The FTSE 100 has fallen by over 200 points since its peak in May. Combined with the market’s strong rises in December and April, this raises the possibility that we might have been too quick to dismiss the possibility that stock markets still have a seasonal pattern.
This pattern is for shares to do well in the spring, fall in the summer and recover later in the year. Since 1997, for example, the All-Share index has risen by an average of more than 2 per cent in April and December (its best months) but has fallen on average in May, June, August and September.
What’s more, almost all FTSE sectors conform to this pattern. Every one of the 24 main sectors has risen on average in December and every one has done better then than in June and September, the two worst months for the market. Every sector has done better in April than in June and only tobacco has done worse in April than in September.
Seasonality, then, is widespread.
Herein, though, lies a puzzle. This fact cannot be entirely explained by the fact that some sectors are simply more sensitive than others to moves in the general market – that they have higher betas.
| Some monthly returns | ||||
| April | June | September | December | |
| All-Share | 2.0 | -1.5 | -1.5 | 2.2 |
| Mining | 2.6 | -0.4 | 0.1 | 3.0 |
| Construction | 1.9 | -1.6 | -1.9 | 3.6 |
| Engineering | 3.9 | -2.1 | -3.3 | 2.0 |
| Tobacco | 0.9 | 0.5 | 1.6 | 2.0 |
| General retailers | 2.6 | -3.1 | -1.8 | 0.9 |
| Banks | 4.1 | -2.9 | -1.7 | 0.8 |
| IT software | 0.2 | -2.3 | -2.2 | 3.8 |
| Based on price moves since 1997 | ||||
Yes, this is the case to some extent. Defensives, for example, have tended to hold up relatively well in the bad months of June and September, while high-beta sectors such as miners, banks and electronics have had good Aprils. But there are several exceptions to this pattern. IT stocks have dome relatively badly in March and April; miners have done relatively well in September; and banks have underperformed in November and December. These patterns are inconsistent with sectors’ seasonal patterns being driven solely by their co-movements with the All-Share index.
There might be a reason for this. In a recent paper David Hirshleifer at the University of California at Irvine and colleagues show that US shares have “mood betas” that are distinct from conventional market betas. Shares that tend to do best in January – traditionally the best month for the US market – tend to underperform in September and October, the worst months. What’s more, they show that those stocks with higher mood betas based on these monthly patterns also tend to do better on Fridays and worse on Mondays, when investors’ moods tend to improve and then worsen.
A similar thing seems true of the UK. For example, there are seven FTSE sectors that have outperformed the All-Share index by at least a percentage point in December. Six of these underperformed it in September; the exception is healthcare.
All this, says, Professor Hirshleifer, is “hard to reconcile” with the idea that shares’ returns are related to risk. Even in the best months, shares must surely be riskier than cash and so should on average outperform it. But, in fact, they underperform it in five months of the year. There are other anomalies. Why, for example, are IT stocks so much riskier at the end of the year than at other times which causes them to deliver great returns then? Why are retailers so risky at the end of July that they tend to do better than other stocks in August but not in any other month? Why are banks so risky at the end of June that they do well in July, but are relatively safe at the end of the year and so offer lower average returns then?
Such anomalies are inconsistent with the conventional idea of a risk-return trade-off. But they are consistent with the idea that investors' moods vary over the year.
Now, you might object here that I’m seeing patterns in what is really random data: it’s called apophenia. I’m not so sure. The difference in returns between April or December and June or September is so large and so persistent – it long pre-dates the last 20 years – that it is unlikely to be due to chance. Instead, we should at least accept the possibility that our moods are seasonal; we become more cheerful in the spring and more pessimistic in the autumn. This is hard for many to accept, even though it is consistent with the evidence.
