Bear markets and crashes are very different things. Bear markets can sometimes be identifiable and often have a recognisable pattern. It may be possible to see the conditions that precede them developing, just not exactly when they will occur.
Market crashes are a very different thing. They remain mysterious and are to market historians much like black holes are to theoretical physicists. They are extremely difficult to predict.
For instance, the most famous crash (1929) occurred when the US stock market was very expensive (Shiller CAPE at around 31x), while the 1987 crash occurred when the market was ‘cheap’ by today’s standards. The Shiller CAPE at the time was around 18x compared with around 40x today. Valuation on its own has historically been of little use in predicting a crash.
Fortunately, crashes are relatively rare. What is most interesting about them is that while they are a mass participation event (i.e. many sellers), even the people living through them often struggle to explain exactly what happened. This is itself fascinating and suggests that there may be (like with black holes) some important hidden truth here.
Bear markets, by comparison, can be more understandable. Valuation can matter and bear markets have historically displayed patterns, and a type of anatomy, that can be studied.
So, what causes a bear market? There is so much that could be written about this – but for an article of around this length there are some important historical observations that stand out.
Bear markets have often occurred sometime after a valuation spike or peak. This occurred ahead of the Great Depression, the 1970s bear market and the bear market of the early 2000s.
They have often followed either a wave of technological innovation (radio in the 1920s, photography and TV in the 1960s, the internet in the 1990s) or a fixed-asset investment boom (housing in the period leading up to 2007). A common characteristic has been overinvestment in a key and economically significant sector.
Unlike a crash, which can take place very quickly, bear markets can take years and are rarely a straight-line event. Historically, major bear markets have often taken several years to complete. One reason is that overinvestment in the booming sector needs to be worked through or reversed. This can contribute to a decline in earnings growth and may involve insolvencies and, in some cases, banking-sector problems, as occurred around 1929 and 2008.
Bear markets have also often ended after much of the valuation increase associated with the preceding boom has reversed. A good example is the period from the 1990s into the early 2000s. The internet-related boom started in the mid-1990s, when the internet really started to take off. The Shiller PE [price-to-earnings ratio] was around 21–22x. The multiple peaked at around 43x in 2000 and subsequently returned towards the 21–22x level.
So where are we now?
A number of the characteristics historically associated with bear markets are present today. There has been a significant valuation increase (the Shiller PE is currently around 40x), and there has also been a very large investment boom associated with AI.
What might happen next – and how long could it take?
If some of the major AI projects ultimately turn out to represent significant malinvestment, this could contribute to a future bear market.
ChatGPT was launched in November 2022. At that time the Shiller PE was around 28x. A return from the current level of around 40x to 28x would, all else being equal, imply a market decline of around 30%. In practice, of course, earnings and valuations change over time, so this should be regarded as an illustration rather than a forecast.
Another way of looking at this is that around the time ChatGPT was launched, the US market was substantially below today's level. A return to those levels would therefore represent a significant decline. Again, this is an illustration of valuation risk rather than a prediction of where the market will trade.
Historically, major bear markets have sometimes taken three years or more to play out. An investor who believed that a bear market was beginning could simply decide to remain out of equities for an extended period.
The obvious difficulty is that this could mean missing substantial gains if that judgement proved wrong, or during the significant rallies that can occur within bear markets.
Bear markets are multi-year processes and can involve substantial swings in both directions before the whole process is complete. This creates both risks and opportunities for active investment managers.
An investment approach that pays particular attention to valuation, market history and downside risk may seek to navigate these periods by reducing exposure to significant market falls while retaining the ability to participate when conditions improve.
This type of environment is particularly relevant to the investment philosophy we have developed at Woodhill. We will see what happens next, of course. The future is inherently uncertain, but current valuations and the scale of investment associated with AI make this an unusually interesting period for investors.
Paul Wood is a fund manager at Woodhill Asset Management. The views expressed above should not be taken as investment advice.