CAPE, and what it does not tell you: known limitation 1

Using it to time an entry or an exit. It has stayed elevated for years at a time while markets continued to rise, and a person who sat out on the strength of the level would have missed a great deal. Nothing in its record supports acting on it in the short run.

CAPE, and what it does not tell you: known limitation 2

Comparing a level today with a level from decades ago as though the measurement were unchanged. Accounting standards, the treatment of intangibles, the prevalence of buybacks and the sector composition of the index have all moved, and each of those shifts the ratio for reasons that have nothing to do with expensiveness.

CAPE, and what it does not tell you: known limitation 3

Comparing one country's CAPE with another's and concluding one is cheap. Different markets have different sector mixes, different accounting and different rates of profitability, and a lower ratio is often a correct price for a different set of businesses rather than a bargain.

Playbooks / Valuation

CAPE, and what it does not tell you

The cyclically adjusted price to earnings ratio divides price by ten years of inflation-adjusted earnings to smooth the business cycle out of the denominator, and it carries useful information about long-run returns while carrying almost none about the next year.

The rule

A price to earnings ratio has a defect that is easy to miss and hard to live with: the denominator collapses in a recession. Earnings fall, the ratio spikes, and the market looks most expensive at roughly the moment it is cheapest.

CAPE fixes that by dividing price not by one year of earnings but by the average of the last ten, adjusted for inflation. Averaging across a cycle takes the cycle out of the denominator, and what is left is a number that can be compared across long periods.

That is the whole idea, and it is a good one. What follows is what it can and cannot support, because the gap between those two is where almost all the misuse lives.

What it is reasonably good at

Ranking decades. Historically, buying when the ratio was high has been followed by lower real returns over the following ten years, and buying when it was low by higher ones. The relationship is real and it is not subtle.

The rough translation is the inverse: a CAPE of 30 implies something like three point three percent real. That number is the centre of a very wide range, not a forecast, and reporting it without the width is the same failure as quoting a gross rental yield and calling it a return, which is the point of net rental yield.

Used that way it is genuinely useful. It changes what a person assumes when they are deciding how much to save, what to withdraw, and how long a plan needs to work, which is where it belongs: in the same conversation as safe withdrawal rate, where the starting valuation is one of the reasons the honest answer is a range rather than a number.

What it is bad at

Timing. Completely.

It has been elevated for years at a stretch while markets kept rising. A person who moved to cash on the level alone would have been out for a long time, and being early is indistinguishable from being wrong for as long as it lasts. Nothing in its record supports acting on it in the short run, and its most prominent advocates have generally not claimed otherwise.

This is worth being blunt about because the ratio is most often cited in exactly the situation it cannot help with: somebody wants a reason to do something now.

The objection that has to be answered

A framework that ignores the strongest argument against it is advocacy.

The measurement has not stayed constant. Accounting standards have changed, the treatment of intangibles has changed, buybacks are far more prevalent than they were, and the sector composition of the major indices looks nothing like it did in the middle of the last century. Each of those moves the ratio for reasons unrelated to whether shares are expensive.

So a level today and a level from 1950 are not quite the same measurement, and anyone drawing a horizontal line across a hundred years of the series and declaring today extreme is comparing things that have drifted apart. This does not make the ratio useless. It makes cross-era comparisons weaker than they look, which is a smaller and more defensible claim than either side usually makes.

The same caution applies across countries. A market with a lower ratio may be cheaper, or it may hold a different mix of businesses with lower profitability, which is a correct price rather than a bargain. Comparing Dubai and London runs into the identical problem in property.

How to actually use it

As an input to expectations, and nothing else.

If the starting valuation is high, assume less. That flows into what a plan requires, and it makes the difference between a plan that survives a mediocre decade and one that assumed the good one. If it is low, the same in reverse.

And apply the discipline in reverse the assumption before leaning on it: ask what would have to be true for this level to be justified. Sometimes the answer is nothing plausible, and that is informative. Sometimes it is a change in the composition of the market that has already happened, and that is informative too.

The other thing it should never be combined with is a bond yield, in the hope that the pair of them produce a signal neither has on its own. That is the Fed model, and the spread has no forecasting record the plain ratio did not already have.

What the ratio should never produce is an action this quarter. A measure built by averaging ten years of earnings is not a signal about the next three months, and using it as one is a misreading of the tool rather than a fault in it.

The arithmetic

The ratio CAPE = Price / Average of the last ten years of earnings, inflation adjusted Why ten years A single year's earnings collapses in a recession, so a plain P/E rises when a market is cheapest and falls when it is most expensive. Averaging a cycle removes that perversity. What it is reasonably good at Ranking decades. High starting CAPE has historically gone with lower ten year real returns, and low with higher. What it is poor at Timing anything. It has been elevated for years at a stretch while markets rose. The honest way to use it As an input to expectations, not a signal to act on. Expected real return, roughly = 1 / CAPE A CAPE of 30 implies about 3.3%. Treat that as a centre of a wide range, not a forecast. The objection that has to be answered Accounting has changed, buybacks have changed, and the composition of the index has changed, so a level from 1950 and a level from today are not quite the same measurement.

Where it breaks

Using it to time an entry or an exit. It has stayed elevated for years at a time while markets continued to rise, and a person who sat out on the strength of the level would have missed a great deal. Nothing in its record supports acting on it in the short run.
Comparing a level today with a level from decades ago as though the measurement were unchanged. Accounting standards, the treatment of intangibles, the prevalence of buybacks and the sector composition of the index have all moved, and each of those shifts the ratio for reasons that have nothing to do with expensiveness.
Comparing one country's CAPE with another's and concluding one is cheap. Different markets have different sector mixes, different accounting and different rates of profitability, and a lower ratio is often a correct price for a different set of businesses rather than a bargain.
Treating the implied return as a forecast. The inverse of the ratio is the centre of a very wide distribution of outcomes, and reporting it without that width is the same error as quoting a gross yield without saying it is gross.
Abandoning it because it has been wrong recently. The claim it supports is about decades, and evaluating a decade-scale claim on a three-year window is the mistake it is most often accused of making in the other direction.

When to use it

When setting expectations for a long horizon, and when somebody uses a valuation level as a reason to act now. It belongs in the same conversation as a withdrawal rate or a savings rate, not in a conversation about this quarter.

Sources

Last reviewed 28 August 2026. Educational research, not personal advice. Disclosure standards.