Is 'sell in May and go away' still a good idea?

Marcus Padley analyses the popular - but dubious - idea that investors should sell in May and go away

It's May and we've all heard the expression "Sell in May and go away", which for me embodies not one but two of the worst elements of financial analysis. The first is that it blindly uses past performance as a predictor of the future and the second is that it is lazy.

There is nothing worse in my mind than bygone statistics masquerading as value-add analysis, especially when it concludes with a hollow prediction or, worse, a recommendation. Where there are numbers there are statistics but whether they will repeat significantly often to advance investment returns is rarely addressed and is conveniently overlooked.

So when a broker rings you up and says you should "sell in May", instead of saying "What a good idea", your more natural response should be to say, "Is that really the best you've got in terms of value-add analysis?" Because if that's what you're paying a brokerage fee for, you might go elsewhere for more bang for your buck or intellect for your commission.

But the odd thing about this sharemarket anomaly is that for some reason it is true not just in Australia but in global equity markets over long periods.

An academic study of 37 stockmarkets by Sven Bouman and Ben Jacobsen in 2001 concluded that in 36 out of the 37 countries, "November to April returns are large in most countries, whilst average returns in the period May-October are not significantly different from zero and are often even negative". They examined a number of possible explanations but declared "none of these appears to convincingly explain the puzzle".

"Sell in May and go away" was an adage developed when the English aristocracy were the only people who could afford to trade shares. The implication was that you should sell out of the stockmarket before the summer social season started on the May Day holiday (first Monday in May) and come back when it wound up in October.

In fact, Bouman and Jacobsen stated that the effect was "robust over time" and in England "the effect has been noticeable since 1694".

In Australia, if you take a chart of the seasonal performance of the All Ordinaries Index from January to December over the past 10 years, the market does rather remarkably peak pretty much on the May Day holiday - May 3 actually.

At least it did in 2006, 2008, 2010, 2011 and 2012 and while the market may have survived May 2007, 2009 and 2013, the effect was simply delayed until June, making May and June two of the worst months, with the run-up in March and April being two of the best.

December is also a good month and there is a bit of a dip from mid-September to mid-October.

And if we take this out a bit further, over the past 75 years the average return on the All Ords in the six months before May (November 1 to April 30) has averaged more than 4.5%, which is more than double the 2.2% average return in the six months after May (May 1 to October 31).

In other words, since 1936 your chances of making money in the market in the six months before May have been more than double your chances of making money in the six months following.

That's surely more than a statistical anomaly and, although Bouman and Jacobsen could find nothing to explain the puzzle, they aren't Australian.

The most obvious driver in our market is that half the money invested is in financials and Telstra. In recent years, the focus on income has boosted the big banks and Telstra as they ran up to their results and dividends in May and November; interest was lost thereafter. On top of that, we have franking, which drives the domestic buyers in March, April and October.

The international institutions that own 40% of our stockmarket enjoy the rise but are naturally inclined to sell into the dividend buying ahead of results, create the peak and come back when the dividends have been paid, the franking premium has evaporated and the enthusiasm has drained away.

But there I go sounding like one of those zero-value-add brokers looking for an easy commission by making the preposterous suggestion that you sell the banks and Telstra before the dividends. What an idiot. No wonder we're being replaced by clicks.

Get stories like this in our newsletters.