Earnings determine stock prices. But which earnings? Those firms have generated in the past or those they are expected to generate in the future? Theory says the latter, but we only know the former with certainty. Fortunately, we can ask analysts what they expect. Recently, they have become exceptionally optimistic. Yet even considering such optimism, stock prices still appear high.
Finance 1.0.1 teaches you to value stocks by discounting firms’ expected future earnings. Nevertheless, we sometimes rationalise stock prices by looking at what firms have actually earned, i.e. their historical earnings. For instance, you often hear commentators arguing that Nvidia’s high stock price—having increased tenfold over the past five years—is justified because the company earns a ton of money. Similarly, I frequently use Robert Shiller’s CAPE (Cyclically Adjusted Price-Earnings) ratio. It takes the current real (inflation-adjusted) value of the S&P 500 and divides it by the average of real earnings over the previous ten years, i.e. it essentially compares current stock prices with historical earnings.
There is at least one good reason to look at historical earnings, even though theory tells us that stock prices should reflect expected future earnings: historical earnings are observable, whereas we cannot directly observe the expectations of the millions of investors out there who trade stocks and thereby determine stock prices.
However, one way to make progress is to ask professional analysts what they expect and aggregate their forecasts into a consensus earnings estimate. Let us see what these expectations imply for current stock market valuations.
Expected earnings
Stock market analysts regularly publish their best estimates of firms’ future earnings. Let’s look at analysts’ expectations for the earnings of the 500 companies in the S&P 500 over the coming year. This is typically referred to as the S&P 500 12-month forward EPS (earnings per share). Its historical development is shown in Figure 1. Not that it matters much for the story here, but I show it both in current prices and in real (inflation-adjusted, 2026) prices, because we are looking at a period spanning forty years during which the general price level has risen substantially.
Figure 1. 12-months forward earnings-per-share (EPS) of corporations in the S&P 500. Nominal and real EPS. Monthly data, 1985–July 2026. Source: Datastream via Refinitiv and J.Rangvid.
Over the past 40 years, analysts’ expected earnings per share for firms in the S&P 500 have increased by a factor of almost twenty, from a consensus estimate of around USD 20 per share in 1985 to more than USD 350 per share today. Adjusting for inflation, expected earnings per share have risen by a factor of six. That is impressive. It is, for instance, more than total economic activity in the US has grown over the same period. US real GDP has increased by a factor of three. No wonder stocks have had such a good run.
Growth in expected earnings
Currently expected earnings of USD 366 per share over the next 12 months may not tell you very much in themselves, so let’s instead look at earnings growth. This reveals an interesting story.
Figure 2 shows how much more—or less—optimistic analysts have become compared with six months earlier, i.e. the growth in consensus earnings expectations over successive six-month periods.
Figure 2. Six-month changes in 12-month forward earnings per share of the S&P 500. The red line marks the latest observation to facilitate comparison with earlier periods. Monthly data, 1985–July 2026. Source: Datastream via Refinitiv and J.Rangvid.
The figure shows that analysts have become much more optimistic over the past six months. In fact, they now expect earnings over the coming year to be 20% higher than they did six months ago. As the figure also shows, this is one of the largest upward revisions to earnings expectations we have seen, surpassed only during periods following deep recessions and the associated collapses in stock prices and earnings, such as the global financial crisis in 2008–09 and the pandemic in early 2020. In other words, analysts have become much more optimistic recently despite us not emerging from a stock market crash. Outside periods following major market crashes, earnings expectations have never been revised upwards by this much over a six-month period. Analysts have, in short, become much more optimistic about firms’ earnings prospects.
Stock market valuation based on expected earnings
Many people have worried that the US stock market has exhibited bubble-like behaviour, and I have even chipped in myself (link, link). Now, with earnings expectations having been revised upwards so dramatically, perhaps those fears are exaggerated? After all, if expected earnings have increased this much, there is surely a good reason why stock prices have risen as well, right?
Perhaps. Or perhaps not.
It turns out that even after taking these optimistic earnings expectations into account, US stock prices still appear high today.
Figure 3 shows the historical development of the S&P 500 relative to expected earnings per share over the coming 12 months.
Figure 3. S&P 500 relative to 12 months forward earnings per share. Monthly data, 1985–July 2026. Source: Datastream via Refinitiv and J.Rangvid.
As you can see, the price you have to pay for the S&P 500 today, relative to what analysts expect the companies in the index to earn over the next year, is high—certainly much higher than the historical average.
On average over the past four decades, the S&P 500 has traded at around sixteen times 12-month forward earnings per share. Today, the price-to-12-month forward EPS ratio stands at 20.4, around 28% above its historical average.
Expressed in yield terms, this corresponds to an expected earnings yield of around 5%, compared with a historical average of around 6.5%.
In short, the S&P 500 is trading at a high multiple today.
Historical earnings
It is relevant to compare this with a valuation of the stock market based on historical rather than expected earnings. Does that paint a different picture?
To answer that question, let’s return to CAPE, which I have used many times before (for instance: link, link, link). As mentioned at the beginning, CAPE takes the current value of the S&P 500 and divides it by the average of the previous ten years of real earnings per share for the companies in the index. An advantage of CAPE is that we can go back 150 years. The result is shown in Figure 4.
Figure 4. CAPE (cyclically adjusted price-earnings ratio). Historical peaks encircled in red to facilitate interpretation. Monthly data, Jan. 1881 – July 2026. Source: R.Shiller’s webpage and J.Rangvid.
When it comes to valuing the stock market, one critique of CAPE is that it can send misleading signals during periods when earnings are growing rapidly and stock prices are rising alongside them. The reason is that the ten-year average of earnings increases only gradually because rapidly rising current earnings are averaged together with older earnings that may have grown much more slowly. This causes CAPE to rise.
This is essentially the situation today. Companies exposed to the AI boom have experienced extraordinary earnings growth, and their stock prices have followed suit. But the ten-year average of earnings has not increased nearly as quickly, leaving CAPE at a high level, as Figure 4 shows.
It is important to note that Shiller was fully aware of this discrepancy between a slowly moving average and rapidly rising current earnings. In fact, he designed CAPE with exactly this in mind. His argument was that earnings are highly volatile, so a smoothed measure of earnings provides a better estimate of the ‘fair’ value of stocks than current earnings, which can be distorted by temporary booms and busts.
Whether you dislike CAPE for this reason or think Shiller was right, the bottom line is the same: the S&P 500 is trading at a high multiple today, whether we compare prices with expected future earnings, as in Figure 3, or with historical earnings, as in Figure 4.
Market timing
At this point, you may be getting worried. Even after taking historically optimistic earnings revisions into account, the S&P 500 still trades at a high multiple. But here comes an important caution: CAPE is not a good market-timing tool.
For example, CAPE has been rising for almost twenty years, since bottoming out in the spring of 2009. If you had sold your equities in 2014, when CAPE first crossed 25—a high level—you would have missed out on a decade of exceptionally strong stock market returns. Nothing tells us that the market has become exhausted when CAPE reaches 25, 30, 35, 40, or any other particular level. As long as the music keeps playing….
At the same time, we pay attention to CAPE and other valuation ratios because they cannot keep rising indefinitely. If they did, we would be in a bubble—and bubbles eventually burst. At some point, CAPE and other valuation ratios will stop increasing. Unfortunately, a large body of academic evidence suggests that when valuations contract, subsequent stock returns tend to disappoint. Historically, valuation ratios have not fallen because earnings growth accelerated sufficiently to justify high prices. At least, that is what the post-war US evidence tells us (link). Evidence from other countries is somewhat more supportive of earnings growth driving valuation-ratio contractions (link). But in this analysis, I look at the US stock market.
So, the reason I still like to look at CAPE is not that it necessarily tells us much about stock market movements over the coming months—or even the next few years—but because it is probably one of the best indicators we have of major stock market turning points, even if it cannot tell us when they will occur.
In particular, CAPE peaked in September 1929, just before the worst stock market crash in modern US history, as Figure 4 shows. It reached its all-time high in the spring of 2000, after which US equities lost around half their value. Today, it stands at its second-highest level ever, exceeding even its peak before the Great Depression. That is noteworthy.
Conclusion
Stock prices are determined by the future earnings that companies are expected to generate. Those future earnings are, of course, uncertain. For that reason, we sometimes compare stock prices with realised historical earnings instead. That may work reasonably well when earnings growth is modest, but it can be misleading when earnings are growing rapidly, as they have been recently.
Over the past six months, analysts have revised their earnings expectations for S&P 500 companies sharply upwards. Indeed, excluding periods immediately following major market crashes, this has been the largest upward revision in the past forty years. Nevertheless, whether we compare today’s value of the S&P 500 with historical earnings or with expected future earnings, the market still trades at a high multiple.
Academic research typically finds that analysts’ earnings forecasts are systematically too optimistic (link, link). If that is also true today, after earnings expectations have been revised upwards by a historically unprecedented amount, the market’s reaction could be painful. Being disappointed is bad; being disappointed after expectations have become exceptionally optimistic is even worse.
But, of course, perhaps this time really is different.
