Bulls are on shaky ground when they argue that low-interest rates justify today’s sky-high P/E ratios.
While they’ve been making that argument for some time, it took on new urgency recently when the stock market rose to yet another new high. The S&P 500’s P/E ratio at that high, based on trailing 12-month earnings, stood at 23.7, higher than 93% of monthly readings since 1871 (according to data from Yale University finance professor and Nobel laureate, Robert Shiller). This otherwise would suggest the market is quite overvalued -- unless low-interest rates can come to the rescue.
The notion that they can has a long history, of course. In the 1990s, for example, this notion was known as the Fed Model. Though the Federal Reserve never endorsed it, Wall Street analysts tried to claim that the model had the Fed’s seal of approval, based on a single paragraph in a Fed report to Congress in 1997. Noting that the 10-year Treasury yield was far higher than stocks’ earnings yield (the inverse of the P/E ratio), that paragraph argued that stocks were overvalued.
Though today few use the name “Fed Model,” the underlying theory remains widely popular. Hardly a day goes by without a talking head on television or in one of the myriad investment newsletters I track referring to it in one form or another.
Cliff Asness, founding principal at AQR Capital Management, has said that he suspects the Fed Model’s popularity in recent years derives from it being almost always bullish. That makes the model invaluable as a sales tool for Wall Street brokers trying to convince nervous investors to put more money into stocks.
To appreciate why the Fed Model will be bullish for the foreseeable future, consider that today, even with the P/E ratio higher than 93% of monthly readings since 1871, the S&P 500’s earnings yield stands at over 4%. The Treasury 10-year yield (currently 1.6%) would have to rise higher than that to turn the Model bearish -- a prospect that seems out of the question for the short and intermediate terms -- if not longer.
If Treasury yields remain low, the only other way the Fed Model could turn bearish would be for the market’s earnings yield to drop to an even lower. That would be equivalent to a P/E ratio of over 60 -- a level not seen even at the top of the Internet bubble in early 2000.
A perpetually bullish model hardly seems like a market timing model worth paying attention to. And that is what a more comprehensive analysis of the historical record also found.
To appreciate what that more comprehensive analysis found, consider a statistic known as the r-squared. It measures the degree to which one data series (in this case, the Fed Model -- the market’s earnings yield minus the 10-year yield) explains or predicts another (in this case, the stock market’s subsequent 10-year return). The r-squared’s highest possible reading is 1.0, while a 0.0 reading would mean that the indicator has no explanatory or predictive ability whatsoever.
The table below reports what I found when analyzing the Fed Model’s track record back to 1871. Notice that the earnings yield by itself does a far better job forecasting the market’s subsequent returns than the Fed Model.
|Forecasting 10-year returns||Forecasting 1-year returns|
Earnings yields minus 10-year yield
Earnings yields minus 10-year yield
In other words, not only does forecast accuracy not improve when taking interest rates into account, it actually gets worse -- much worse.
And that, in turn, means that high P/E ratios are bearish for stocks, whether or not interest rates are low.
The bottom line? Interest rates don’t enable you to wriggle out from underneath the force of equities’ extreme overvaluation. While you nevertheless are free to argue that the bull market will continue, you can’t base your argument on interest rates being so low.