Shiller P/E is a useful market-timing tool Commentary: Popular valuation method can help measure stock exposure
NORTHVALE, N.J. (MarketWatch) — According to many, the Shiller P/E ratio has proven to be a decent measure of predicting future 10-year returns (at least in the U.S. and U.K. markets). When the number is high, future 10-year returns tend to be muted; when the number is low, future 10-year returns tend to be bountiful.
Click to Play
Has the market rallied too far, too fast?
The stock market's race to all-time highs has impressed and surprised many on Wall Street. But even bulls are wondering whether the rally is due for a fall. Jonathan Cheng reports.
But Shiller P/E has its detractors. The recently published Credit Suisse Investment Returns Yearbook, for example, contains an article arguing that using valuation metrics such as the Shiller P/E to determine stock exposure is ineffective. Read more; What Shiller P/E says about market's top.
As a reminder, Shiller P/E (popularized by Yale University economist and professor Robert Shiller and known more formally as CAPE ratio or P/E10) is the current price of the Standard & Poor’s 500-stock index SPX -0.55% divided by the past decade’s average inflation-adjusted earnings of the constituents of the index.
The long-term average Shiller P/E is 16, meaning investors have paid 16x past 10-year inflation-adjusted average earnings to own the market. Read more: Brace yourself for meager stock returns.
According to the article in the Credit Suisse handbook, however, we have the benefit of hindsight when looking at metrics like the Shiller P/E. In other words, in, say, the late 1990s, we wouldn't have known that the Shiller P/E would have screamed into the 40s, previously uncharted territory for the metric. Using the Shiller P/E as a signal to sell stocks in, say, 1997 would have caused investors to exit well before stocks peaked in price.
We only know what's cheap and expensive with the benefit of hindsight. Because of this problem, the authors assert that a simple buy-and-hold strategy or a dollar-cost-averaging strategy would have done better than letting Shiller P/E dictate your exposure to stocks.
Moreover, Shiller P/E and a Price/Cyclically Adjusted Dividend Yield metric (P/D10), which the Credit Suisse authors substitute (given their preference to the cold, hard cash of dividends over the accounting vagaries of net income), doesn't seem to work worldwide — though there is strong evidence of the predictability of future returns in the U.S. and U.K.
But don’t overburden the Shiller P/E. The metric is quite useful, as long as you don’t use it to get totally in or totally out. Selling in the mid-or late 1990s, for example, would have caused some pain. However, good investors at that time didn't simply dump stocks and move to cash.
Rather, they held some cash and also bought small-cap stocks, which back then looked comparatively much cheaper than large-cap stocks. In retrospect, it seems that valuation-conscious managers such as Steve Romick of FPA Crescent FPACX -0.33% and Don Yacktman of Yacktman YACKX -0.42% and Yacktman Focused YAFFX -0.35% were proved correct.
The Credit Suisse article doesn't examine whether cheap areas of the market can be reliably identified in this way. It also doesn't discuss the possibility of moderating one's exposure — never selling out completely or going all in, for example.
One can (and probably should) use Shiller P/E to adjust one’s exposure more gently, perhaps never going below 25% stock exposure or above 75% stock exposure as Benjamin Graham suggested for more sophisticated investors not locked into a policy portfolio. An investor can pick other prearranged parameters, but the important thing is to use a valuation metric to adjust one’s exposure on a scale rather than in the heavy-handed (completely invested or completely in cash) way the authors suggest.
The issue of the "Gambler's Fallacy" that the piece raises is important as well. Just because you've flipped a coin heads 20 times in a row doesn't mean the odds are any different than 50/50 that it will come up heads again on the 21st flip; there’s no reason to expect any reversion.
Similarly, investors shouldn't think a raging bull market can't continue to rage. However, a certain amount of common sense dictates that there are limits to how long and how much stock prices can become divorced from business fundamentals.
A coin flip, completely independent of flips that have come before it, is not quite the same thing as a stock price, which can’t stay disconnected from fundamentals forever or whose risk, given its detachment from fundamentals, may be judged to some extent.
At least in the U.S. and U.K. the evidence shows that some reasonable risk-reward profile of the market can be measured. It would have been hard to predict a Shiller P/E in the 40s in early 2000 when the previous high was in the mid-30s in 1929, but that doesn’t make the Shiller P/E or the authors’ P/D10 utterly useless at all.
John Coumarianos is a managing member of Hamilton Research and Management, an investment advisory firm in Northvale, N.J.
10 signs your stocks are about to tumble