Join our community of smart investors

The Halloween signal

Halloween is a strong buy signal.
October 31, 2017

We all know that our tastes change as winter approaches. We eat less salad and fish and more stews (and parkin!), and I listen to less swamp pop and more Townes Van Zandt and Leonard Cohen. This poses the question: if our tastes in music and food change with the seasons, why shouldn’t our taste for risk also do so, if only slightly?

Centuries of history tell us that it does. Halloween – and before that Samhain – marks a time of the year when people are anxious and fearful. And spring is a time of optimism, when “a young man's fancy lightly turns to thoughts of love.*”

What’s more, we know that people don’t sufficiently anticipate that their tastes will change. When we’re miserable it’s hard to foresee better times, and when we’re cheerful we don’t envisage tougher ones. We are prone to what George Loewenstein and colleagues have called projection bias; we project our current tastes into the future. A team of economists have shown that this costs us money. They’ve found that people pay more for houses with swimming pools and for convertible cars in the summer than in the winter. They don’t sufficiently anticipate that such goods won’t be so enjoyable in the winter, and don’t sufficiently anticipate in the winter that they’ll be worth having in the summer.

All of which prompts the question: if our tastes for risk change, and if we don’t anticipate a reversal of those changes, aren’t we likely to pay more for equities in the spring and less in the autumn?

Yes. Lisa Kramer, Mark Kamstra and Maurice Levi have shown that stock markets are swayed by variations in seasonal affective disorder; prices fall as the nights draw in and rise as they get longer. (And yes, the pattern is the same in the southern hemisphere). And Ben Jacobsen and Cherry Zhang have shown that in almost all national stock markets since they began trading returns have been higher from Halloween to May Day than from May Day to Halloween.

In the UK, for example, total returns (that is, including dividends) have averaged 8.6 per cent from Halloween to May Day since 1966 after inflation but minus 0.5 per cent from May Day to Halloween.

Of course, average returns disguise a large dispersion. But this dispersion is skewed; high returns are more likely in winter, and low ones more likely in summer. For example, the last 52 winters have seen only two losses of more than 10 per cent. But the last 52 summers have seen 11 losses of that magnitude. And winters have seen 21 gains of 10 per cent or more, whereas summers have seen only 11 such gains. 

All this tells us that prices are often too low in the autumn and too high in the spring – consistent with investors being too optimistic in the spring and too fearful in the autumn.

If we judge only by theory and evidence, “buy on Halloween” is one of the strongest rules of equity investing we have.

Nor is it the case that decent summers for equities cause this rule to break down. On each of the last four occasions when the All-Share index gained more than 10 per cent from May Day to Halloween (in 2003, 2005, 2009 and 2016), we had good returns in the following six months too – of an average of 6 per cent.

You might reply that the market doesn’t feel cheap right now.

That’s the point. Shares have usually been cheap in the autumn precisely because investors haven’t believed they were, and have instead erred on the pessimistic side.

As for what might drive prices up, there are some candidates. Among the more foreseeable ones (which are only a subset of the total) are: perhaps President Trump will get corporate tax cuts through congress; maybe tensions with North Korea will subside; global growth could continue for a while yet; rate rises in the US and UK might not spook shares as much as some fear; and perhaps Brexit negotiations won’t go as terribly as some expect. Yes, these are all maybes. But the point is that we’re at that time of year when we tend to underweight nice possibilities and overweight nasty ones.

All this makes me think I should be buying around now, even though other leading indicators of the market point to shares falling slightly over the next 12 months if not the next six..

I appreciate, though, that many of you remain sceptical. I think I sympathise. Instinctively, the “buy on Halloween” rule doesn’t feel right, despite the theory and evidence behind it. Personally, though, I don’t trust my instincts. These – not just on this but on many other matters – were formed in hunter-gatherer societies and might not be well adapted to life in complex advanced capitalist ones.

* In Tennyson’s poem the object of the young man’s love rejected him, which fits my story.