Last year, in 2011, the FTSE 100 index rose 3.9 per cent over the final five trading days of the year. By contrast, the previous year the market fell over the same period. So what, you may think. In the short term market moves are random - some years it will be up, some years down.
However, in the 15 years prior to 2010 the FTSE 100 rose every year over the final five trading days of the year. And since the formation of the FTSE 100 index in 1984, it has fallen in those final five days only four times. The average change in the index in this five-day period has been 1.4 per cent over those 28 years. Given that the average annual return for the index since 1984 has been 7 per cent, it is remarkable that a fifth of that return is attributable to just the final five days of the year.
But the FTSE 250 index is even more impressive. Since 1986 the index has fallen in only one year (1987) over this final five-day period.
Of course, there's no guarantee that the FTSE 250 will increase, say, in the final five days of this year, but history would suggest that there's a very good chance it will.
The end of the year is one of the strongest short periods for the market; and this is followed quickly in the New Year by one of the weakest periods. Since 1984, the FTSE 100 index has fallen in 18 years over the first eight trading days of the New Year, with an average return of minus 0.32 per cent. Which suggests a strategy of being long the market over the final five days of the year, then reversing the position to a short at the close of trade on the final trading day of the year, and closing out on the eighth trading day of the New Year.
Monthly seasonality
If we extend the period of study to a month, we can find further interesting seasonality effects. The strongest month for the market is December - by quite a wide margin. Since 1970, the market has positive returns in December in 86 per cent of years, increasing by an average of 2.6 per cent in the month.
Surprisingly, perhaps, the second strongest month in the year is October, a month often associated with large falls. At this point it is useful to refer to a statistical measure called standard deviation - this measures the spread of actual values around the calculated mean (the average). So, although the average market return in October is positive, the standard deviation is very high, which means that in any one year the actual return can be a long way from the average. In other words, the market has been very volatile in October.
The weakest months in the year are June and September.
It can be useful to also look at relative performance by month. Market sectors can display seasonality behaviour that results in periods of under- and over-performance of the general market. For example, the pharmaceuticals and biotechnology sector has historically been weak relative the FTSE 100 index from October to February, but relatively very strong in May and September. Another example we could look at would be the chemicals sector: since 2000 this has been weak for the three-month period August to October, but has outperformed the wider market for the rest of the year.
We can also study the relative monthly performance of international markets.
Although since 1984 the S&P 500 index has overall greatly outperformed the FTSE 100 index, there are months in the year when the FTSE 100 fairly consistently outperforms the S&P 500 (in sterling terms). The three months that are relatively strong for the FTSE 100 are: April, July, and December. The greatest monthly difference in performance is seen in May, when the S&P 500 on average beats the FTSE 100 by 1.4 percentage points each year.
This suggests a strategy of investing in the FTSE 100 in the months April, July and December and in the S&P 500 index for the rest of the year. The chart shows the result of operating such a strategy every year from 1984. For comparison, the chart also includes the portfolio returns from continuous investments in the FTSE 100 and S&P 500 (sterling-adjusted), rebased to 100.
The result: the FTSE 100 portfolio would have grown 564 per cent, the S&P 500 risen 751 per cent, but the FTSE 100/S&P 500 monthly switching portfolio would have increased 1,874 per cent. Switching six times a year would have incurred some commission costs, but these would not have dented performance significantly.
The six-month effect
If we now look at market behaviour over six-month periods, we will come to the most famous seasonality effect of them all – the six-month effect (also known as the 'Sell in May' or Halloween effect). This describes the tendency of the market over the period November to April (sometimes called the winter period) to greatly outperform the period May to October (summer period).
This rather extraordinary effect can be exploited by investing in the FTSE 100 during the winter period and then switching into sterling T-bills for the summer period. Such a strategy will outperform a standard FTSE 100 index fund invested all year round. The strategy can be further enhanced by using the MACD indicator to time the entry and exit point to the market.
The six-month effect does not exist just in the UK. A paper published this year found evidence for the effect in 81 out of 108 countries. Its strongest impact was in western developed countries in the past 50 years.
Momentum
Besides seasonality effects, there are many market anomalies that can be interesting to look at.
Research shows that the more periods the market rises, the more likely it is to continue rising. For example, if the market rises four months in a row, there is a 67 per cent probability that the market will continue to rise in the fifth month. In addition, such momentum effects are stronger for longer time frequencies - for example, momentum effects for months are stronger than those for weeks or days.
What might be called 'reverse momentum' can also exist. This is the basis of the Dogs of the Dow strategy. In the UK a simple strategy is to buy at the end of the year the 10 stocks in the FTSE 350 that have performed the worst and hold them for three months to the end of March. These stocks have a tendency to bounce back; and a portfolio of such stocks would have outperformed the FTSE 350 index in the first three months in every year since 2002.
How persistent are these effects?
An important question to ask is how persistent are these effects? It would be no good if they occurred just for a year or two then disappeared. And, strictly, none of these effects should last long as the actions of arbitrageurs should work to eliminate the anomalies. Some effects do indeed come and go. For example, the UK market used to be strong on Fridays and weak on Mondays; but that has no longer been true in recent years. Elsewhere, up to 2007 the month of January was the fourth strongest month in the year, but it has now slipped to eighth position.
But many effects also have astonishing staying power. Another academic paper (by the same authors as above) studied the persistency of seasonality effects in the UK market from 1693 (I think we can safely call that long-term).
They found that the months of July and October had the greatest persistency of under-performance; while April and December have been strong months for long periods. In contrast to the monthly performance, which has varied over time, the six-month effect has shown robustness over the whole 300-year period, and has been significantly strong since the 1950s. They also found that over the long term the stock market in the summer period underperformed the risk-free rate, and that a strategy based on the six-month effect beats the market over 90 per cent of the time for a 10-year time horizon.
How to exploit this analysis?
The instruments of choice to exploit these effects and anomalies would be futures, contracts for difference (CFDs) and spread betting, although some of the analysis may be helpful in timing stock buys and sells. With the more arbitrage-oriented strategies, the benchmark should be the risk-free rate, rather than an equity index. Leverage the strategies a bit, and you'll be running your own hedge fund.
Needless to say, this quantitative analysis is a long way from the buy-and-hold investing approach of a Warren Buffet. It might be regarded as a complement to, rather than a replacement for, more traditional methods of investing.
