The Price-to-Earnings Ratio
The most quoted number in equities, and the most misread. It compares price to one year of profit and says nothing at all about the years after it.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- The P/E ratio is share price divided by earnings per share.
- Trailing uses reported earnings; forward P/E uses estimates, which can be wrong.
- A low P/E is not cheap and a high one is not expensive without knowing growth and risk.
- It breaks entirely when earnings are negative, near zero, or distorted by one-offs.
- A multiple only means something against a reference: history, peers, or the index.
- A multiple is a compressed discounted cash flow, so comparing two multiples is comparing two sets of assumptions.
MAD Academy Training Video · 0:46
Cheap Is Not a Number
A P/E is a ratio of price to one year of profit. It tells you what is being paid, never whether it is worth it.
This lesson is part of a Stock Alerts + Tools plan.
What it says
P/E = share price / earnings per share
- equivalently market capitalization divided by net income
A P/E of 20 means the market is paying twenty dollars for each dollar of annual profit. Inverted, it is an earnings yield of five percent, which is the more useful framing when comparing equities against a bond yield.
That inversion is worth doing routinely, because it puts the multiple on the same scale as the alternative. A market at 25 times earnings offers a four percent earnings yield, and whether that is attractive depends entirely on what a Treasury security pays.
Trailing and forward
| Version | Earnings used | The weakness |
|---|---|---|
| Trailing | The last four reported quarters | Backward-looking; a company can have changed completely |
| Forward | Analyst estimates for the next year | Estimates are frequently wrong, and optimistic more often than not |
| Adjusted | The company's own non-GAAP figure | Definitions vary, so cross-company comparison is unsound |
Forward estimates carry a documented bias. Consensus forecasts start optimistic and are revised down through the year far more often than up, which means a forward P/E computed today is usually being divided by a number that will turn out to have been too high.
Why low is not cheap
A P/E is a single number standing in for a whole stream of future cash flows. It is low when the market expects those earnings to fall, and high when it expects them to grow. Screening for the lowest multiples reliably surfaces companies whose earnings the market believes are about to decline.
The cyclical trap
A cyclical business looks cheapest at the top of its cycle, when earnings are at a peak that will not repeat, and most expensive at the bottom, when depressed earnings inflate the ratio. For cyclicals the multiple often reads exactly backwards.
This is why cyclical companies are frequently analysed on a mid-cycle earnings estimate rather than on reported earnings, and why the sectors where the P/E ratio is least reliable are precisely the ones where it looks most attractive at the wrong moment.
Scroll the chart sideways to see all of it.
- Price
- Trailing EPS x 10
- P/E
Where it stops working
- Negative earnings: the ratio is meaningless and is usually shown as not applicable.
- Near-zero earnings: a tiny denominator produces an enormous, uninformative multiple.
- One-off items: a large gain or charge distorts the denominator for four quarters.
- Heavy financial leverage: interest expense sits inside net income, so capital structure drives the multiple.
- Cross-border comparison: accounting and tax differences make raw multiples less comparable.
The fourth is the one most often forgotten. Two companies with identical operations and different debt loads have different P/E ratios purely because of the interest line, which is a financing fact rather than an operating one.
Using it properly
A multiple is only informative against a reference. The three that carry information are the company's own history, its direct peer group, and the sector or index average.
A stock at 18x is neither cheap nor expensive; a stock at 18x against a ten-year median of 11x and a peer group at 14x is telling a story. Even then the story is about expectations rather than about value: the market is paying more for this company's earnings than it used to, and the question is what it expects that the past did not deliver.
Deep Dive 2 builds a peer set and lines up multiples and growth side by side, which is the comparison this ratio needs.
Peer comparison — for membersWhat a multiple actually represents
A price-to-earnings ratio is often described as the number of years of earnings being paid for, which is a useful first approximation and is not what it is. A multiple is the compressed output of a discounted cash flow: it embeds an expected growth rate, an expected duration of that growth, a required return, and a view on how much of the earnings can actually be distributed.
P/E = payout ratio x (1 + g) / (r - g)
- g is the expected long-run growth rate of earnings
- r is the required return, being the risk-free rate plus a risk premium
- the payout ratio is the share of earnings that reaches shareholders
Reading it this way explains several things that otherwise look arbitrary. Multiples fall when rates rise because r rises. Two companies with identical growth can deserve different multiples because one needs to retain more of its earnings to fund that growth. And a small change in an assumption about g produces a large change in the multiple, because g appears in a denominator that is a difference between two similar numbers.
The consequence is that a multiple is a summary of assumptions rather than a measurement. Comparing two multiples is comparing two sets of assumptions, and the useful question is which assumption differs rather than which number is larger.
Where the earnings figure comes from
The denominator is not one number, and two quoted price-to-earnings ratios for the same company on the same day can differ by half. The difference is always in what earnings figure was used.
| Version | The denominator | Character |
|---|---|---|
| Trailing GAAP | The last four reported quarters, as filed | Factual, and includes every one-off |
| Trailing adjusted | The same period, on the company's adjusted basis | Smoother, and defined by the company |
| Forward consensus | Analysts' estimate for the next twelve months | Almost always a lower multiple, and almost always adjusted |
| Normalised | An estimate of mid-cycle earnings | A judgement, and the only one that handles cyclicals |
| Shiller CAPE | Ten years of inflation-adjusted earnings | Used for indices, and unsuited to single companies |
Forward multiples are systematically lower than trailing ones, for two reasons that have nothing to do with the company being cheap: earnings are usually expected to grow, and estimates are usually optimistic. Comparing a forward multiple against a trailing one is therefore not a comparison, and it is the most common way a stock is made to look inexpensive.
The normalised version is the only one that handles a cyclical business sensibly. At the top of a cycle, a cyclical trades at a low multiple of peak earnings, and at the bottom at a very high multiple of depressed ones, which is exactly the inverse of what the number appears to say.