MarketWatch Premium

July 24, 2021, 9:45 a.m. EDT

The price-to-earnings ratio is flawed as a stock market indicator. This is how to make it more accurate

By Mark Hulbert

There is a simple way to improve the forecasting ability of the price-to-earnings ratio.

That’s good news because, despite perhaps being the most widely followed valuation indicator on Wall Street, the P/E ratio has a mediocre track record, at best. In fact, when tested on data back to 1871 from Yale University professor Robert Shiller , the P/E’s ability to forecast the S&P 500’s subsequent total real return is only barely statistically significant, regardless of whether one is focusing on the subsequent one-, five- or 10-year periods.

The modification that improves the P/E’s forecasting ability is to base its denominator — the “E” — on earnings in the best three of the trailing four calendar quarters.

That change, which an alert reader recently asked me to take a look at, was suggested last year in the Journal of Portfolio Management . Its authors are Thomas Philips, an adjunct engineering professor at New York University, and Adam Kobor, managing director of investments at NYU.

The motivation for their change is earnings’ extreme volatility from quarter to quarter, which causes the P/E ratio to itself be quite volatile and thereby reduces its ability to forecast the stock market’s long-term subsequent return. The authors suggest a modified P/E ratio in which the denominator is equal to the sum of the best three of the trailing four quarters — multiplied by 4/3 in order to make the altered denominator similar in magnitude.

This modification doesn’t always lead to major differences. Currently, for example, the S&P 500’s (S&P:SPX) traditional P/E ratio is 28 versus 27.5 for the modified version. But sometimes there’s a huge difference. At the March 2009 bear market bottom, for example, the traditional P/E was over 100, while the modified version stood at 18.8.

The table, below, shows the improvement in the P/E ratio’s track record when calculated in this modified way. The tests were run based on data back to 1871 from Yale’s Shiller. (The values in the table are the r-squareds of the correlations.)

  Extent to which original P/E ratio explains or predicts the S&P 500’s real total return Extent to which modified P/E ratio explains or predicts the S&P 500’s real total return
During subsequent year 0.3% 1.6%
During subsequent 5 years 1.0% 6.5%
During subsequent 10 years 5.4% 17.4%

Note carefully that the authors of this recent Journal of Portfolio Management article suggest other ways as well to improve on the P/E ratio’s track record, and interested readers should consult that article.

In the meantime, these results reinforce their conclusion that you can improve the P/E’s track record by ignoring the worst of the trailing four quarters’ earnings.

You will notice from the table, below, that I have replaced the traditional P/E ratio with this modified P/E ratio. While the modified P/E ratio doesn’t paint quite as bearish a picture as the traditional one, it is still suggesting that the stock market is overvalued.

This table, which I report each month in this space, reviews the status of various valuation indicators with good track records predicting the stock market’s 10-year returns.

  Latest End of last month Beginning of year Percentile since 2000 (100% most bearish) Percentile since 1970 (100% most bearish) Percentile since 1950 (100% most bearish)
P/E ratio (based on the best 3 of the trailing 4 quarters) 27.50 27.30 24.40 81% 87% 91%
CAPE ratio 37.98 37.05 33.76 100% 96% 97%
P/Dividend ratio 1.35% 1.36% 1.58% 97% 95% 96%
P/Sales ratio 3.08 3.06 2.76 100% 100% 100%
P/Book ratio 4.58 4.55 4.05 100% 97% 97%
Q ratio 3.38 3.33 2.64 99% 100% 100%
Buffett ratio (Market cap/GDP ) 2.06 2.03 1.82 100% 100% 100%
Average investor equity allocation 49.5% 49.5% 47.6% 100% 98% 99%

 Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at .

Link to MarketWatch's Slice.