Sell in May and go away: Is it true?

Alex Kalina | iStock/360 | Getty Images

There's an old saying in the stock market: sell in May and go away. Now that the first of May is here should we sell?

A lot of studies have been done on this saying. A famous study published in the American Economic Review in 2002 found that, yes, this phenomenon does exist and that returns on stock markets in 36 out of 37 countries studied from 1970 to 1998 were higher in the November to April period than they were in the May to October period.

But a study published in Econ Journal Watch in 2004 attributed the higher return to a couple of extreme data points, in particular the October 1987 crash in world equity prices and the August 1998 collapse of Long Term Capital Markets. These people found no economically exploitable opportunity in the S&P 500 futures market based on that strategy. However, it did find that more major economic and political events seem to take place during May to October period.

Read MoreWhat does 'sell in May and go away' mean?

That didn't stop researchers, though. Yet another paper published in the Financial Analyst Journal in 2013 studied this effect in the period following that observed in the first paper and found that the phenomenon did indeed exist for 1998 to 2012.

"On average across markets and over time, stock returns are roughly 10 percentage points higher in November-April half-year periods than in May-October half-year periods," the paper concluded.

Of course, to say returns are lower in May-October does not mean that they are negative. They could be positive but just not as large. However, this paper found that returns did tend to be negative during the period they studied for many of the euro zone countries -- Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands and Portugal – as well as the U.S., U.K. and Japan and a few others. Yet in no countries were the returns on average negative for November-April.

What are the implications this year?

I don't have a strong view on equity markets from a fundamental point of view – I'm an FX strategist by profession – but I think it's worth noting the comment that the authors of the second paper made: "a preponderance of major economic and/or political events that negatively impacted world equity prices have occurred during the May-October periods."

Read More 'Sell in May': Market bull and bear duke it out

Given the number of critical issues bubbling away right now, such as the crisis in Ukraine, the credit crunch in China, European elections this month, and the gradual end of quantitative easing in the U.S., I think caution is advised. Stock market investors might want to buy some out-of-the-money puts as an insurance policy, or hedge their portfolios in the futures or contract-for-difference market.

For the FX market, looking at quarterly correlations between stock markets and FX markets since 1998, we find the commodity currencies – Canadian dollar, New Zealand dollar, and Australian dollar – are strongly correlated with stock markets, particularly the S&P 500. These currencies (particularly CAD and AUD) tend to strengthen when stock markets strengthen and weaken when stock markets weaken. (Note that since CAD is quoted USD/CAD and AUD is quoted AUD/USD, the correlations have opposite signs, but nonetheless the implication for how the non-dollar currency moves in relation to stocks is the same.)

The British pound and Swedish krona have somewhat weaker but still significant correlations. Meanwhile, the Japanese yen is unique in that it is negatively correlated with the Japanese stock market but is not strongly correlated with U.S. stocks. The euro and Swiss franc show only weak correlation.

Read MoreWhy strategists think 'sell in May' is ridiculous

Thus, if stock markets do perform badly during the summer, the Australian and Canadian dollars would probably be the main losers, along with the New Zealand dollar, while the Japanese yen could strengthen. The euro and British pound might be relatively little affected, unless of course Ukraine were the trigger.

The paper that found the "sell in May" phenomenon held from 1998 to 2012 also found a similar phenomenon in the FX market with regards to carry trades, namely that carry trades tend to perform better during November-April than they do May-October. The return for both periods was positive on average, just lower in one period than in the other.

Nonetheless, carry trades are extremely sensitive to risk aversion, which is why they are often referred to as "picking up nickels in front of a steam roller." Given the warning above about unusual events tending to occur during this period, we should be careful about emerging market carry trades in relatively illiquid currencies and exercise strong risk control. The euro, Swiss franc, or British pound would probably be safer funding currencies than the Japanese yen.

The author is the Global Head of FX Strategy at IronFX Global, an on-line trading firm specializing in Forex, CFDs on U.S. and U.K. stocks, and commodities. He was previously Head of the Forex Committee at Deutsche Bank Private Wealth Management.