If you had switched from shares into cash every May Day, and then switched back into shares on Halloween, £100 invested 30 years ago would have grown to over £5,000 by now. The same sum left in equities all the time would have grown to just over £2,800. If you'd begun seasonal investing five years ago, you'd have made almost 25 per cent while the All-Share made nothing. A seasonal investor would be wondering what all the talk of financial crisis and recession is about.
These numbers pose the question: why do so many investors ignore the May Day sell signal? We know they do because if people had sold at the end of April - in anticipation of others selling the following day - share prices would fall and so April's returns would be bad. But, in fact, April is the best month for the market. Since 1966, the All-Share has returned 3.1 per cent on average in April, compared with an average of 1.2 per cent for all months.
The very fact that there are seasonal patterns in returns shows that investors don't take advantage of seasonal investing.
We can't blame the dealing costs of selling in May and buying again in Halloween for this. For one thing, many investors are happy to incur high costs, by either trading actively or by holding costly actively managed funds. And for another, seasonal investing need not be costly. Many unit-linked pension schemes allow customers to switch between equity and cash funds at no cost. And investors don't need to sell shares to reduce their exposure. They can instead buy put options as insurance.
Instead, I suspect there are other reasons why people aren't seasonal investors.
One is the same reason why such seasonal patterns exist in the first place. The likeliest explanation for such patterns is that our appetite for risk follows the same trends recognised by our pagan ancestors. As the days get longer in March and April, we become more optimistic and look forward to days of sun and plenty. Such optimism drives share prices up – often too far up, to levels from which they subsequently fall back. Hence the May Day sell signal. Conversely, as the nights draw in in the autumn we become fretful about the future, which pushes shares down to levels from which they often recover. Hence the Halloween buy signal.
But here's the thing. These same changes in optimism cause investors to ignore the May Day sell signal. They figure instead - being optimistic - that there are reasons to stick with shares. And because evidence about future returns is always scant and ambiguous, you can always read it in an optimistic way; if the economy looks like weakening, you can take comfort in the prospect of monetary easing or in the possibility that shares are discounting bad times.
Amplifying this seasonal optimism bias is another cognitive bias - base rate neglect, our tendency to ignore background probabilities. This mistake can be costly. It can cause companies to make disastrous takeovers, because they ignore the background probabilities which tell us that mergers often fail. And it can cause people to over-invest in newly-floated companies, oblivious to their high probability of doing badly. In similar fashion, investors pay so much attention to the futurological runes that they forget the background probabilities which tell us that the odds turn against shares in the summer. For example, in the last 46 years, the All-Share index has fallen 16 times between 30 April and 31 October, but only seven times between 31 October and 30 April.
Another error that causes people to ignore the May Day sell signal is a form of the representative heuristic. Forecasting future asset returns looks very complicated because there are so many factors to consider; economic conditions, money flows, investor sentiment, valuations, and so on. And it's natural to suppose that the solution to a complex problem must itself be complex - that is, that solutions must be representative of problems.
But this is wrong. As Gerd Gigerenzer at the Max Planck Institute has shown, simple rules of thumb can give adequate answers to complex problems.
For example, if you want to catch a cricket ball you could solve the equations which describe its motion. Or you could follow the rule: 'keep your eye on the ball and your hands soft'. Most cricketers prefer the latter. 'Sell in May' works the same way. But our urge to over-intellectualise investing stops us seeing this.
There's a further problem – a mix of the heuristic of social proof and regret aversion.
If, for the above reasons, nobody is a seasonal investor then it is very easy to suppose that there must be good reasons for this. It’s difficult not to go along with the herd.
One reason why it's difficult is that the psychological costs of being wrong are asymmetric. If we follow others and stay in the market and lose money we can comfort ourselves with the knowledge that everyone else made the same mistake. But if we break with the crowd and sell and then see the market rise, we'll kick ourselves. When he said that "it is better to fail conventionally than succeed unconventionally", Maynard Keynes was exaggerating – but not perhaps by much. The desire to minimise regret causes us to stick with the crowd and not sell in May.
Herein, though, lies a problem. While 'sell in May' has worked well on average over the long run – the All-Share index has returned an average of only 1.4 per cent between May Day and Halloween since 1966, which is worse than cash – it does not work all the time. There have been 12 summers in the last 46 years when the All-Share rose more than 10 per cent – the last in 2009. There is, therefore, a reasonable chance that 'sell in May' will be wrong. But then, all decisions about the future have a good chance of being wrong, and it is not obvious that 'sell in May' is worse than the alternatives.
