Cormac Gleeson and Paul ODwyer ask whether managers should incorporate the well-known trading adage into their portfolio strategies
It was American science fiction author Philip K Dick who once said: "Reality is that which, when you stop believing in it, doesn't go away."
It has been about 15 years since Jacobsen and Bouman published their work on the so-called 'sell in May' or 'Halloween effect', predicated on the old stock market adage 'Sell in May and go away, but buy back by St Leger Day'. The pair concluded that in 35 out of 37 markets tested, risk could be halved, without affecting returns, by selling in May and buying again six months later. These results were discussed by Shane Whelan in an article for The Actuary (Beating the markets, 2002). Whelan, alongside Brian Lucey, further tested the hypothesis on new data, reporting similar success (Figure 1). The strategy was revaluated in a 2012 publication in the Financial Analysts Journal by University of Miami professors Andrade, Chhaochharia, and Fuerst, who updated the data and found the results held for the previously untested period 1998-2012.
Is this pithy rhyme a reality?
Using MSCI Investable Market Indices (large, medium and small capital) for each country since October 2001, we have broken down cumulative returns into two half-year periods (as is consistent with the literature), May-October and November-April, and the average was taken for each of the 23 indices. All indices were taken in their local currency, with gross dividends reinvested.
Figure 2 shows the comparison between the average return in the two six-month periods for each country. Graphically, we can see that returns are concentrated in the November-April period, with an unweighted average of 7.46% across each country, compared to 1.58% in the May-October period. The concentration of returns in November-April holds for all countries tested except for Spain. We can see that the effect is also present in emerging markets, although less pronounced than in more developed markets such as the UK and the US.
One notable exception consistent in all countries examined was 2009, where cumulative returns from November 2008-April 2009 were significantly lower than the following consecutive six-month sell period. This could certainly be attributed to the global financial crisis following the collapse of Lehman Brothers, which shows that the sell-in-May effect cannot be considered in a vacuum without proper economic context.
Even without conducting the statistical analysis which Jacobsen and Bouman, Whelan and Lucey, or Andrade, Chhaochharia and Fuerst carried out in their investigations of the sell-in-May effect, we can see that this anomaly is still present in the market today. So why does it remain while most, if not all, of the other known calendar anomalies dissipate after they have been discovered (see Murphy's Law and Market Anomalies, Dimson and Marsh, 1999). Bouman and Jacobsen proposed that perhaps summer holidays played a role in the presence and persistence of the effect, however the effect is still visible in markets in the Southern Hemisphere.
While the litany of evidence hasn't gone completely unnoticed, its impact on the world of investment has been close to imperceptible. Figure 3 shows the slight uptick in mentions of 'sell in May' in Google's collection of published texts as a proportion of overall mentions of 'stock returns'. Why does this level of unpopularity persist?
The answer may lie in the nature of fund management; specifically the return-driven incentive structure that is part and parcel of the industry. The sell-in-May effect does not deliver abnormal returns, merely significantly reduces the riskiness of returns. While this may not be as attention-grabbing on an annual report, the actuarial profession possesses a unique understanding of the costs of risk-taking in the market. Surely this is a strategy that can aid responsible risk management. By not focusing on the reduction in risk, and instead looking at the potential opportunity cost (halving the returns on investment), are we ignoring a potential trading strategy to remove unrewarded risk?
In truth, we have little to no idea why the effect persists. Any narrative spun thus far has been pure conjecture. That said, the weight of evidence now available is irrefutable; the sell-in-May effect has existed for some time, and continues to exist. Were we testing a drug for licensing, it would have passed all stages with flying colours. At a time when outperforming passive funds has never been more challenging, investors would do well to incorporate this strategy into their portfolios.
Cormac Gleeson is a final year actuarial student at University College Dublin
Paul O'Dwyer is a final year actuarial student at University College Dublin
