P/E Ratio Explained: What 150 Years of Data Say a Good P/E Is

The price-to-earnings ratio is the most quoted number in investing and probably the most misunderstood. Every stock screener shows it, every earnings headline mentions it, and almost every argument about whether the market is "expensive" eventually comes down to it. Yet the same P/E of 30 can describe a bargain, a fair price or a bubble depending on what is happening to the earnings underneath it.

This guide explains what the P/E ratio is, how to calculate it, the different versions you will see quoted, and what a "good" P/E actually looks like. Rather than repeating textbook rules of thumb, we tested them against more than 150 years of S&P 500 price and earnings data compiled by Yale economist Robert Shiller, and we calculated today's P/E ratios for six well-known US companies using the earnings they reported to the SEC. The short version: the long-run average market P/E is about 16, the median is about 15, starting valuations have mattered for ten-year returns, and a single P/E number on its own tells you far less than most people assume.

What is the P/E ratio?

The P/E ratio compares a company's share price with the profit it earns for each share. The formula is simple:

P/E ratio = share price ÷ earnings per share (EPS)

If a stock trades at $50 and the company earned $2.50 per share over the past year, its P/E is 20. Another way to say the same thing: investors are paying $20 for every $1 of annual profit the business currently generates. If earnings stayed exactly flat forever and every cent were handed back to shareholders, it would take 20 years of profits to "pay back" the share price.

That framing is useful because it makes clear what the number is measuring. A P/E ratio is not a measure of quality, safety or growth on its own. It is a price tag expressed in years of current profit. A high number means the market is paying a lot for each dollar of today's earnings, usually because it expects tomorrow's earnings to be much larger. A low number means investors are paying little for each dollar of profit, usually because they expect those profits to shrink, stagnate or be at risk.

You can also flip the ratio upside down. Earnings divided by price gives the earnings yield. A P/E of 20 is an earnings yield of 5% (1 ÷ 20). A P/E of 10 is an earnings yield of 10%. Expressing valuation as a yield makes it easier to compare stocks with bonds and cash, which we will come back to later.

Where the "E" comes from

Earnings per share is net income attributable to common shareholders divided by the number of shares outstanding. Public companies report it every quarter. In the United States, the figure you see in a 10-K or 10-Q is calculated under generally accepted accounting principles (GAAP), and companies report two versions:

• Basic EPS divides net income by the weighted average number of shares actually outstanding.
• Diluted EPS also counts shares that could be created from stock options, restricted stock units, convertible bonds and warrants. Diluted EPS is the more conservative figure and is the one most analysts use for P/E calculations.

Because diluted EPS includes potential new shares, it is always equal to or lower than basic EPS. For companies that pay a lot of stock-based compensation, the gap can be meaningful. All company calculations in this article use diluted EPS taken directly from filings submitted to the SEC.

A worked example

Suppose a company earned net income of $8 billion last year and had 2 billion diluted shares outstanding. Its diluted EPS is $4.00. If the stock trades at $96, its P/E is 96 ÷ 4 = 24. Its earnings yield is 4 ÷ 96 = 4.17%.

Now imagine two things can happen next year. In scenario A, profits rise 25% to $10 billion and the share count is unchanged, so EPS becomes $5.00. If the stock price stays at $96, the P/E falls to 19.2 without the stock getting any cheaper in dollar terms. In scenario B, profits fall 25% to $6 billion, EPS drops to $3.00 and, at the same $96 price, the P/E rises to 32. Same price, very different valuations. That simple illustration explains most of the confusion around P/E ratios: the ratio moves when either the price or the earnings move, and earnings can move a lot.

The different types of P/E ratio

When someone says a stock "trades at 25 times earnings," the first question to ask is: which earnings? There are several common versions, and they can give wildly different answers for the same company on the same day.

Trailing P/E (TTM)

The trailing P/E uses earnings from the past twelve months, usually the sum of the last four reported quarters. This is what most finance websites display by default. Its strength is that the earnings are real: they have been reported and audited or reviewed. Its weakness is that it looks backwards. A company coming out of a bad year will look expensive on trailing earnings even if profits are about to recover, and a company at the peak of a boom will look cheap just before profits fall.

Forward P/E

The forward P/E divides the price by analysts' estimates of earnings over the next twelve months or the next fiscal year. It is the version most professional investors talk about, because markets price the future rather than the past. The obvious problem is that forecasts are forecasts. Consensus estimates tend to be revised down as a year progresses, and in recessions they can be badly wrong. A forward P/E is only as good as the estimate in the denominator.

Fiscal-year P/E

Some sources, including this article's company table, use the most recent full fiscal year's EPS from the annual report. It is less timely than a trailing-twelve-month figure but is easy to verify and is drawn straight from the company's 10-K. When fiscal years end at different times of year, as they do for Apple, Microsoft and Nvidia, comparisons between companies need a little care.

Adjusted or "operating" P/E

Many companies publish adjusted EPS that strips out items management considers unusual: restructuring charges, acquisition costs, impairments, gains or losses on investments, and frequently stock-based compensation. Adjusted earnings are almost always higher than GAAP earnings, which makes adjusted P/E ratios look lower. Wall Street estimates are usually based on adjusted figures. There is nothing wrong with looking at them, but you should know which version you are comparing, and you should be suspicious of a company whose "one-off" charges recur every year.

Shiller P/E or CAPE

The cyclically adjusted price-to-earnings ratio, popularised by Robert Shiller and John Campbell, divides the current price by the average of the past ten years of earnings, with both adjusted for inflation. Averaging a decade of profits smooths out the boom-and-bust swings of the business cycle, so CAPE is mainly used for whole markets rather than individual stocks. We look at it in detail below because it reveals something the ordinary P/E hides.

Version	Earnings used	Best for	Main weakness
Trailing (TTM)	Last four reported quarters	Factual, comparable screens	Backward-looking; distorted by unusual years
Forward	Analyst estimates, next 12 months	Growth companies, market pricing	Estimates are often wrong
Fiscal-year	Last annual report	Easy verification from filings	Can be up to a year stale
Adjusted	Company-defined "core" earnings	Seeing through genuine one-offs	Flatters results; definitions vary
CAPE (Shiller)	10-year average, inflation-adjusted	Whole-market valuation	Slow to react; affected by accounting changes

What 150 years of data say about the market's P/E

The most common question about P/E ratios is "what is a normal level?" The best public dataset for answering it is Robert Shiller's monthly series of S&P 500 (and pre-1957, its predecessor Composite index) prices, dividends and earnings going back to January 1871. We downloaded the file directly from Shiller's Yale data page; the latest edition we could retrieve contains earnings through June 2023, so the long-run statistics below run from January 1871 to June 2023. That is more than 1,800 monthly observations.

The long-run average P/E is about 16

Calculating price divided by trailing twelve-month reported earnings for every month in the data gives these results:

Measure (S&P Composite, monthly)	1871–2023	1950–2023
Mean trailing P/E	16.0	18.5
Median trailing P/E	14.9	17.5
Mean CAPE	17.4 (all available months)
Median CAPE	16.5 (all available months)

Source: MarketsHQ calculations from Robert Shiller's online data (ie_data.xls), monthly observations January 1871–June 2023.

Two points stand out. First, the often-quoted "historical average P/E of about 15 to 16" holds up: the full-period mean is 16.0 and the median is 14.9. Second, the market has spent most of the post-war era above that level. Since 1950 the mean has been 18.5 and the median 17.5. Whether that reflects a permanent shift or a long stretch of unusually favourable conditions is one of the biggest debates in finance.

There are reasonable arguments that higher modern P/E ratios are justified. Information is cheaper, trading costs have collapsed, index funds make diversification nearly free, and the mix of companies in the index has shifted towards software and services businesses with higher profit margins and lower capital needs. Interest rates also fell for roughly four decades after 1981, and lower rates make future profits worth more today. There are equally reasonable arguments that some of the post-1990 premium simply reflects investors paying more than they should. The data cannot settle that debate, but it does tell you that "the market is above its 150-year average" has been true for a large share of the past 35 years.

The extremes: from 5 to 124

The range of the trailing P/E is enormous. The lowest reading in the data is about 5.3, in December 1917, when wartime profits were high and stock prices depressed. The highest is about 123.7, in May 2009. That 2009 spike is the single best illustration of why you should never read a P/E without looking at the earnings behind it.

Stocks were not expensive in May 2009. The S&P 500 had just fallen more than 50% from its 2007 peak. What happened is that trailing reported earnings collapsed: the twelve-month reported EPS figure in Shiller's data fell to about $7.29, as banks and insurers booked enormous write-downs. Divide a price of about 902 by earnings of 7.29 and you get a P/E above 120. Anyone who concluded that the market was in a bubble at that moment would have missed one of the best buying opportunities of the century.

Month	S&P Composite price	Trailing EPS	Trailing P/E	CAPE	What was happening
Sep 1929	31.30	1.55	20.2	32.6	Just before the crash
Jun 1932	4.77	0.51	9.4	5.6	Great Depression low
Jun 1949	13.97	2.40	5.8	9.1	Post-war pessimism
Jul 1982	109.40	13.97	7.8	6.6	Start of the 1980s bull market
Dec 1999	1,428.68	48.17	29.7	44.2	Dot-com peak
May 2009	902.41	7.29	123.7	16.0	Earnings collapse after the financial crisis
Dec 2020	3,695.31	94.13	39.3	33.8	Pandemic earnings dip
Dec 2021	4,674.77	197.87	23.6	38.3	Earnings rebound, prices near a peak

Source: Robert Shiller online data; monthly average prices, trailing twelve-month reported earnings. Figures rounded.

Look at the table's last two rows. In December 2020 the trailing P/E was a frightening-looking 39.3, because pandemic-hit earnings were still in the denominator. One year later, with the index more than 25% higher, the P/E had fallen to 23.6 because earnings had more than doubled. Prices then fell sharply in 2022. The ordinary P/E said the market got cheaper through 2021; the CAPE, which rose from 33.8 to 38.3, said it got more expensive. In this episode the CAPE was the better warning sign.

The 1929 row is equally instructive. On trailing earnings, the market at the peak before the Great Crash was on a P/E of about 20, which does not look extreme. The CAPE of 32.6 told a very different story, because earnings in 1929 were far above their ten-year average. Peak-cycle profits make the ordinary P/E look reasonable at exactly the wrong moment.

Does the P/E ratio predict returns?

This is the question that matters to investors. Valuation is only useful if paying a lower price for each dollar of earnings has historically led to better outcomes.

To test it, we used Shiller's real (inflation-adjusted) price and dividend series to build a total-return index for the S&P Composite, then calculated the annualised real total return over the following ten years for every starting month from January 1881 to June 2013. That gives 1,590 overlapping ten-year windows. We then grouped each starting month by its CAPE.

Starting CAPE	Number of start months	Median 10-year real return (per year)	Worst	Best
10 or below	229	11.0%	1.8%	20.0%
10 to 15	461	7.8%	-4.2%	17.6%
15 to 20	514	6.2%	-4.6%	16.1%
20 to 25	246	4.6%	-4.0%	14.2%
25 to 30	83	5.2%	-1.2%	6.6%
Above 30	57	-1.1%	-5.9%	4.5%

Source: MarketsHQ calculations from Robert Shiller online data. Real total return with dividends reinvested monthly, before fees and taxes. Overlapping windows; start months January 1881–June 2013.

The pattern is clear and broadly consistent. Starting from a CAPE of 10 or below, the median ten-year real return was 11.0% a year, and not a single ten-year window lost money after inflation. Starting from a CAPE above 30, the median was -1.1% a year and the best outcome was only 4.5%. Across all windows, the correlation between starting CAPE and the next decade's real return was -0.52: a meaningful, though far from perfect, inverse relationship.

Before drawing conclusions, notice what the table does not say.

• It says nothing about the next year. Valuation has a weak record of predicting short-term returns. Markets can stay expensive for years, and the dot-com bubble's CAPE kept rising long after it first passed 30.
• The ranges overlap heavily. A CAPE between 15 and 20 produced ten-year real returns anywhere from -4.6% to +16.1% a year. Plenty of other things, from inflation to interest rates to wars, also drive returns.
• The high-CAPE sample is small. Only 57 starting months had a CAPE above 30, and they cluster around 1929, the late 1990s and the late 2000s. These are overlapping windows, so the effective number of independent episodes is much smaller than 57.
• The window ends in 2013. We can only measure complete ten-year periods, so the most recent starting points, including those since CAPE went back above 30 in the late 2010s, are not in the table. Some of those years went on to produce strong returns, which is one reason critics argue CAPE has become less reliable.

The honest summary: buying the market when it is cheap relative to its earnings history has generally produced better long-term outcomes, and buying when it is very expensive has generally produced worse ones. It is a tool for setting expectations, not a timing signal. That distinction matters, and it is consistent with what our own research on time in the market versus timing the market found: missing a handful of strong days can cost more than a modest valuation edge gains.

Real-company P/E ratios: six examples from SEC filings

Market-wide numbers are one thing; most people look up P/E ratios for individual stocks. To show how much the number varies, we took each company's most recent full-year diluted EPS as reported in its annual 10-K filing to the SEC and divided its closing share price on October 20, 2026 into it.

Company	Close, Oct 20, 2026	Fiscal year ended	Diluted EPS (10-K)	Fiscal-year P/E	Earnings yield
JPMorgan Chase (JPM)	$335.26	Dec 31, 2025	$20.02	16.7	6.0%
ExxonMobil (XOM)	$177.86	Dec 31, 2025	$6.70	26.5	3.8%
Microsoft (MSFT)	$531.87	Jun 30, 2026	$17.95	29.6	3.4%
Nvidia (NVDA)	$216.47	Jan 25, 2026	$4.90	44.2	2.3%
Apple (AAPL)	$339.50	Sep 27, 2025	$7.46	45.5	2.2%
Tesla (TSLA)	$373.71	Dec 31, 2025	$1.08	346.0	0.3%

Sources: diluted EPS from SEC EDGAR XBRL company filings (us-gaap EarningsPerShareDiluted, 10-K annual values); closing prices from Yahoo Finance. P/E uses the latest full fiscal year, not trailing twelve months, so figures will differ from those shown on quote pages. Not a recommendation to buy or sell any security.

These six numbers range from under 17 to over 340. None of them, on its own, tells you whether the stock is a good investment. What each one does tell you is what the market is implicitly assuming.

JPMorgan: a low P/E is normal for banks

Banks have traded on lower P/E ratios than the overall market for most of modern history. Their earnings are tied to the credit cycle, they are highly leveraged, and they need to retain capital to satisfy regulators. A bank P/E in the teens is the sector norm rather than a bargain signal. Investors in banks often look at price-to-book value and return on equity alongside the P/E for that reason.

ExxonMobil: when falling earnings push the P/E up

Exxon's diluted EPS declined in each of the last three fiscal years in our data: $8.89 in 2023, $7.84 in 2024 and $6.70 in 2025, as reported to the SEC. Lower oil and gas prices than the 2022 peak reduced profits. With the share price higher, the P/E has risen to the mid-20s, which is high for an energy company. For cyclical businesses like oil producers, steelmakers and chemical companies, the P/E is often lowest at the top of the cycle (when profits are huge) and highest near the bottom (when profits are depressed). Buying cyclicals on a low P/E can be a trap; the earnings may be about to fall.

Microsoft and Nvidia: growth carries the multiple

Nvidia's reported diluted EPS rose from $1.19 for the fiscal year ended January 2024 to $2.94 a year later and $4.90 for the year ended January 2026, roughly a fourfold increase in two years. A P/E in the 40s for a company that has been compounding earnings that quickly is a different proposition from a P/E in the 40s for a company growing at 5%. Investors are paying for expected future growth. If that growth continues, today's P/E will look cheap in hindsight; if it slows sharply, the P/E can compress quickly, taking the share price with it. We examined that risk in our analysis of whether the AI trade is a bubble.

Apple: a premium for consistency and buybacks

Apple's P/E in the mid-40s is well above its own long-run history. Part of what investors pay for is the durability of its earnings and the cash it returns. Apple has repurchased more than $800 billion of its own stock since fiscal 2013, which steadily raises EPS by reducing the share count. We covered that mechanism in our guide to stock buybacks. Buybacks matter for P/E analysis: if a company is growing EPS mainly by shrinking its share count rather than by growing net income, the quality of that growth is lower than it looks.

Tesla: when the P/E stops being meaningful

Tesla's diluted EPS fell from $4.30 in 2023 to $2.04 in 2024 and $1.08 in 2025, according to its 10-K filings. At a share price above $370, that produces a P/E of roughly 346. A number that large is essentially telling you that the share price is not being set by current earnings at all. Investors are valuing the company on expected future businesses and profits far larger than today's. Whether that is justified is a judgment call, but the P/E itself has stopped doing useful work. In cases like this, analysts usually switch to other measures such as forward earnings several years out, price-to-sales or discounted cash flow models.

So what is a "good" P/E ratio?

There is no single good P/E ratio. A P/E is only meaningful in comparison with something else. Here are the comparisons that actually help.

1. Compare with the company's own history

A stock's P/E relative to its own five- or ten-year range tells you whether investors are currently more or less optimistic than usual. A durable consumer-staples business trading at the bottom of its historical range may be worth investigating; one trading at the top may already price in the good news. This comparison works best for mature, stable companies whose business has not changed fundamentally.

2. Compare with peers in the same industry

Different industries have structurally different P/E ratios. Utilities, banks, energy producers and carmakers usually trade on lower multiples. Software, medical devices and premium consumer brands usually trade higher. Comparing a bank's P/E with a software company's is like comparing a house's price per square foot with a car's: the units match, but the comparison is meaningless.

3. Compare with growth: the PEG ratio

The price/earnings-to-growth (PEG) ratio divides the P/E by the expected annual EPS growth rate in percent. A company on a P/E of 30 expected to grow earnings 30% a year has a PEG of 1.0. One on a P/E of 15 growing at 5% has a PEG of 3.0. Peter Lynch popularised the idea that a PEG around 1 is fair value for a growing company. It is a useful sanity check, but it inherits all the problems of growth forecasts, and it ignores how long the growth can last. A company growing 30% for two years is worth far less than one growing 30% for ten.

4. Compare with interest rates: the earnings yield

When bond yields are high, investors can earn a decent return without taking stock-market risk, so they tend to demand a higher earnings yield (a lower P/E) from stocks. When bond yields are very low, they will accept a lower earnings yield. Comparing the market's earnings yield with the yield on 10-year Treasury bonds is sometimes called the "Fed model." It has weak theoretical foundations, because earnings are a real quantity that grows with inflation while bond coupons are fixed, but it does capture why P/E ratios rose so much as rates fell after 1981 and why higher rates since 2022 have put pressure on high-multiple stocks.

5. Compare with the market's own history

For the whole market, the Shiller data gives a sense of scale. A trailing P/E below about 12 or a CAPE below about 12 has historically been cheap. A P/E around 15 to 18 has been roughly typical. A CAPE above 30 has occurred only around 1929, the dot-com era and the past several years. None of those thresholds is a rule; they are a description of where the market has spent its time.

A simple framework for reading any P/E ratio

When you look up a stock's P/E, run through these questions before you form a view.

1. Which earnings? Trailing, forward or adjusted? GAAP or non-GAAP? Make sure you are comparing like with like.
2. Are the earnings normal? Was last year unusually good or bad? A one-off gain, write-down or tax benefit can distort EPS dramatically.
3. Is the business cyclical? If profits rise and fall with commodity prices or the economy, a low P/E at the peak may be a warning rather than an opportunity.
4. How fast are earnings growing, and for how long? High growth can justify a high P/E, but only if it lasts.
5. Where is the growth coming from? Revenue growth, margin expansion and share buybacks all raise EPS, but they are not equally durable.
6. What do peers trade at? Is the company priced at a premium or discount to similar businesses, and is there a reason?
7. What does the balance sheet look like? Two companies with the same P/E but very different debt levels carry very different risks. That is why many professionals also use enterprise value to EBITDA.
8. What is the market implying? Work backwards: what earnings growth would make this price reasonable? If the answer requires everything to go right, the stock is priced for perfection.

Working backwards: what does a P/E imply?

Step 8 is the most useful habit you can develop. Here is a simplified way to do it. Suppose you believe a mature, average-quality company should eventually trade at a P/E of about 18, close to the post-1950 market median. If a stock trades at a P/E of 45 today and you expect it to reach that "mature" multiple in ten years, its earnings need to grow enough that the P/E falls by a factor of 2.5 (45 ÷ 18) just to keep the share price flat. That requires EPS to grow about 9.6% a year for ten years (because 1.096 to the power of 10 is about 2.5) before shareholders make any money from multiple compression alone.

If you want the stock to return, say, 8% a year over that decade on top, earnings need to grow by roughly 2.5 × 2.16 = 5.4 times in ten years, or about 18% a year. That is achievable for an exceptional company, but very few companies sustain that pace for a decade. Running this kind of arithmetic does not tell you whether to buy or sell. It tells you what you are betting on.

Starting P/E	Annual EPS growth needed for flat price if P/E reverts to 18 in 10 years	Annual EPS growth needed for an 8% annual price return
15	-1.8%	6.0%
18	0.0%	8.0%
25	3.3%	11.6%
30	5.2%	13.7%
45	9.6%	18.4%
60	12.8%	21.8%

Illustrative arithmetic only. Ignores dividends, buybacks and changes in share count. Required growth g solves (1+g)^10 = (starting P/E ÷ 18) × (1.08)^10 for the right-hand column.

The table shows why investors who buy high-P/E stocks are making a very specific bet. A P/E of 60 needs earnings to grow almost 13% a year for a decade just to stand still if the multiple normalises. That is not impossible; it is a high bar.

Common P/E ratio mistakes

Treating a low P/E as automatically cheap

A low P/E is often low for a reason. The market may expect earnings to fall, the company may face a structural decline, debt may be dangerously high, or last year's profit may have included a one-off gain that will not repeat. Investors call stocks that look cheap but keep getting cheaper "value traps." Strategies that systematically buy cheap stocks, such as the Dogs of the Dow, work only some of the time, precisely because some cheap stocks deserve to be cheap.

Treating a high P/E as automatically expensive

Some of the best-performing stocks of the past few decades traded on P/E ratios that looked absurd at the time. Amazon spent many years with little or no reported profit, which made its P/E either undefined or enormous, while it reinvested heavily in growth. A high P/E is a statement about expectations. It is expensive only if those expectations turn out to be too optimistic.

Ignoring negative earnings

When a company loses money, its P/E is negative, and most data providers simply display "N/A" or a dash. Loss-making companies are not "cheap" or "expensive" on a P/E basis; the measure does not apply. That is why many early-stage technology and biotech stocks are valued on revenue multiples or cash runway instead.

Comparing across sectors

As noted above, P/E ratios differ structurally across industries because of differences in growth, capital intensity, leverage and risk. Comparing a utility's P/E with a software company's will mislead you almost every time.

Forgetting about inflation and interest rates

In periods of high inflation, reported earnings can be flattered by inventory gains and understated depreciation, and investors demand higher returns to compensate for uncertainty. P/E ratios were very low in the late 1970s and early 1980s, when inflation and interest rates were in double digits; the Shiller data shows the market's CAPE at about 6.6 in July 1982. When rates fell, multiples expanded dramatically. Any P/E comparison across time needs to account for the interest-rate environment.

Mixing GAAP and adjusted earnings

If you compare one company's adjusted forward P/E with another's GAAP trailing P/E, the result is meaningless. Always check the definition. The gap can be large: companies with heavy stock-based compensation or frequent acquisitions often report adjusted EPS well above GAAP EPS.

Ignoring the share count

EPS can rise even if net income is flat, simply because the company buys back shares. It can fall even if net income rises, because the company issues new shares to fund acquisitions or pay employees. Always look at net income growth alongside EPS growth to see what is really driving the number.

P/E ratio versus other valuation measures

The P/E ratio is popular because it is simple and widely available, but it is one tool among several. Here is how it compares with the main alternatives.

Measure	Formula	Why use it	Limitation
P/E	Price ÷ EPS	Simple, universal, ties price to profit	Distorted by accounting, cycles, one-offs; useless with losses
Forward P/E	Price ÷ estimated EPS	Reflects expected profits	Depends on forecasts
PEG	P/E ÷ EPS growth %	Adjusts for growth	Growth estimates are unreliable; ignores duration
Price-to-sales	Market cap ÷ revenue	Works for unprofitable firms	Ignores margins entirely
Price-to-book	Price ÷ book value per share	Useful for banks and asset-heavy firms	Poor for intangible-heavy businesses
EV/EBITDA	Enterprise value ÷ EBITDA	Accounts for debt; compares firms with different capital structures	EBITDA ignores capital spending and taxes
Free cash flow yield	Free cash flow ÷ market cap	Focuses on actual cash generation	Volatile year to year for capital-intensive firms
Dividend yield	Dividend ÷ price	Simple income measure	Many companies return cash via buybacks instead

Professional analysts typically use several of these together. If a stock looks cheap on P/E but expensive on EV/EBITDA, for example, the explanation is usually debt: the company's equity is cheap because it carries a lot of borrowing. If it looks cheap on P/E but expensive on free cash flow, its reported profits may not be turning into cash, which deserves a closer look.

How to find and calculate a P/E ratio yourself

Almost every quote page shows a P/E ratio, but it is worth knowing how to check one, because providers use different earnings definitions and update at different speeds.

1. Get the share price. Use the latest closing price from the exchange or a reliable quote service.
2. Get diluted EPS. For the trailing figure, add the diluted EPS from the last four quarterly reports. For a fiscal-year figure, take diluted EPS from the latest annual report. In the US, all filings are free on the SEC's EDGAR system, and EPS appears on the income statement.
3. Check for unusual items. Read the earnings release or the notes to the accounts for large one-off gains or charges that distort the figure.
4. Divide. Price divided by EPS gives the P/E. One divided by the P/E gives the earnings yield.
5. Put it in context. Compare with the company's history, its peers and the market.

For an index or a fund, the process is the same in principle, but providers calculate it by aggregating the earnings of every component. That is why the S&P 500's P/E is quoted by S&P Dow Jones Indices and by data vendors rather than calculated by individual investors. If you invest through index funds, our comparison of ETFs and index funds explains how the vehicle you choose affects costs, though not the underlying valuation you are buying.

The P/E ratio for the whole market today

Readers often want to know whether the market is expensive right now. The S&P 500 closed at 7,712.57 on October 20, 2026. The index's official P/E depends on which earnings series you use, and the figures quoted by S&P Dow Jones Indices, FactSet and other providers differ because they use different combinations of reported, operating and estimated earnings. We do not quote a current market P/E here because we could not verify a single authoritative figure for this date against a primary source.

What we can say from the long-run data is this: the market has traded above its 150-year average P/E for most of the period since the mid-1990s, and the CAPE has been above 30 for long stretches of the past decade. History suggests that elevated valuations have tended to coincide with lower long-term returns, not with imminent crashes. Investors who use that information sensibly tend to adjust their expectations and diversification rather than abandon stocks altogether.

How traders use the P/E ratio

Long-term investors use P/E ratios to judge value. Shorter-term traders tend to use them differently.

• Earnings season reactions. High-P/E stocks are priced for strong results, so even a small earnings miss or a cautious outlook can trigger a large price drop. Low-P/E stocks sometimes rally on results that are merely "less bad" than feared. Knowing a stock's multiple helps you anticipate how sensitive it is to surprises.
• Rate sensitivity. High-multiple stocks tend to be more sensitive to changes in interest-rate expectations because more of their value depends on profits far in the future. When bond yields rise sharply, these stocks often fall more than the market.
• Sector rotation. Traders watch the gap in P/E between growth and value stocks, or between large and small caps. When that gap is historically wide, some look for opportunities in the cheaper group; when it narrows, momentum often fades.
• Sentiment cross-checks. Valuation is a slow-moving fundamental indicator. Traders often pair it with faster sentiment measures such as the put-call ratio to judge whether a stretched market is also showing signs of excessive optimism.

If you prefer to follow trade ideas rather than build your own valuation models, our free trading signals cover major indices, stocks, forex and commodities, with entry, stop and target levels published in advance. Signals are not personal advice, and every trade carries risk.

Limits of the P/E ratio: what it cannot tell you

Even used carefully, the P/E ratio has blind spots that are worth spelling out.

It ignores debt. Two companies with identical P/E ratios can have very different balance sheets. A highly indebted company may look cheap on P/E while carrying far more risk.

It ignores cash quality. Accounting profit and cash are not the same thing. Revenue recognition, capitalised costs and working-capital changes can all make earnings look better or worse than the cash a business actually generates.

It ignores capital needs. A company that must reinvest most of its earnings just to stay competitive is worth less than one that can distribute most of them, even at the same P/E.

It ignores duration. The P/E captures one year of earnings. It says nothing about how long a company's advantage will last, which is the main driver of long-term value.

It is affected by accounting rule changes. Changes to how companies account for goodwill, stock options, leases and investment gains have shifted reported earnings over time, which complicates historical comparisons, including the long-run Shiller series.

It cannot tell you about timing. As the CAPE study above showed, valuation has historically said something about decade-long returns and very little about the next few months.

Key lessons

The P/E ratio is a starting point, not an answer. It tells you how much the market is paying for each dollar of a company's or an index's current earnings, and by implication what it expects those earnings to become. The data supports a few clear conclusions:

• The S&P 500's long-run average trailing P/E is about 16 and its median about 15 (1871–2023). Since 1950, the average has been about 18.5.
• Extreme P/E readings can be caused by collapsing earnings as well as soaring prices. The record trailing P/E of about 124 came in May 2009, near a market bottom.
• Starting valuations have mattered over ten-year horizons. Starting from a CAPE of 10 or less, the median ten-year real return was 11.0% a year; from above 30, it was -1.1%.
• Valuation has said little about short-term returns. Use it to set expectations, not to time trades.
• For individual stocks, the P/E only makes sense alongside earnings growth, cyclicality, balance-sheet strength, peers and the quality of earnings.

Treat any single P/E as a question: what has to happen to the earnings for this price to make sense? Answering that question honestly is most of what valuation analysis is.

When you are ready to put this into practice, you can open a trading account with our regulated partner and track the stocks discussed here. Trading involves risk, and you can lose more than you invest in leveraged products.

Methodology and sources

Long-run market data. Monthly S&P Composite prices, dividends, earnings, CPI and CAPE come from Robert Shiller's online data file (ie_data.xls), downloaded from his Yale University page in October 2026. The edition available contained earnings through June 2023. Trailing P/E is the monthly average price divided by trailing twelve-month reported earnings as recorded in that file. Means and medians use every month with both values available (January 1871–June 2023).

Forward returns by CAPE. We built a real total-return index from Shiller's real price and real dividend columns, assuming monthly reinvestment of one-twelfth of the annual dividend. For each start month from January 1881 to June 2013, we computed the annualised real return over the following 120 months. Results are before fees, taxes and trading costs. Windows overlap, so they are not independent observations.

Company data. Diluted EPS for Apple, Microsoft, Nvidia, JPMorgan Chase, ExxonMobil and Tesla comes from the SEC EDGAR XBRL API (us-gaap: EarningsPerShareDiluted), selecting full-year values filed on Form 10-K. Closing prices for October 20, 2026 come from Yahoo Finance. P/E figures use the latest full fiscal year, so they differ from trailing-twelve-month figures on quote pages.

Illustrative tables. The worked examples and the required-growth table are hypothetical arithmetic and do not describe any specific company.

This article is for educational purposes and is not investment advice. Past performance is not a reliable indicator of future results.