The CAPE ratio — the cyclically adjusted price-to-earnings ratio, also called the Shiller P/E — is one of the most widely watched measures of whether the stock market is cheap or expensive. Economist Robert Shiller popularised it, and it is part of the dataset he publishes. This guide tests how well it has lined up with the returns that followed.

What CAPE measures

An ordinary price-to-earnings ratio divides a stock index's price by one year of company earnings. Earnings swing a lot over the business cycle, so a single year can make the market look cheap at the peak of a boom or expensive in a recession.

CAPE smooths this out. It divides the price by the average of the past ten years of earnings, adjusted for inflation. A higher CAPE means investors are paying more for each dollar of long-run earnings.

Since 1881, the S&P 500's CAPE has ranged from below 5 to a peak of 44.2 in December 1999, with a long-run median of about 16.5.

CAPE and the next 10 years

We grouped every month from 1881 to mid-2013 by its CAPE level, and measured the S&P 500's real total return (after inflation, with dividends reinvested) over the following 10 years:

Starting CAPE Months Average real return, next 10 years Worst Best Chance of a real loss
Below 10 227 11.3% a year 1.8% 20.0% 0%
10 – 15 463 7.6% −4.2% 17.6% 9.7%
15 – 20 514 6.0% −4.6% 16.1% 8.9%
20 – 25 246 4.3% −4.0% 14.2% 25.6%
25 – 30 83 4.4% −1.2% 6.6% 6.0%
30 and above 57 −1.0% −5.9% 4.5% 59.6%

The pattern is clear. Cheap starting points were followed by strong decades, and expensive starting points by weak ones. Every 10-year period that began with a CAPE below 10 delivered a positive real return. More than half of the periods that began with a CAPE of 30 or more lost money after inflation.

Across all the months, the correlation between starting CAPE and the following 10-year real return was −0.52 — a meaningful relationship for something as noisy as stock returns.

Some famous starting points

Month CAPE Real return over the next 10 years
September 1929 32.6 −1.4% a year
January 1966 24.1 −1.8% a year
August 1982 6.6 +14.3% a year
December 1999 44.2 −3.2% a year
March 2009 13.3 +14.3% a year

The two best decades in this short list began when the market was cheapest; the three worst began when it was expensive.

What CAPE can't do

It is tempting to read this as a timing signal. The evidence does not really support that:

  • It says little about the next year or two. Expensive markets can keep getting more expensive for years. CAPE rose above 30 in 1997 and kept climbing until 2000.
  • The high-CAPE sample is small. Every month with a CAPE of 30 or more in this test comes from just two episodes, 1929 and 1997–2002. That is not much evidence to generalise from.
  • The range within each bucket is wide. Even from a CAPE between 20 and 25, the best decade returned 14% a year. A high CAPE shifts the odds; it does not decide the outcome.
  • The world may have changed. Accounting rules, tax rates, interest rates, and the mix of companies in the index have all shifted over 140 years. Some researchers argue that a "normal" CAPE today is higher than in the past.

How to use CAPE sensibly

  1. Use it to set expectations, not to time the market. When CAPE is well above its history, it is reasonable to plan with lower long-run return assumptions. In our calculator, that might mean choosing a more conservative risk profile when testing whether your plan still works.
  2. Pay extra attention around retirement. Starting withdrawals from an expensive market is risky because poor early returns do the most damage. All three historical failures of the 4% rule began with an above-average CAPE.
  3. Stay invested, but diversify. Rather than leaving the market, investors worried about valuations often hold more bonds, cash, or markets with lower valuations.

Notes

Data: Robert J. Shiller, U.S. Stock Markets 1871–Present, via datasets/s-and-p-500. CAPE (PE10) as published in the dataset, available from 1881. Ten-year forward returns are real total returns using monthly average prices, dividends reinvested and Shiller's CPI series; the last starting month with a full ten years of data is June 2013. The dataset's CAPE figures end in June 2023, when CAPE stood at 29.9. Calculations are our own.

This guide is educational and not investment advice. Past performance does not guarantee future results.