How forward and trailing P/E are calculated
Trailing P/E = share price ÷ earnings per share over the last four reported quarters
Forward P/E = share price ÷ expected earnings per share over the next twelve months
Take Northfield Tools, a fictional industrial company trading at $60. Its last four quarters add up to $3.00 of earnings per share, so its trailing P/E is 60 ÷ 3.00 = 20.0×. Analysts expect $3.60 over the next twelve months, so its forward P/E is 60 ÷ 3.60 = 16.7×. Same price, same company, two different multiples, because the denominator changed from what was earned to what is expected.
Data providers differ in detail. "Trailing twelve months" (TTM) is the sum of the last four quarters; some sites use the last full fiscal year instead, which can be several months staler. "Forward" may mean the next twelve months, the current fiscal year or the next fiscal year, and early in a company's year those are very different numbers. Before comparing two stocks, make sure both multiples were built the same way.
Why they differ: the growth hiding in the gap
The two multiples are linked by one line of arithmetic. If earnings are expected to grow by g over the next year, then:
Forward P/E = trailing P/E ÷ (1 + g)
Northfield's 20.0× trailing becomes 16.7× forward because 20 ÷ 1.20 = 16.7: the forward multiple assumes 20% earnings growth. Turn that around and the gap between the two numbers is a forecast you can read directly. At a trailing P/E of 20×:
| Expected EPS growth, next 12 months | Forward P/E |
|---|---|
| −20% | 25.0× |
| 0% | 20.0× |
| 10% | 18.2× |
| 20% | 16.7× |
| 30% | 15.4× |
So "it only trades at 17 times forward earnings" is not a statement about cheapness on its own. It is a statement that the stock is reasonable if earnings grow 20%, which is a claim about the business, not the price. The implied growth rate article extends the same idea over many years rather than one.
Worked example: when the estimate is wrong
A forward multiple borrows its credibility from the estimate underneath it. Suppose Northfield's year goes badly and it earns $3.00 again instead of the $3.60 analysts expected. At $60, the stock that "traded at 16.7× forward" was really trading at 20× the earnings it would actually deliver. If investors still think 16.7× is the right multiple for a company growing this slowly, the price falls to 16.7 × 3.00 = $50, a drop of about 17%, with no change in the multiple and no surprise in the business beyond the one that was already in the estimate.
| Expected | Delivered | |
|---|---|---|
| EPS, next 12 months | $3.60 | $3.00 |
| Forward P/E at $60 | 16.7× | 20.0× |
| Price at 16.7× delivered EPS | $60 | $50 |
This is the ordinary mechanism behind a stock that falls after an estimate cut: the multiple did not change, the denominator did. It is also why analysts' revisions matter more than their level. A forward P/E built on estimates that have been drifting down for two quarters is less trustworthy than the same number built on estimates that have been rising. VALYSTR's company pages show the consensus next to the growth the price implies, so the two can be compared rather than taken on trust; the earnings beat article covers the same mechanism on results day.
Cyclicals and turnarounds: when the two disagree
For a steady business the forward P/E is a little below the trailing P/E and nothing more needs saying. For a cyclical business the two can tell opposite stories. Imagine a steelmaker at the top of its cycle: it earned $8 a share last year and trades at $48, a trailing P/E of 6×. Next year analysts expect $2 as prices normalise, so the forward P/E is 24×. The trailing figure says bargain; the forward figure says the market already knows the good year is over. Buying a cyclical because its trailing multiple is low at peak earnings is the most common way to pay a full price while believing you paid a low one, a pattern the value trap article sets out in more detail.
Turnarounds run the other way. A company coming out of a bad year has depressed or negative trailing earnings, so its trailing P/E is huge or meaningless, while its forward P/E may look ordinary because the recovery is in the estimate. Here the forward multiple is the more useful one, provided you have a view on whether the recovery is real. In both cases the rule is the same: when the two multiples disagree sharply, the disagreement is the thing to investigate, because it marks where earnings are expected to change the most.
GAAP vs adjusted: check the basis before comparing
Trailing earnings are usually the GAAP figures from the company's filings. Forward estimates are usually "adjusted" earnings, which leave out stock-based compensation, amortisation of acquired intangibles and whatever else the company classifies as non-recurring. For a company that pays a lot of its staff in shares, adjusted earnings can be far above GAAP, so a forward P/E on adjusted earnings sitting next to a trailing P/E on GAAP earnings overstates the growth being forecast and understates the price being paid. Compare like with like: GAAP with GAAP, or adjusted with adjusted, and know which the number in front of you is.
Which one to use
Use both, and use them for different jobs:
- Trailing P/E for what has happened. It is built from audited numbers and cannot be revised down by a change of mood. It is the right anchor for a stable business and for checking a stock's own history.
- Forward P/E for what is being promised. It tells you what the market expects and what it is paying for that expectation. Read it with the revision trend and with the forecast growth made explicit.
- The gap for the question. Divide one by the other and you have the growth the price assumes for the coming year. Compare it with the company's record and with the industry. A forward multiple that needs 30% growth from a business that has managed 8% is not cheap; it is optimistic.
Neither multiple values a business on its own. The good P/E ratio article shows how growth, returns and risk set the multiple a company deserves, and the stock valuation guide puts P/E alongside the cash-flow methods that do the real work. For the multi-year version of the same question, Belief Gap solves for the growth a price implies and sets it next to what forecasts expect.
See the growth a price implies
Belief Gap works out the multi-year growth a share price already assumes and sets it next to what forecasts expect, for any listed company.
Frequently asked questions
Is forward P/E or trailing P/E better?
Neither is better; they answer different questions. Trailing P/E measures the price against earnings that have actually been reported, so it cannot be revised. Forward P/E measures it against a forecast, so it shows what the market expects but inherits every error in the estimate. Look at both, and treat the gap between them as the growth being assumed.
What does it mean when forward P/E is lower than trailing P/E?
It means analysts expect earnings to rise over the next twelve months. Forward P/E equals trailing P/E divided by one plus the expected growth, so a forward multiple of 16.7× against a trailing 20× implies 20% growth. The lower forward number is not evidence of cheapness; it is the forecast restated as a multiple.
Where do forward earnings estimates come from?
From sell-side analysts who cover the company; data providers average their forecasts into a consensus. Companies usually guide those forecasts with their own outlook. The consensus is most reliable for large, steadily covered companies and least reliable for small, cyclical or rapidly changing ones, where a handful of analysts may disagree widely.
Is forward P/E based on GAAP or adjusted earnings?
Usually adjusted earnings, which exclude items such as stock-based compensation and amortisation of acquired intangibles, whereas trailing P/E is often quoted on GAAP earnings. Mixing the two makes a stock look cheaper and faster growing than it is. Check which basis each figure uses before comparing them.
What is a good forward P/E?
There is no single number. A forward P/E is reasonable when the growth it assumes is plausible for that business and the required return is fair for its risk. The same 17× can be cheap for a company that can compound earnings at 15% for a decade and expensive for one whose earnings are peaking. Judge the growth assumption, not the multiple in isolation.