Monday, October 29, 2012
Old Halloween investment adage still holds true
In Europe it’s a piece of market wisdom referred to as ‘Sell in May and go away’; in the United States they call it the
Halloween effect. No one is entirely sure why, but researchers at Massey University have shown the old adage holds true
– even when all of the world’s stock markets are considered.
The Halloween effect suggests that stock market returns during the Northern Hemisphere winter (November to April) are
significantly higher than during the summer (April to October). A Halloween strategy says investors should sell their
stocks in May, and return to the market after Halloween (October 31).
Massey University Professor Ben Jacobsen and Sven Bouman first tested for the phenomenon in 2002. Their resulting paper
found that the stock markets in 36 out of the 37 countries studied showed substantially higher returns during November
to April.
Professor Jacobsen and PhD candidate Cherry Zhang have now confirmed the effect’s existence by crunching the numbers for
stock markets in 108 countries over all the periods for which data is available. Their results show not only that the
Halloween effect continues to exist, but that it is also increasing in strength.
“We analysed 55,425 monthly observations over 319 years, and the average difference between November-April and
May-October returns is 6.25 per cent over the past 50 years,” says Ms Zhang. “A Halloween strategy beats the market more
than 80 per cent of the time over a five-year horizon, and this increases to 90 per cent if you expand the horizon to 10
years.”
The two researchers found that November-April returns were higher in 81 out of 108 countries, including in New Zealand,
but the size of the effect does vary geographically.
“The Halloween effect is strongest in developed Western European countries, but is also significant in North America and
Asia,” says Ms Zhang. “If you wanted to use a Halloween investment strategy, you would do best to focus on those
regions, entering the market around November each year, then selling up in May and putting your money into Treasury
bonds or the bank.”
While Ms Zhang says her PhD supervisor Professor Jacobsen has invested in this way for many years with success, she is
more conservative.
“We haven’t found the root of the anomaly, so all we have observed is past information,” she says. “You can’t say the
effect will definitely persist, even though the trend suggests it is increasing in strength. If we identified a really
good economic reason underlying the anomaly, then I would be confident to invest my money using the strategy.
“It’s also possible that, if everyone started to use this strategy, the pattern might change. But I do think, at the
moment, investors could do well by assuming the effect does exist.”
Finding the underlying cause of the Halloween effect is something that continues to preoccupy Ms Zhang. In another piece
of research for her PhD thesis she is analysing outbound travel data to see if there is a correlation between the
Halloween effect and the holiday period.
“The hypothesis is that lower summer returns are caused by investors taking vacations away from stock markets during the
summer months, which changes market risk aversion, liquidity demand and reduces the trading activities of the market.
“Interestingly, the peak season for New Zealand outbound travel is the winter months in the middle of the year, which
are the same months, but different season, to the European countries. While outbound travel data is a bit noisy as a
proxy for the holiday period, it could explain why the Halloween effect exists for New Zealand, even though it’s in the
Southern Hemisphere.”
‘The Halloween Effect: Everywhere and all the time’ by Professor Ben Jacobsen and Cherry Zhang can be downloaded from: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2154873