“Sell in May”: time to spring clean your investments?

It’s the month of May, which means that the investment press is pumping out the obligatory spew of articles on the wisdom (or otherwise) of adopting a “sell in May” strategy – also known, in the US, as “the Halloween effect”. The basic idea is that markets historically do better in the winter months (end-October to end-April) than in the summer months. Therefore, you should dump your stocks in early May, sit in cash (or short-term bonds), and not buy back in until pumpkin-carving season is over.

Is there any validity to the idea? Perhaps surprisingly, the data says “yes”. Since 1950, according to Ned Davis Research, the S&P 500 index has gained an average of 1.4% (2.1% using the median average) between May and October, while from November to April it’s gained 7% (6.6% median). The hit rate is better for the winter months too – the market has risen four times out of five, compared with three out of five for the summer months.

That looks pretty compelling. Once upon a time, fears over trading costs and the sheer hassle involved might have put you off acting on it, but in these days of index funds and online brokers, there’s far less friction involved in switching. So should you be emailing your broker (assuming you haven’t already sold out)? Well, no. New research shows that the “sell in May” effect “can be traced to only a few years”, notes Mark Hulbert on MarketWatch – “specifically, the third year of the presidential four-year term”.

The finding comes from research by Kam Fong Chan of the University of Queensland, and Terry Marsh, finance professor at the University of California, Berkeley, who used data going all the way back to 1897. They found that in the third year of a presidential term, the stockmarket shot up by more than 10% during the winter months, and declined a little in the summer. In years one, two and four, there was no statistically significant difference (3.3% for winter versus 2.3% for summer). We’re still only in year one of Donald Trump’s reign.

It’s worth noting that this also applies to non-US markets. For example, since 1966, the FTSE All-Share index has declined by an average of 0.6% between May Day and Halloween. But the researchers found that this is also driven by gains and losses made in the presidential third year – and this held true for markets from Canada, to France, to Japan, to the UK. None of this means that markets are “safe” for the next six months – they could still decline, or even crash. But the point is that the “sell in May” indicator offers no historical justification for bailing out now.

What drives the “sell in May” effect?

The saying “sell in May, go away, and come back again on St Leger Day” is thought to have its roots in the days when stockbrokers would up sticks from the City for “the season” – basically a round of partying that would end with the St Leger horse race in early to mid-September. In other words, markets were weak because all the big players were on holiday. Other theories include the idea that investors suffer from seasonal depression in the darker months; or are affected by systematic overoptimism in the spring; or that it’s all down to fluctuations in money flows caused by the timing of the tax year.

However, the new research outlined above dovetails beautifully with another well-known cyclical phenomenon – the “presidential cycle”. GMO’s Jeremy Grantham (see above) is credited with popularising the cycle, observing that markets tend to do better in the third year of a president’s term than in the other three combined. Using data going back to 1932, Grantham showed that US stocks gained an average of 0.2% a month in the first, second and fourth years of a presidential term; in year three, monthly returns averaged from 0.75% to 2.5%.

Grantham argued that this was driven by the Federal Reserve always deciding (“completely innocently”) to help out the party in power. The US central bank would stimulate the economy with interest-rate cuts in the third year of the term, boosting the economy ahead of election year, thus helping the incumbent’s efforts to retain power. Grantham now believes the cycle is dead – today, the Fed constantly stimulates the economy, which means the actions of politicians matter far less than they did. If he’s right, then it may not be long before “sell in May” dies out too – regardless of when St Leger Day falls.