Equity markets have started the new year on a weak footing, and “as goes January, so goes the year” – so they say. But like “sell in May and go away”, this seasonal motto is not always useful.
Residents of the United States, please read this important information before proceeding
Developed stocks have fallen in January (by around 3% in dollars). The stock market saying suggests January’s move is a good predictor of the market’s direction for the year as a whole. Should we worry? Tactically no, since we’re no more than neutral on developed stocks, and are waiting for a setback to add to positions, as noted above. But
strategically, maybe: we’re expecting a positive year. How accurate is the saying?
There is some evidence of seasonality in stock market returns – but it has varied over time
It needs some clarification at least. Stocks rise in most months (because profits grow in most months) and so a positive January wouldn’t tell you anything special. For the MSCI World index since 1970, there have only been 11 negative years (1 in 4); for the S&P500 since 1964 there have been 13 (roughly the same proportion).
Of the 15 instances of a negative January for the MSCI World index, 7 turned into a negative year (so the retrospective odds of a full year decline rise to around 1 in 2, and almost 2 in 3 negative years had a negative January in them). In the 19 instances of a negative S&P January, 9 turned into a negative year (again, almost I in 2, and 2 in 3, respectively). So if history is a guide (and as we are bound to point out, it isn’t always), the odds of a full year decline after a negative January are broadly even – a coin flip. The odds rise further if we look for negative Januarys being followed by below average years – but since the averages are double digit gains, this is not too troubling.
Selling in May has not been a good tactic – but November-April has been the stronger half-year
The most familiar seasonal investing advice is ‘Sell in May and go away’. It originated in the City of London: the full saying is “Sell in May and go away; come back on St. Leger’s Day”. The St. Leger Stakes is the oldest of England’s five horseracing classics and the last to be run, around mid-September. The advice is closely related to the academically researched Halloween effect, which suggests buying in October.
Clearly, selling in May didn’t work last year: the MSCI World index rose 10% during May to October. In fact, the advice hasn’t worked more often than it has: developed, US and UK markets have usually risen in the May-October period (roughly 2 times out of 3). But as the charts suggest, on average the November to April half year has been more positive than May-October – by roughly 7%, 6% and 11% for the MSCI World index from 1970, the S&P500 from 1964 and the FTSE All-Share from 1965 respectively.
More generally, Figures 1-3 show that there are no individual months with consistently positive or negative stock returns on average in each period shown, though December comes close (the exception there being the S&P in the 1964-9 period). There are some clear patterns in implied stock market volatility (Figure 4) – it is lowest in June, and
seems to spike in September and October – but the VIX index is relatively young.
There are few investment sayings in bond markets, but some apparent seasonality nonetheless. For Treasuries, bunds and gilts since 1980, August to November are the stronger months (Figure 5). There is also some (less surprising, perhaps) seasonality visible in commodity prices since 1951 (Figure 6). Agriculture and Energy have been weaker in September and October.
There could be good reasons for these various apparent seasonal effects – the timing of holidays, temperate weather and harvests, for example – but they are not supposed to exist in “efficient” markets. A more rigorous test of this however would require taking transactions and opportunity costs into account. For the time being, we simply note that January may not be telling us much about stocks in the rest of 2014.