While efficient-market theorists will cringe at the thought, the old market adage “sell in May and go away” has not gone away, and the calendar summons all financial journalists to take a fresh look at the latest thinking on this issue.
The idea that stocks outperform and underperform in seasonally predicable ways makes no sense to efficient-market types, since rational profit-seeking investors would exploit this effect and arbitrage away the advantage.
Nevertheless, as May kicks in, the financial blogosphere is charged with updated analysis mainly reaffirming the phenomenon.
In a research note from Wesley Gray, who blogs at Turnkey Analyst, the portfolio manager and Drexel University finance professor cites research showing that the “Sell in May” effect outperforms in 36 of 37 developed and emerging markets.
An updated study he cites showed the effect held — with winter-season stocks beating summer-season stocks by an average 6.25% — even under an expanded time horizon of 319 years across 108 markets.
“This effect is prevailing and statistically significant everywhere and all the time!” Gray writes, noting it has actually strengthened in recent years.
Gray’s money management team added their own tests of the strategy using 13 different asset indexes — ranging from equal-weighted S&P 500 to MSCI emerging markets to a 10-year bond index — over an 87-year period.
Gray compared selling in May (and buying back in November) to the opposite strategy (buying in May and selling in November) and both to a buy-and-hold strategy over an 87-year time horizon.
“The results show that November-April period returns substantially outperform the May-October period returns, with an average difference of 8.22% in equity markets, and 3.42% in alternative markets,” he writes. “Only the bond market shows insignificantly underperformance of [the] ‘sell in May strategy’ (-0.73%).”
In dollar terms, the buy in May and sell in November (i.e., the strategy that is opposite the stock market folk wisdom), performed worst. Over a 34-year period, from 1980 to the present, $1 million invested in the Russell 2000 would have increased to nearly $3 million, far less than the buy-and-hold portfolio, which would total more than $33 million. Yet the sell-in-May portfolio would have more than doubled that, reaching nearly $79 million.
Digging deeper, Gray next compares Sharpe ratios for various asset indexes and finds that, except for the bond portfolio, the Sharpe ratios are “much greater” for stocks between November and April — indicating a higher risk-adjusted return in the winter months.
Further, about 80% of the negative Sharpe ratios (which investors would want to avoid) occur during the summer period.
While finding that the effect exists, Gray offers no explanation of his own as to its cause, citing instead diverse opinions ranging from changing risk aversion and liquidity constraints during summer vacation to mood changes caused by seasonal affective disorder.
Indeed, the debate rages on. Writing on his blog, Lance Roberts of STA Wealth Management points out that the force of the sell-in-May strategy is its success over time as opposed to its success every time.
“As with all Wall Street axioms, they are viewed by the media to be ‘valid’ only if they work every single year,” he writes. “The reality is that no axiom, investment discipline or strategy works all the time. It is the cumulative effect over long periods of time which defines success or failure.”
In his blog post, Roberts cites different opinions as to whether to follow the sell-in-May adage. Two reasons he cites for adherence this particular May include the Dow’s closing at a 52-week high on the last day of April and a study showing sell-in-May takes on greater force in midterm election years.
A third financial blogger, Jeff Miller, similarly emphasizes that not every May is an auspicious time to follow Wall Street’s folk wisdom and indeed his view of today’s market suggests that current economic trends’ bullishness outweigh what he deems a small seasonal effect.
“The Great Recession has been followed by a slow and plodding recovery,” Miller writes. “We have an extended business cycle with plenty of central bank support. Since I am expecting the current cycle to feature (eventually) a period of robust growth, I do not want to miss it. The 1% seasonal effect will be minor in a month where we get a real economic surge.”