S&P Up 9% In A Year. Its Forward P/E Fell From 22.7 To 17.7
Mathematical Money | November 13, 2026
The S&P 500 is about 9% higher than a year ago. Over the same year its forward P/E fell from 22.7 to 17.7. FactSet says that is the lowest in three years.
Forward P/E is the price divided by the earnings analysts expect over the next 12 months. A year ago it sat 1.7 points under the dot-com peak. Today tech stocks as a group are at 17.3, which is below the index itself. Nvidia is at 10.5 and Coca-Cola is at 27.6.
So is the market cheap now? A forward P/E has two parts, a price and a forecast, and only the price is a fact. Before I call anything cheap I split it, and you can run the same test on any stock you own.
Split the P/E first
A P/E can fall two ways. The price drops, or the earnings number goes up. So the check starts with which one happened, and you can do it from two closing prices and two P/Es.
The S&P closed at 6,720.32 on 6 Nov 2025. On 5 Nov 2026 it closed at 7,353.73. That is up 9.4%.
The earnings forecast behind the P/E was about $296 a year ago. Now it is $415.82. That is up about 40%.
So none of the fall came from the price. All of it came from the earnings number, which rose about four times as fast as the index. On the 12 Nov close of 7,592.40 the same forecast gives 18.3.
Where the 40% came from
It came from two places, and they are not the same kind of thing.
The first is real earnings. A year ago analysts expected 2026 earnings to grow 13.7%. They are now on track for 33.3%. Third-quarter earnings are up 31.4% on a year ago, with 88% of companies reported.
Energy and chips did most of that. Energy earnings are up 149.4%, because oil averaged $85.68 in the third quarter. A year before it averaged $64.97. Chip-maker earnings are up 133%.
The second is the calendar. A year ago "the next 12 months" was mostly 2026. Today it is mostly 2027, and analysts have 2027 another 16.1% above 2026. Nothing was earned there. The window just moved forward into a bigger forecast.
Delivered, or still promised
This is the part that decides it, and the part worth checking on anything you hold. How much of that $415.82 has already been earned?
FactSet adds up what companies earn each quarter. The last four quarters come to $347.95. The next four are forecast at $406.08. So earnings have to grow about 17% from here for the forward P/E to be true.
Is 17% a lot? I took the growth analysts expected at the start of each year since 2010. The middle figure is 10.2%. Only two years started higher, 2010 and 2021.
Those two both delivered. Earnings grew 40.4% in 2010 and 48.8% in 2021, more than forecast both times. But both came straight after earnings had collapsed, and bouncing off a crash is the easy kind of growth. This time the 17% has to come on top of a 33% year.
The wider record is worse. FactSet counted the 25 years from 2000 to 2024. Earnings met the start-of-year forecast in 8 of them. The average miss was 6.2%.
The 39% trap
One number going around needs a warning. A Reuters column on 26 Oct was reported as saying the forward P/E only holds if earnings grow 38.6%. You get a figure like that by dividing the trailing P/E by the forward P/E.
The trouble is the two P/Es are not built from the same earnings. FactSet's trailing P/E is 24.7, which works out to about $298 of earnings. Its own last four quarters add up to $347.95. Same company, two different counts.
So the trap is mixing sources. If you compare a forecast with an actual, take both on the same count, from the same place. Do that and the gap is 17%. It is still high, but it is not 39%.
Against 2000
FactSet's forward P/E for the S&P peaked at 24.4 on 24 Mar 2000. A year ago we were at 22.7. Today's 17.7 is still above the 25-year average of 16.4, so this is not a bargain-basement number.
The leaders are the bigger difference. Goldman Sachs compared the seven biggest tech stocks of 2000 with today's seven, as reported by Benzinga in October 2025. The 2000 group traded at 52 times forecast earnings. Cisco alone was at 101.7.
The Magnificent 7 were at 26.8 in that table. I get about 20 for them now, using Yahoo Finance estimates and 12 Nov closes. By name:
Nvidia 10.5.
Meta 17.5.
Alphabet 19.4.
Microsoft 21.6.
Amazon 24.5.
Apple 33.7.
Tesla 184.1.
On the same source Procter & Gamble is at 22.7. So the biggest chip company in the world trades at under half the forward P/E of a soap maker.
P/E against growth
A lot of people go one step further and divide the P/E by the growth rate. That is the PEG. Lower means you pay less for each point of growth.
For the S&P that is 17.7 divided by 16.1. It comes to 1.1. A year ago the P/E was 22.7 and the growth forecast was 13.7%. That came to 1.7. By name, on Yahoo's forecast for the next fiscal year:
Nvidia 0.15.
Meta 0.9.
Microsoft 1.0.
Alphabet 1.1.
Amazon 1.8.
Apple 2.8.
Coca-Cola 4.9.
Tesla 7.3.
On that list Nvidia looks thirty times cheaper than Coca-Cola. Now run the split on both.
Nvidia earned $7.91 a share over the last 12 months. Its forward P/E uses $15.85. Half of that number has not happened yet. On what it has actually earned, Nvidia's P/E is 21.0.
Coca-Cola's P/E is 28.7 on what it earned and 27.6 on the forecast. Nearly all of it is already in the bank. If you have ever bought a stock because the forward P/E looked low and then watched the estimate get cut, you know which of those two numbers can move on you.
The PEG has the same weak spot. It puts the forecast in twice, once in the P/E and once in the growth rate. If the forecast is wrong, both halves are wrong in the same direction.
I could not find clean growth forecasts for the 2000 leaders, so I have no PEG for them. But Cisco at 101.7 would have needed earnings to double in a year just to score 1.
The P/E split, for any stock
You can run this yourself on the index or on one company.
1. Split the move. Take the price change over the year and the P/E change over the year. Whatever the price does not explain is the earnings forecast moving.
2. Split the forecast. How much has been earned in the last 12 months, and how much is still promised? Use the same source for both.
3. Compare the promised part with what is normal. For the S&P the usual start-of-year figure is about 10%. For a stock, use its own last few years.
4. Only then call it cheap.
The inputs are free. FactSet posts its Earnings Insight report every week, with the forward P/E and the quarterly earnings. For a stock, Yahoo Finance's Analysis tab shows next year's estimate next to the year-ago figure.
Cheap on forecasts, not on history
One more number, so nobody reads this as a buy call. The Shiller P/E uses ten years of real earnings and no forecast at all. It is 40.09 now. In March 2000 it was 43.22.
In the Fed post I wrote that S&P companies earn about $3.86 a year for every $100 of index price. The 10-year pays $5.10. That was on earnings already made. On the forecast it is $5.48, so the same market is under the bond on one count and over it on the other.
What I'm watching
I'm not calling it cheap and I'm not calling it a bubble. The whole case rests on one number, the 2027 forecast, which FactSet has at $424.96. If that starts falling while the price holds, the P/E climbs back and nobody had to buy anything.
Which growth number do you put under your own PEG, next year's or the five-year one? I used next year's because I can check it in twelve months. If there is a better one, I want to know why.
S&P forward and trailing P/E, forecast and quarterly earnings, sector figures and the oil price are from FactSet Earnings Insight, 7 Nov 2025 and 6 Nov 2026. The year-ago forecast of about $296 is the 6 Nov 2025 close divided by FactSet's 22.7. The forecast record is from FactSet, 5 Dec 2025; the 10.2% middle figure is my arithmetic on FactSet's yearly tables, 17 year-starts. The 24.4 peak is from FactSet, 21 Feb 2020. The Goldman Sachs figures are as reported by Benzinga, October 2025. The Reuters column is by Joachim Klement, 26 Oct 2026, read through a syndicated copy. Company P/Es and growth forecasts are Yahoo Finance consensus at the 12 Nov close; fiscal years differ by company, and Alphabet's and Amazon's trailing earnings include one-off investment gains. The Shiller P/E is from multpl.com. S&P closes are from yfinance. Everything else is my own arithmetic.
#SPX #valuation #earnings #PEratio #investing
Stop guessing. Start calculating.
Live to fight another day. 🤙
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

