CAPE Fears Miss the Point

Joseph G. Paul

With the recent wobbles in the market, scary headlines about stock bubbles are back in the spotlight. We think all this foreboding overlooks an important point: In today’s low-return world, where better than stocks to put your money?

We’re particularly reacting to a recent article in the New York Times, entitled “Time to Worry About Stock Market Bubbles”, featuring comments from Yale professor Robert Shiller, who shared the Nobel prize in economics last year for his work on asset prices. Shiller observes that, based on his cyclically adjusted price earnings (CAPE) ratio—which looks at current stock prices relative to an inflation-adjusted average of the past 10 years of earnings—stocks have been this expensive only thrice in the past: the 1920s (before the Great Crash of 1929), the late 1990s (the Internet bubble) and 2007 (the eve of the global financial crisis). Those are disturbing comparisons.

In a preemptive strike, the article notes that stock boosters always say, “It’s different this time,” and that they typically cite counterarguments about the muting of the business cycle or the advent of the Internet to justify exceptionally high valuations. As we all know, the stock boosters have been proved wrong time and again.

We offer another perspective. If you look behind the averages and into the historical drivers of earnings and the multiples that investors are willing to pay for those earnings, the picture looks less grim.

First, by our calculation, Shiller’s current CAPE multiple of 25 times implies adjusted earnings for the S&P 500 Index of about $75, a full 30% below the market’s current trailing earnings. A drop of that magnitude would take a recession-level collapse in revenues and margins. Is that a plausible expectation?

Much has also been made of the fact that profit margins are at record levels. But margins have been rising for nearly 25 years, and many of the factors levitating them—globalization, lower borrowing costs and lower debt balances—are unlikely to reverse any time soon.

Second, we point to a critical driver of earnings multiples: the discount rate used to value earnings.  Discount rates are related to the prevailing rate of interest. Today, rates are well below average, whether long term or recent. Lower rates argue for higher multiples.

Shiller isn’t completely wrong—stocks do look expensive relative to history. We simply believe that this point doesn’t matter nearly as much as most people think. We cannot go back and buy stocks in the past. So rather than asking “Are stocks attractively priced relative to history?” investors should be asking “Are stocks attractively priced relative to my other investment choices today?” And the answer to that question is yes. From that vantage point, we’d say that stocks remain the best game in town.

Our forecast for the 10-year return of global stocks is 7.1% a year, based on our proprietary forecasting model, which considers both historical and current economic, business and capital-market conditions.1  This is materially below our “normal” expectation of about 9.2% a year (Display 1). But returns are low everywhere. Our 10-year expected annual return from bonds is 2.8%, well below the “normal” expectation of 6%, reflecting today’s ultralow interest rates, the starting point of our investment horizon.

No one likes low returns. But in a world where returns are below normal across every asset class, stocks shine. We see this in the equity risk premium (ERP)—the extra compensation that investors demand for taking on the greater risks of owning equities over bonds. Based on our 10-year projections shown above, the spread is 4.3%, or roughly a third higher than the long-term average. Based on current 12-month forward earnings yields (a proxy for stock returns), the global ERP is 7.9%, or nearly double the long-term average of 4.3% (Display 2).

Low returns are a fact of investment life today. It’s important that investors adjust their perspective to accommodate this new reality and evaluate stocks relative to their alternatives—not simply relative to history—to improve the chances of meeting their long-term goals.

1The AllianceBernstein Capital Markets Engine is a Monte Carlo model that simulates 10,000 plausible paths of return for each asset class and rate of inflation, producing a probability distribution of outcomes; however, it goes beyond randomization by taking the prevailing market conditions into consideration and basing the forecasting on the building blocks of asset returns, such as inflation, yield spreads, stock earnings and price multiples. The analysis incorporates the linkages that exist among the returns of the various asset classes and factors in a reasonable degree of randomness and unpredictability. There is no guarantee that the returns presented will actually be realized.

The views expressed herein do not constitute research, investment advice or trade recommendations and do not necessarily represent the views of all AllianceBernstein portfolio-management teams.

MSCI makes no express or implied warranties or representations, and shall have no liability whatsoever with respect to any MSCI data contained herein. The MSCI data may not be further redistributed or used as a basis for other indices or any securities or financial products. This report is not approved, reviewed or produced by MSCI.

Joseph G. Paul is Chief Investment Officer of US Value Equities at AllianceBernstein (NYSE: AB).


  1. Joe Smith

    I think Cliff Asness wrote something on the lower discount rate thing(Fed Model)- even though from an asset pricing prospective the market does value securities based on RF+ERP it doesnt have any predictive power.

    Re CAPE- you have a months in ”09 where corporate earnings were under $10 on the S&P 500, that”s the real problem w/ CAPE. That”s not normal, and that”s the whole point of CAPE, to normalize earnings- you need to throw those out to normalize the normalized earnings, so its the CAPE PE not the T12 P/E that”s skewed right now.

    • Joseph G. Paul

      To your comment about the lack of predictive power of the equity risk premium: It depends on what period you’re looking at and the market conditions in play at that time. We did some research into the frequency of US stocks beating US Treasuries over one-, three-, five- and 10-year rolling periods since the 1940s, based on 10-year cyclically adjusted S&P 500 earning yields (an inversion of Shiller’s CAPE) . We found that when the earnings yields were above the 10-Year US Treasury yield, the odds of stocks beating bonds were very high, especially for long holding periods [Display]. This is precisely the point we were making in our blog. Based on a Shiller CAPE of 25x, the current earnings yield is 4%, far surpassing the Treasury yield of 2.5%.

      Still, as we said in our post, we agree with Shiller’s basic premise—that stocks look expensive relative to history, which from a predictive point of view suggests that future returns will be subpar. But the potential for equities still looks far better than that for bonds.

      You also point out that that the 10-year earnings used to calculate CAPE have been distorted by the Great Recession, but the multiple we apply to those earnings may not be. However, higher multiples would be justified if interest rates stay low for longer, as we expect. Stocks also look attractive relative to bonds if rates rise because of a stronger economy, which would bolster earnings and cause bond prices to fall. Falling profits and low rates would favor bonds but it seems unlikely. The only really harmful setting for stocks would be if the economy weakens while rates rise—but that would be equally bad for bonds. Bottom line, we think stocks offer better potential than bonds almost any way you slice it.

  2. This reminds me of a friend who bought Microsoft near the end of 1999 because it was one of the best relative values in the technology sector–which it probably was, but it still slumped by almost two thirds within one year. Absolute valuations do matter, and historic patterns must be respected. Those who buy U.S. equities today are going to get exactly what they suffered following similar peaks in 2000 and 2007.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>