A Recession Says Nothing About Future Stock Returns
(Thanks to Morningstar’s John Rekenthaler for including one of my emails to him in a column consisting of reader reponses while he tended to his wife who, as he reports, suffered a fainting spell. We wish both of them well, of course.)
Do we need a recession or another credit event similar to 2008 to tell us stocks are overpriced and cause them to tumble? John Rekenthaler of Morningstar seems to think so. I sent him an email in response to an article he wrote doubting the verdict of recent bubble-callers like GMO and Research Affiliates. I said stocks were objectively expensive (using the Shiller PE), and that meant future returns would likely be low.
But John thinks that a turn in the economic cycle will determine a downturn in the stock market, and tell us, after the fact, if stocks are overpriced. Since we don’t know when that will occur or what it will look like, we must remain agnostic as to the future returns of the stock market. As he responds to my email in a new article:
“One of these years the economic cycle will turn, thereby making projected corporate earnings wildly overstated rather than moderately so. Stocks will get crushed. If that happens in 2018 or 2019, then equity prices will indeed have been high, and returns will indeed be low. If the economy holds out until 2020 or longer, though, then today’s values should look reasonable.”
Unfortunately, while stock markets tend to tumble when the economy goes South, since the Great Depression there’s scant evidence that single recessions tell us anything about how stocks are priced or indicate anything about their future 10-year returns. For that all-important forecast, one must consult starting valuations more than recessions or moment in the economic cycle.
Consider the 50% decline the S&P 500 Index suffered from 2000 through most of 2002. The recession in 2000 was minor. In fact, it didn’t’ even meet the standard definition of two straight quarters of GDP contraction. GDP contracted in the second quarter of 2000, then again in the fourth quarter of that year, and never again.
Did that recession warrant a 50% price reduction in stocks? Did it somehow prove that stocks were overpriced? Or were stocks just wildly overpriced to begin with, as the Shiller PE hit 44 in early 2000?
The point isn’t that we may or may not have a recession over the next 2, 3 or 5 years. The point is stocks are at a Shiller PE seen only twice before in history – 1929 and the run-up to 2000. Come recession or not, over the next decade investors in the S&P 500 will capture a 2% dividend yield. They may also capture 4%-5% earnings-per-share growth. That puts nominal returns at 6%-7%, which isn’t bad at all. Unfortunately, the third component of future returns consists of where the future PE ratio will sit. Will the Shiller PE maintain itself above 30? Or will it contract to something resembling the historical average of nearly 17? Even if that average is outdated, is the new norm 32? Or is it more like 20 or 22?
Whether a recession comes within the next 5 years or not has little to do with these questions. And though it may send stocks down for most of its duration, it ultimately will have told us nothing about longer term returns compared to how much starting valuation can tell us. In fact, the two features of the Shiller PE are that it’s based on a prior decade’s worth of earnings and is pretty good at forecasting the next decade’s worth of returns. It’s not based on short-term earnings, and it’s not good at forecasting short-term stock returns. A recession doesn’t matter one whit insofar as it’s a typical part of a full cycle that the Shiller PE aims to capture in its earnings calculation and in its stock return forecast.
It’s possible we might wake up in a decade to a 32 Shiller PE. And it may have remained there all along, or it may have arrived there again as the result of any number of gyrations. The question is what should financial writers be telling their readers (and financial advisers telling their clients) about that possibility?