There’s a way of looking at the cyclically adjusted price-earnings ratio that makes the Standard & Poor’s 500 Index seem a little less expensive.
That’s the view of economists at the Federal Reserve Bank of Cleveland, who say adjusting the signal by applying a moving average could be useful in normalizing the valuation indicator, a brainchild of Robert Shiller.
“This is an attempt to get at what the underlying trend really is,” Joseph Haubrich, a vice president at the Cleveland Fed and one of the study’s authors, said by phone. “If you think the underlying growth or trend moves around over time, this would say we’re not amazingly far from the historical experience, and maybe things aren’t particularly high.”
The CAPE, as it’s known, differs from regular P/Es by using 10 years of earnings data instead of one. It’s been ringing alarm bells for the past three years and is at “one of the highest levels ever in history,” Shiller, the Yale University finance professor and Nobel laureate, said in April.
Central to the view that the ratio is portending catastrophe is its position today compared with history. Using 10 years of earnings data that encompass the financial crisis, the S&P 500 trades at a ratio of more than 26 -- way above its 134-year average of 16.6 and at a level that in the past signaled subpar returns in equities.
Not so fast, says Haubrich. Comparing the cyclical P/E to an average gleaned from more than a century of stock fluctuations fails to account for changes in the economy. What once was an appropriate valuation for equities might have changed, making old ways of thinking about stock prices obsolete.
To account for this, Haubrich proposed a different comparison: get a mean value for the CAPE that resembles a moving average weighted to smooth the impact of business cycles, using something known as a Hodrick-Prescott filter. With that, you can establish a range for the CAPE that isn’t anchored in distant history, but to movements that give greater influence to its recent levels.
In that context the CAPE’s level today looks less alarming -- it’s within one standard deviation of Hodrick-Prescott mean, which currently goes as high as 28.7, the study shows.
“There’s always the danger of choosing the statistic that fits your priorities, but this is a way of presenting another tool to look at,” said Haubrich. “I see it as taking the CAPE philosophy to the next level.”
Shiller, the technique’s creator, said the Cleveland Fed’s model is similar to an enhancement known as relative CAPE used in some exchange-traded funds that, among other things, employs a 20-year average valuation.
“When I apply the relative CAPE measure to predicting the U.S. stock market since 1901, I find that relative CAPE works similarly to CAPE itself, though somewhat less well,” Shiller said. “The idea is that there are differences across countries or industries or through time in what earnings mean and what kind of accounting is used.”
The Cleveland Fed’s analysis “sounds plausible,” he said in a phone interview. “It could be a good thing to do. I still think the CAPE ratio looks high and I’m worried about it, but on the other hand the issue is that interest rates are at rock bottom so the alternative from moving from stocks to bonds doesn’t look very attractive.”