great deal of work that individual stock analysis cannot see.
For the denominator itself, see earnings per share explained; for what earnings leave out, see net income.
