Past improvements in the predictive power of the P/E ratio

Empirical evidence has shown that short-term stock market forecasting is a perilous undertaking. However, some valuation indicators, and the traditional price/earnings (P/E) ratio in particular, have shown a modest historical ability to forecast long-run returns.

In the late 1990s, professors Robert Shiller and John Campbell proposed an adjustment to improve the traditional P/E ratio’s predictive power. Instead of dividing a market’s current stock price by the average earnings per share generated over the prior four quarters, they divided it by the average inflation-adjusted earnings per share generated over the prior 10 years. That modification captures the earnings power of the market over a business cycle better than a single year of earnings does.

As the figure below illustrates, the Shiller CAPE (cyclically adjusted price-to-earnings) ratio provided a fairly accurately forecast of 10-year-ahead returns—for a while. Since the mid-1990s, however, the predictive power of this metric has deteriorated. Contrary to expectations, the ratio has failed to revert to its long-term average for long periods. As a result, the metric has projected lower returns than those actually produced by the stock market. The notable absence of a reversion toward the Shiller CAPE’s long-run average has raised doubts about its continuing usefulness in forecasting stock returns in real time.  

Another step forward for forecasting: Vanguard’s “fair-value” approach

Vanguard Investment Strategy Group’s ”fair-value” CAPE approach started with an observation: A long-term decline in interest rates and inflation depresses the discount rates used in asset-pricing models. The upshot is that investors are willing to pay a higher price for future earnings, thus inflating P/E ratios. The critical implication is that there is no reason for the CAPE ratio to revert to its long-term average. Instead, it should be expected to revert to a level that reflects current economic conditions.

To test our hypothesis, which breaks with the standard assumption that the CAPE will mechanically revert to its fixed long-run average, we used real interest rates, expected inflation, and measures of financial volatility as proxies for the state of the economy to arrive at a “fair value” for the CAPE ratio.

More about our methodology and calculations can be found in our paper, Improving U.S. Stock Return Forecasts: A “Fair-Value” CAPE Approach, published in the Winter 2018 issue of the Journal of Portfolio Management. The results in the chart below strongly suggest that adjusting the “long-term average” P/E ratio to account for macroeconomic conditions results in more accurate real-time projections of actual 10-year-ahead returns. 

The fair-value CAPE has proved better at forecasting
1970 through 2016

Notes: For the real-time analysis, the regression coefficients are determined recursively, starting with 10-year trailing annualized returns from January 1901–December 1959 data and re-estimating the regression coefficients with the addition of data for each month thereafter.
Source: Davis, Joseph, Roger Aliaga-Diaz, Harshdeep Ahluwalia, and Ravi Tolani, 2018. Improving U.S. Stock Return Forecasts: A “Fair-Value” CAPE Approach. Journal of Portfolio Management 44 (3): 43-55. © 2018 Institutional Investors LLC. All rights reserved.
It’s worth noting that our fair-value CAPE would appear to explain both elevated CAPE ratios and robust stock returns over the past two decades.

What the different measures tell us now

At the end of 2017, the Shiller CAPE was flirting with all-time highs, surpassed only by the peaks that preceded the collapse of the dotcom bubble. It stood at about 33; its historical average is about 16. A reversion to the ratio’s long-term average would bode ill for future stock returns.

Vanguard’s fair-value CAPE ratio paints a less alarming picture. It suggests that valuations are indeed high but not in the “bubble” territory implied by the conventional CAPE. We expect the returns of U.S. stocks to fall below their historical averages over the next 10 years. The most likely outcome, according to our projections, is annualized returns of 3%–5% for the coming decade.

That outlook is subdued, but not nearly as bleak as those suggested by tools that fail to account for today’s low interest rates and inflation.


All investing is subject to risk, including possible loss of principal.

IMPORTANT: The projections or other information generated by Vanguard’s fair-value CAPE approach regarding the likelihood of various investment outcomes are hypothetical in nature, do not reflect actual investment results, and are not guarantees of future results.