It is likely that the S&P 500 Technology Hardware, Storage & Peripherals indexis in a bubble at risk of bursting.
That’s the conclusion I draw from a study by three Harvard University researchers: Robin Greenwood, a finance and banking professor who chairs the institution’s Behavioral Finance and Financial Stability project; Andrei Shleifer, an economics professor; and Yang You, a Ph.D. candidate. In their study,“Bubbles for Fama,”published in the January 2019 issue of the Journal of Financial Economics, they analyzed U.S. stock market history back to 1926 in search of ways to forecast a bubble that was about to burst.
Applying the formula the researchers derive, I calculate there is an 80% chance that the Technology Hardware, Storage & Peripherals index will be 40% lower than today at some point in the next two years. Among some of the better-known firms in this industry are Apple(ticker: AAPL),Seagate Technology(STX), and Western Digital(WDC).
Though no other industries satisfy the researchers’ definition of a bubble, two others come close. They are also in the technology arena: Semiconductors and Semiconductor Equipment, and Software.
Why focus on an industry that may be in a bubble, rather than the market as a whole? Prof. Greenwood toldBarron’sthat he and his fellow researchers learned from their study of the history of bubbles that they “rarely are marketwide” events. Far more common, he said, is for a bubble to manifest in certain pockets of the market even as other sectors remain undervalued.
This was certainly the case at the top ofthe dot-com bubble, the mother of all bubbles. Greenwood reminds us that, even as dot-com stocks soared to outrageous valuations in the late 1990s and early 2000, other sectors of the market—notably value stocks—were either fairly valued or even undervalued. Some of those other sectors actually gained ground during the bear market that accompanied the bursting of the dot-com bubble stocks.
The researchers define a bubble to be any industry whose two-year return is at least 100 percentage points greater than the overall market’s. This is a high standard indeed—among all industries for which they had performance data between 1926 and 2016, just 40 satisfied the definition at any point over this 90-year period.
Not all bubbles burst, of course, and those that do don’t always burst right away. The researchers imposed a strict precondition here as well: Once an industry satisfied their definition of a bubble, they considered it to have burst if, within the subsequent two years, it lost at least 40% of its value. Of the 40 industries that satisfied the researchers’ definition of a bubble, 21—or 53%—burst.
What this means, assuming the future is like the past: There’s a slightly better than one out of two chance that any industry that outperforms the market by 100 percentage points in any two-year period will lose 40% or more over the subsequent two years.
The researchers also studied how the probabilities of a crash changed when they tightened or loosened their definition of a bubble. When they set the criterion to be just 50 percentage points ahead of the market, instead of 100, the odds of a crash fell to just 20%. When they tightened the criterion to 150 percentage points, the probabilities of a crash rose to 80%.
This latter probability is what applies to the Technology Hardware, Storage & Peripherals index. Over the past two years, according to FactSet, it has outperformed the S&P 500 by 151 percentage points.
One is tempted to apply the professors’ formula to individual securities, as I myself have done in the past. In November 2017, for example, I used the formula to argue thatthe odds of Bitcoin crashingwere greater than 80%. It lost 67% over the next 12 months. I used the professors’ formulaonce again in February 2020to argue that the odds of Tesla(TSLA) crashing were 80%. The stock lost 59% over the next six weeks.
Since then, of course,TeslaandBitcoin have skyrocketed, as have any of a number of other highflying assets. Should I once again forecast that there is a high probability of their crashing, I asked Greenwood? He demurred, stressing that further research is needed into the various factors that affect the odds of an individual stock crashing.
Yet he added he believes that not only is the broad stock market overvalued, there are individual pockets of the market that are “incredibly frothy and bubbly.”