TL;DR. The P/E ratio compares a stock’s price to its earnings per share. The number is easy. The hard part is knowing when it lies — and it lies more often than most investors think. This piece walks through the formula, the difference between trailing and forward P/E, why sectors trade at very different multiples, what Robert Shiller’s CAPE adjusts for, and the five most common ways P/E misleads.
The formula
The U.S. Securities and Exchange Commission’s investor education site puts it cleanly: a price-to-earnings ratio “is calculated by dividing the current stock price by the current earnings per share,” where earnings per share are “the earnings for the past 12 months” divided by common shares outstanding (SEC investor.gov).
So:
P/E ratio = Price per share ÷ Earnings per share (EPS)
A P/E of 20 means investors are paying $20 today for every $1 of last year’s earnings. That’s a relative-pricing yardstick, not a verdict. The same multiple can be cheap, fair, or expensive depending on the business, the cycle, and the interest-rate backdrop.
Two flavors: trailing vs forward
- Trailing P/E uses the past 12 months of reported EPS. It is a fact.
- Forward P/E uses the next 12 months of consensus analyst EPS. It is a forecast.
Trailing tells you where you stand. Forward tells you what the market is paying for. When the two diverge sharply, the gap is information. If forward P/E is far below trailing, the market expects earnings to grow into the price. If forward P/E is far above trailing, the market is bracing for an earnings decline.
A simple worked example
Assume a company earned $5 in EPS over the last twelve months and trades at $100 per share.
- Trailing P/E = 100 ÷ 5 = 20.0
- If analysts expect $6 next year, forward P/E = 100 ÷ 6 = 16.7
That spread between trailing 20.0 and forward 16.7 is the “earnings growth discount.” If next year’s earnings instead come in at $4 (a decline), forward P/E would be 25.0 — higher than trailing — and the stock would look more expensive at the same price.
Same multiple, different meaning — sector context
A 20x P/E means very different things in different industries. Software earnings scale at near-zero marginal cost; bank earnings come from leveraged spreads that mean-revert; energy earnings whip with commodity prices. The result: long-run “normal” multiples differ wildly across sectors. NYU Stern’s Aswath Damodaran maintains a widely cited dataset of trailing and current P/E ratios by industry. The snapshot below uses his January 2026 update.
| Industry | Current P/E | Trailing P/E |
|---|---|---|
| Software (System & Application) | 122.49 | — |
| Semiconductor | — | 100.18 |
| Aerospace & Defense | — | 92.80 |
| Computers / Peripherals | 80.95 | — |
| Oilfield Services & Equipment | 37.98 | — |
| Food Processing | 30.97 | — |
| Brokerage & Investment Banking | 26.81 | — |
| Utility (Water) | 21.99 | — |
| Utility (General) | 21.03 | — |
| Banks (Money Center) | 17.58 | — |
| Retail (Grocery & Food) | 17.46 | — |
| Oil & Gas (Integrated) | 13.01 | — |
The takeaway: “cheap” and “expensive” only have meaning relative to (a) the company’s own history and (b) the sector’s structural multiple. A bank at 12x is roughly normal; a software platform at 12x is either a bargain or a value trap, almost never an ordinary observation.
Fixing the cycle problem — the Shiller CAPE
A single year of earnings can be wildly above or below normal. Bank earnings collapsed in 2009. Energy earnings spiked in 2022. Using a one-year EPS in the denominator makes the index-level P/E flicker with the business cycle — right when you would most like a steady reading.
Robert Shiller’s solution, the Cyclically Adjusted Price-to-Earnings ratio (CAPE), divides price by the trailing ten-year average of inflation-adjusted earnings. The smoothing cancels out single-year noise and lets you compare today’s valuation to a century of history.
CAPE smooths the cycle, but it does not smooth structural change. Critics fairly note that today’s S&P 500 carries far more software and services — higher-margin, lower-capital-intensity businesses — than the index of the 1950s, so an apples-to-apples comparison across eras is harder than the chart suggests. Still: at 41, today’s CAPE sits in the top one percent of all readings since the 1880s, second only to the dot-com peak.
When the P/E lies — five failure modes
1. Earnings near a cyclical peak
The most dangerous P/E is often the lowest one. Cyclical companies (autos, homebuilders, semiconductors, energy producers) post their lowest P/E ratios at peak earnings — precisely when the multiple should be compressing to reflect what comes next. A homebuilder at 6x trailing earnings in a housing boom can be far more expensive than the same name at 25x at the bottom of a downturn. Look at five-to-ten year average earnings before declaring a cyclical stock cheap.
2. Negative or near-zero earnings
A money-losing company has no meaningful P/E (you can technically compute a negative number; nobody uses it). A barely-profitable company has a wild P/E that swings on tiny EPS changes. For early-stage growth names, unprofitable cyclicals, and turnaround stories, use price-to-sales, EV/EBITDA, or free-cash-flow multiples instead.
3. Buybacks and accounting noise
Aggressive share repurchases lower the share count and lift EPS — lowering the P/E even when total earnings have not grown. The business is not cheaper; the math is just different. Similarly, the gap between GAAP and non-GAAP earnings can exceed twenty percent at companies with heavy stock-based compensation, restructuring charges, or one-time items. A P/E built on the non-GAAP figure is a different ratio than one built on GAAP, and the two are rarely the same.
4. Capital-light vs capital-heavy businesses
Sector context matters because business models matter. A REIT’s earnings are depressed by depreciation that does not represent real economic loss — analysts use AFFO multiples instead. Bank earnings can be inflated by reserve releases or depressed by reserve builds. Insurance earnings are smoothed by reserve assumptions. The headline P/E quietly carries each industry’s accounting quirks.
5. Different growth rates
Two companies with the same P/E and different growth rates are not the same price. A 25x P/E on a 25% grower is roughly comparable to a 10x P/E on a 10% grower — that’s the intuition behind the PEG ratio (P/E divided by expected earnings growth). The comparison breaks down once growth slows, since growth itself is what the multiple is paying for.
Interest rates: the silent input
The P/E ratio is the inverse of the “earnings yield” (E/P). A 20x P/E equals a 5% earnings yield. When the 10-year Treasury yields 5%, paying 20x for an equity earnings yield of 5% is no risk premium at all. When the 10-year yields 2%, that same 20x represents a healthy 300-basis-point spread over the risk-free rate.
This is why rate cuts tend to lift multiples and rate hikes tend to compress them — the same dollar of earnings is being discounted against a lower (or higher) risk-free rate. Any analysis of whether the market is “cheap” or “expensive” on P/E that ignores rates is half the analysis.
Common mistakes
- Comparing across sectors without adjustment. Software at 30x is not the same as autos at 30x.
- Treating forward P/E as fact. Sell-side consensus EPS is systematically optimistic at cycle peaks and revised down as the cycle turns.
- Anchoring only on the dot-com peak. Plenty of bear markets started from much lower multiples (1969, 1973). High CAPE is a long-term return headwind, not a timing signal.
- Ignoring buybacks. A falling P/E driven by shrinking share count is not the same as a falling P/E driven by rising earnings.
- Confusing trailing and forward. Always know which you’re looking at, and check the spread.
What to learn next
- EV/EBITDA — strips out capital-structure differences, useful for cross-company comparison.
- PEG ratio — adjusts P/E for expected growth.
- Earnings yield — the inverse of P/E, expressed as a yield for direct comparison to bond yields.
- Shiller CAPE — cycle-adjusted, for index-level valuation.
- Price-to-book and price-to-sales — alternatives when earnings are negative, volatile, or noisy.
Sources
- SEC Investor.gov — Price-Earnings (P/E) Ratio
- NYU Stern, Aswath Damodaran — P/E Ratios by Industry (January 2026)
- multpl.com — Shiller PE Ratio (data updated through June 25, 2026)
- Robert Shiller — Online Data (Yale) for the CAPE methodology and historical series
- Federal Reserve H.15 Selected Interest Rates for the Treasury yield context
Disclosure: This article was produced with AI assistance and reviewed before publication. It is for informational purposes only and is not investment advice.