By James Mackintosh
Oct. 19, 2017 2:06 p.m. ET

19 COMMENTS

Everyone’s a China-watcher these days, as red-blooded capitalists try to extract policy clues from hourslong speeches by aged Communist leaders.
A more fruitful pursuit for investors might be to remind themselves of the two big lessons from recent history in China’s markets: Being early is as good as being wrong, and buying into a bubble doesn’t usually work out well.
Investors who bought into Chinese stocks when they first started listing in Hong Kong in the early 1990s can congratulate themselves on anticipating years of powerful economic growth. They can also kick themselves if they stuck with the trade, as mainland companies listed in Hong Kong, known as H-shares, or those incorporated outside the mainland but listed in Hong Kong, called red chips, were both lower a decade later.
The best way to profit from being early to China was similar to the smart way to trade new themes today: Sell out as soon as everyone else catches on to the idea. In 1996-97, everything China-related soared, with red chips peaking at 250% year-over-year gains around the handover of Hong Kong to China before collapsing as the Asian Tiger bubble popped.

Mature themes are more about earnings prospects and less about bubbles. But investors thinking about three big themes—bitcoin and blockchain; artificial intelligence, or AI; and treatment of aging—would do well to learn from the China experience.
The three are on a spectrum of frothiness. At the clearly bubbly end is anything cryptocurrency-linked, in which the joyous echoes of bubbles past are being heard by millennials. As in 1999 with dot-coms, a mere change of company name to something blockchain-related can boost the stock price.
Twenty-somethings are raising tens of millions of dollars for mad-sounding projects with flaky business models, celebrities are backing startups, and regulators are watching with horror but doing little. Real-world projects wanting no-strings-attached financing are piling in—one even proposes a digital currency exchangeable into sand—and due diligence is rare. The bubble is well advanced already.
Much less extreme is the excitement about AI. Real companies are making money from real products using machine-learning tools, but much of the money is being made by software engineers able to charge pretty much what they want for their ability to code using new branches of statistics. Companies have realized that they need to know—and talk about—machine learning, but valuations of the stocks involved are merely high, not wildly implausible.


At the far opposite end is aging. This isn’t exactly novel, with the search for immortality dating back thousands of years. But companies are starting to get serious about extending life, with Google, now called Alphabet, starting research in 2013. With the science at an early stage, many of the ventures being set up will fail, and most of the biotechnology companies involved are still private anyway.
Jim Mellon, a biotech investor and author of “Juvenescence,” says it is probably too early for the general investor to get into the small stocks he thinks might make it big, as their products may well fail. “But I have no doubt that in three or four years’ time there will be a speculative bubble in this area,” he says.

The problem for investors is timing such bubbles. When bubbles pop, losses can be very large and very rapid. A year after peaking in 1997, Chinese red chips were down 85%; the Hang Seng China Enterprises Index of H-shares was lower in January this year than at its peak in 1993.
But working out when a bubble will pop is difficult, to put it mildly. Richard Taffler, a finance professor at Warwick Business School in the U.K., says bubbles tap into “unconscious fantasies,” so gauging investor emotion is key, albeit hard. “All one can do is measure the underlying sentiment and the extent to which reality intrudes,” he says.

Part of the problem is that each bubble is different. Work by Boston fund manager GMO LLC a few years ago found bubbles varied widely both in size and duration. After meeting GMO’s statistical bubble test of two standard deviations above trend, the time required to reach the peak was almost four years for large Japanese companies up to 1989, but less than six months for the South Sea Bubble of 1720 or the first emerging-market bubble, in 1825.
GMO found that bubbles in emerging-market stocks, on average, pop more quickly and don’t get as big as in developed countries, and that the biggest bubbles occur in commodities. The 1974 sugar bubble remains one of the biggest ever.
Bitcoin has already jumped 10-fold in just over a year, and other cryptocurrencies have had even more extreme moves, so if they pop, they might take the record. True believers argue bitcoin is different, and its price rise is justified because it is the future of money. The history of bubbles suggests that’s unlikely but not impossible: GMO found that one in seven price moves that look like a bubble turns out to be a paradigm shift.
Investors trying to spot the next big theme need to beware of thinking they are early when they are late, as they might be with bitcoin. They also need to be ready to sell if it looks like a bubble is developing and to have the strength not to be sucked back in if the bubble keeps going for a long time. Both are hard to do, but no one ever said investing was easy.