Should investors ignore the bubble-watchers?

A Harvard study reveals clues to help investors avoid stock-market crashes

Markets exhibiting sharp price increases have a “substantially heightened” probability of crashing, a Harvard study says.

Nobel economist Robert Shiller, who famously spotted the dotcom bubble that burst in 2000 and the housing bubble that climaxed in the global financial crisis, is once again warning that the US market is "way overpriced".

On the other hand, Eugene Fama, who shared the 2013 Nobel award with Shiller, essentially argues that investors should chill out, once saying that most bubbles are "20/20 hindsight" and that there is no reason to think that what goes up must come down.

Who is right? Do most market bubbles lead to crashes? Do sharp price increases predict unusually low returns going forward? And can investors profit by getting out in time, or should they ignore the bubble-watchers?

That question is addressed in a new Harvard study, Bubbles for Fama, which examines Fama's contention that the word "bubble" is a "treacherous" term best ignored by investors.

READ MORE

Fama's belief in efficient markets can seem puzzling. In 2010, he rejected the idea there had been a US housing and credit bubble. He has also defended internet valuations in the late 1990s, prompting behavioural economist Richard Thaler to quip that Fama was "the only guy on Earth who doesn't think there was a bubble in Nasdaq in 2000".

As the Harvard study points out, however, Shiller first said markets were displaying irrational exuberance in December 1996, and prices proceeded to double over the next three years.

Prices eventually crashed, but even at the market low in 2003, US prices remained higher than when Shiller first called the bubble.

Furthermore, commentators tend to focus on past bubbles that were followed by crashes, but this neglects the fact that many market booms don’t end in tears.

Essentially, Fama contends that investors should trust the markets, saying there is no reason to fear rapidly rising sectors. To test that thesis, the Harvard researchers examined US stock returns since 1926 as well as international data covering 31 countries between 1987 and 2012. A bubble was defined as a period where a sector doubled in price over a two-year period; a crash was defined as a fall of at least 40 per cent in the subsequent two-year period.

Soaring sectors

Sector bubbles are not uncommon; the researchers found 40 such episodes in the US and 107 in the international data. Importantly, there is no truth to the idea that post-bubble returns are invariably awful. Fama is “mostly right” when he says sharp price increases do not predict subsequent subnormal returns.

There are two reasons for this. Firstly, many soaring sectors continue to gain. US healthcare stocks, for example, more than doubled between 1976 and 1978; and instead of crashing, they continued to rise by more than 65 per cent annually over the next three years. Secondly, “bubble peaks are notoriously hard to tell, and prices often keep going up, at least for a while, before they crash, leading to good net returns for an investor who stays all the way through”.

However, this doesn’t mean investors should be casual about buying into high-flying sectors. Markets exhibiting sharp price increases have a “substantially heightened” probability of crashing. More than half of the US industry bubbles – 21 of the 40 cases – subsequently suffered a share price crash.

Non-US markets were little different; 53 of the 107 bubbles were followed by crashes. A crash is “nearly certain” in sectors witnessing especially rapid gains; when a US industry rose by at least 150 per cent over a two-year period, a subsequent crash occurred in 80 per cent of cases. Among non-US sectors that recorded 150 per cent run-ups, prices crashed two-thirds of the time.

For investors, the obvious question is, why do markets crash in some instances but not in others? Is there a difference between some “good” bubbles and “bad” bubbles?

Tell-tale signs

There are some tell-tale signs that allow investors to distinguish between industries likely to crash compared with sectors likely to continue gaining.

There is, the researchers note, “much more to a bubble than a mere security price increase”. Crash candidates are much more volatile than stocks that tend to avoid crashes, experiencing a notable increase in volatility in the period leading up their peak.

They are also much more likely to issue stock and to register higher-than-usual numbers of initial public offerings (IPOs); to give one extreme example, almost half (48 per cent) of stocks in the software industry in 1999 had either gone public or issued stock over the previous year.

At-risk sectors are more likely to feature younger stocks; price-earnings ratios are likely to be higher; share price acceleration is likely to be especially rapid as the rally peaks.

Spotting these distinguishing features is a valuable exercise that can beat a buy-and-hold approach by more than 10 percentage points over the following years, according to the study. Importantly, industry bubbles tended to coincide with broader market rallies that were similarly fragile.

“When an industry bubble is called correctly, it is best to avoid the stock market altogether,” the authors say, rather than switching one’s money to a different stock market sector.

Still, bubble peaks are “extremely hard to call”, and betting against bubbles remains a risky task, even if you correctly identify that a crash is coming. During the dotcom era, the German stock market first crossed into bubble territory in March 1998; anyone who got out then would have had to watch the index soar for another 23 months. On average, sectors that enjoyed gains of at least 100 per cent over a two-year period continued gaining for another six months. Gains averaged 30 per cent over this period, confirming that the final stage of market bubbles tends to be an especially euphoric affair.

So what are the takeaways for ordinary investors? Well, although Shiller tends to be widely lauded for having called past market peaks and Fama is often derided for his atheistic attitude to bubbles, the latter is right to say that booms should not necessarily be associated with subsequent busts. On average, high-flying sectors do not underperform in subsequent years.

However, the more prices rise, the greater the probability of a subsequent crash. Furthermore, timing bubbles is possible. Not all high-flying sectors are the same, with crash candidates different in some crucial respects. Spotting these differences will help investors to outperform by avoiding market crashes.

However, while timing is possible, it’s a tricky task. Actively betting against bubbles is, as the authors caution, best left to those with “extremely deep pockets” and a “high tolerance for volatility”.