Thursday, October 12th, 2017

A Bolt from the Blue

Can a bear market in US equities emerge out from nowhere? Can it strike like lightning from a clear, blue sky, without any warning signs from within US financial markets? The answer, unfortunately, is yes, but only if the cause is outside the US. A quick read of the literature surrounding the 1987 crash suggests that there is no agreed narrative about what caused it, though there is plenty of agreement that programme trading in US equities made it worse. Having been in the market at the time, our recollection relates to US oil tankers being attacked by Iran, a simmering foreign exchange crisis and a huge storm in the UK, none of which could be easily forecast by looking at US financial markets.

However, when we look at the last two crashes, 2000-03 and 2007-09, we find that the behaviour of volatility (particularly the difference between equity volatility and bond volatility) can act as a useful warning signal. By definition, US volatility is not the best predictor of trouble elsewhere in the world; we will look at bear markets in other countries in the near future. The focus of this note is on US equities and the probability of another “home-grown” bear market, like the last two. Other things aside, these tend to fall harder and last longer. The 1998 LTCM crisis, originally caused by a Russian debt-default, doesn’t even qualify for our definition of a bear market – total return down 20% from the peak on a weekly close.

Using the weekly close definition allows us to date the 2000 bear market from the beginning of September, not March which is the date of the Tech bust. This is not just sophistry; for several months many investors believed that there was no read-across from Nasdaq to the broader market. In the same way, in late 2007 they did not believe that a problem with the securitisation of US mortgages could cause a general financial crisis. This explains why looking at excess volatility can be helpful.

In both examples there was period of several months when investors knew there was a problem, but could not be sure of its severity. Equities were past their peak but had not entered bear market territory. During this time equities became progressively more risky compared with bonds. Eventually excess volatility rose to a point where equities were not generating an adequate risk-adjusted return under any reasonable assumptions. This led to a disorderly sell-off which took the index into bear market territory. This process can be iterated two or three times as in 2000-03, or it can happen catastrophically as in 2007-09.

In both cases it was the preceding increase in excess volatility which mattered. An increase of more than 10 percentage points seems to be the critical threshold: from 8% in April 1999 to 22% in July 2000 and from 3% in April 2007 to 14% in April 2008. Without this rise in excess volatility it is very hard for a correction to turn into a full-scale bear market. Regarding the current situation, excess volatility is now 2%. A rise of 10 percentage points would it take from the top of the first to the middle of the fourth quartile (since 1974). It has happened before, and it will happen again, but it won’t happen overnight. A correction could happen at any time. Converting that to a full-scale bear market will require a significant increase in excess volatility and that cannot happen without a warning signal.

Leave a Reply

Full Blog Archive:


Other Recent Posts