I did point out at the start of the year that short of an exogenous shock it was difficult to know what would derail the equity market. Such shocks are by their nature difficult to foresee but who would have thought that the catalyst for change in market thinking would have come in the form of something we cannot see but whose presence we are aware of – a veritable ghost at the feast? Equities have just posted their biggest weekly correction since 2008, and having experienced similar corrections in the past, I know the futility of trying to call the market bottom. The extent of market concern can be gauged from the VIX index of implied equity market volatility which has shot up to a level of almost 48 (recall that three weeks ago I expressed astonishment that it was running so low), taking it to its highest level since 2011 (chart below).
Whatever the longer-term health implications, there is
clearly going to be a period of intense economic disruption. It could last for days,
weeks or even months, but it is clearly going to impact on activity rates at
the end of February and into March. Such is the power of the unknown triggered
by the virus that face-to-face client meetings are being cancelled as
businesses test their disaster recovery procedures; Switzerland has banned
gatherings of more than 1000 people, with the result that two major trade fairs
including the Geneva Motor Show have been cancelled, and travel restrictions
are being ramped up. Naturally this will adversely affect corporate earnings,
which explains the collapse in markets over the past week (I would not like to
be in the insurance business at the present time). This raised a question in my
mind regarding the information content of the equity market collapse for events
in the wider economy. After all, investors focus on the slope of the 2-10 curve
in the bond market, but is there a corresponding equity indicator?
The information content of market corrections for the real economy
In order to assess the severity of the market collapse we need an indicator which measures both the extent and duration of the collapse. In order to do this, I looked at all trading days since 1940 and calculated those periods when the S&P500 declined for five consecutive sessions, and measured the resulting 5-day change in the index (I excluded the period 1928 to 1939 due to the volatility of the index over this horizon). I reduced the sample still further to select the subset of periods where the fall in the index cumulated to more than 7% (admittedly an arbitrary value). This resulted in 15 episodes (not counting the current one). To put some values on it, I measured the sum of peak-to-trough declines across all such episodes per calendar quarter. For the most part these are zero but in 13 cases there was one such event per quarter and in 1974 and 2009 there were two, resulting in index values of between -15 and -20 (chart below).
As a leading indicator, the index is by no means perfect. It has provided three false recession signals (1962, 1986 and 2015) and did not foresee the recessions of 1969-70 and 1980. But it did provide useful information in 1974, 2000 and 2008. In this sense it is not that much different from the 2-10 curve which often flashes false recession signals. And it may be possible to improve it by being more systematic about measuring the decline threshold.
The information content of market corrections for the real economy
In order to assess the severity of the market collapse we need an indicator which measures both the extent and duration of the collapse. In order to do this, I looked at all trading days since 1940 and calculated those periods when the S&P500 declined for five consecutive sessions, and measured the resulting 5-day change in the index (I excluded the period 1928 to 1939 due to the volatility of the index over this horizon). I reduced the sample still further to select the subset of periods where the fall in the index cumulated to more than 7% (admittedly an arbitrary value). This resulted in 15 episodes (not counting the current one). To put some values on it, I measured the sum of peak-to-trough declines across all such episodes per calendar quarter. For the most part these are zero but in 13 cases there was one such event per quarter and in 1974 and 2009 there were two, resulting in index values of between -15 and -20 (chart below).
As a leading indicator, the index is by no means perfect. It has provided three false recession signals (1962, 1986 and 2015) and did not foresee the recessions of 1969-70 and 1980. But it did provide useful information in 1974, 2000 and 2008. In this sense it is not that much different from the 2-10 curve which often flashes false recession signals. And it may be possible to improve it by being more systematic about measuring the decline threshold.
It would thus be too much of a stretch to suggest that the
equity market is pointing to a recession in the US, but given the expected
impact on activity as a result of what has been going elsewhere in the world,
some slowdown in growth is likely. Moreover, given the duration of the US
business cycle, which is the longest in recorded history, it may also be vulnerable
to shocks. One transmission mechanism from the market is the consumer wealth effect.
Estimates of this effect vary but a study produced by the IMF in 2008 suggested
that the long-run elasticity of US real consumption with respect to equities is
around 3.5%. In other words, each one dollar decline in the value of equity
holdings will reduce consumption by 3.5 cents. If the market holds at current
levels (13% down), this would imply a reduction of around 0.4% in consumption.
If this spills over into other assets, such as housing, the impact will be even
bigger since the US housing wealth elasticity of demand was estimated at 13.7%.
Is there value out there?
We are, of course, getting ahead of ourselves. Anecdotal evidence suggests that real money investors have not sold off to anything like the extent to which the headline index suggests. If true, it might indicate that the selloff has been exacerbated by algorithmic trading. An academic study published in 2017[1] suggested that the rise of exchange traded funds (ETFs), which are essentially passive investment funds which track the market, means that investors derive “lower benefits from information acquisition”, thus reducing their incentive to undertake it. This in turn reduces the efficiency with which investment decisions are taken and raises the risk that market swings may be larger than would otherwise happen in the event of a market where investors are forced to do their own due diligence. Once the dust settles, regulators will undoubtedly take a closer look at this issue given their mounting concerns over the impact of black-box trading models on market swings.
For now, however, investors are flying blind. Whether the coronavirus effect turns out to be a flash in the pan or a prolonged problem, the time for taking risks is over. As winter slowly gives way to spring, the next few weeks are going to be interesting. There is no doubt that the recent shakeout has taken a lot of air out of the balloon and on the basis of Robert Shiller’s long-run CAPE measure, we are now starting to approach less toppy valuation levels (chart below). This long-run P/E measure is now close to 27x versus 31x before the rout started. But if this is a trigger for a cyclical correction as in 2000-01, there could be another 10-20% market downside as the CAPE heads towards 23x.
Is there value out there?
We are, of course, getting ahead of ourselves. Anecdotal evidence suggests that real money investors have not sold off to anything like the extent to which the headline index suggests. If true, it might indicate that the selloff has been exacerbated by algorithmic trading. An academic study published in 2017[1] suggested that the rise of exchange traded funds (ETFs), which are essentially passive investment funds which track the market, means that investors derive “lower benefits from information acquisition”, thus reducing their incentive to undertake it. This in turn reduces the efficiency with which investment decisions are taken and raises the risk that market swings may be larger than would otherwise happen in the event of a market where investors are forced to do their own due diligence. Once the dust settles, regulators will undoubtedly take a closer look at this issue given their mounting concerns over the impact of black-box trading models on market swings.
For now, however, investors are flying blind. Whether the coronavirus effect turns out to be a flash in the pan or a prolonged problem, the time for taking risks is over. As winter slowly gives way to spring, the next few weeks are going to be interesting. There is no doubt that the recent shakeout has taken a lot of air out of the balloon and on the basis of Robert Shiller’s long-run CAPE measure, we are now starting to approach less toppy valuation levels (chart below). This long-run P/E measure is now close to 27x versus 31x before the rout started. But if this is a trigger for a cyclical correction as in 2000-01, there could be another 10-20% market downside as the CAPE heads towards 23x.
Brave investors will
likely step in at some point soon. As Warren Buffett, the grand old man of
value investing, once said, “Widespread
fear is your friend as an investor because it serves up bargain purchases.”
But Buffett also knows the value of waiting until the price is right.
[1] Israeli,
D., C. Lee and S. Sridharan (2017) ‘Is There a Dark Side to Exchange Traded
Funds? An Information Perspective’ Review of Accounting Studies (22), pp
1048-1083
