Seasonal sectors

My table shows one answer, based on price changes since 1987. Three things stand out here.

.

First, almost all sectors are seasonal in the sense that they do better in the winter months (November to April inclusive) than in the summer. The only exception to this - and this is only slight - is the utilities sector.

Secondly, the least seasonal sectors are defensive ones: utilities, tobacco, food retailers and (these days) telecoms and pharmaceuticals. This is partly what you'd expect; if the market does badly in summer, low-beta sectors should outperform then.

But there's a quirk here. Sectors' betas are low but not negative, which implies that they should fall less than the market in bad times. However, our least seasonal sectors have risen on average in the summer while the market has fallen on average.

Thirdly, the most seasonal sectors are not really high-beta ones. IT stocks are only slightly more seasonal than the market, and banks are actually less seasonal. Instead, it is cyclical sectors that are the most seasonal.

One reason for this could be that economic activity is itself seasonal. GDP falls sharply in the first half of the year, and investors need a high risk premium on cyclicals in winter to compensate for this risk.

However, this cannot be the whole story. This cannot explain why cyclical sectors tend to fall on average in the summer. Something else must be happening.

This something could simply be that investors become over-optimistic in the spring, and this raises the prices of cyclicals to levels from which they are vulnerable to disappointment about economic growth.

The key word here is 'vulnerable'. The fact that many sectors fall on average in the summer disguises the fact that, more often than not, the market does actually rise then; it's done so in 15 of the last 26 summers. The problem is that, when the market does fall, it occasionally falls a lot and these few occasions have dragged down average returns. For example, the All-Share index fell by almost 30 per cent between May Day and Halloween in 2008 and by over 22 per cent in the summer of 2002.

It would, of course, be silly to dismiss these as outliers. To do so would be to commit the same error as the actor who asked: "Apart from that Mrs Lincoln, what did you think of the play?"

They do, however, tell us something about the sort of investor who should worry most about seasonality. These are loss-averse types - the sort for whom a (say) 5 per cent chance of a 20 per cent loss is more worrying than a 10 per cent chance of a 10 per cent loss; the two have equal expected monetary value. If you are this type of investor - the sort that can tolerate smallish falls but not big ones - then for you it is especially risky to hold cyclical stocks over the summer months.