Join our community of smart investors
OPINION

September's troubles

September's troubles
August 27, 2014
September's troubles

Since 1966 the All-Share index has on average lost 0.6 per cent in September, which means the month just beats June as the worst one for shares. This contradicts common sense, which says that risky assets should outperform safe ones on average.

It’s unlikely that such poor returns are due to mere chance. If September’s true returns and volatility had been the same as all months since 1966, we would have seen the market fall in 10 of the past 48 Septembers. In fact, it has fallen in 22 of them. And we should have seen just one September in which the market lost 10 per cent or more, but in fact we’ve seen four such occasions.

In recent years, September has been especially bad for smaller stocks. Since 1998, the FTSE small cap index has lost an average of 2.8 per cent in the month and the 250 has lost 2.4 per cent, whereas the 100 index has dropped only 1.6 per cent. These falls, though, tend to be reversed in November and December.

Monthly returns since 1998
All-ShareFTSE 100FTSE 250Small capsAimGiltsOutput
Jan-1.2-1.50.51.92.90.3-0.1
Feb1.61.43.01.72.50.3-0.3
March1.31.21.90.0-1.80.310.2
April2.32.22.53.90.70.1-8.8
May-0.4-0.60.00.3-0.60.40.9
June-1.2-1.2-0.9-1.1-1.3-0.10.9
July0.50.50.70.0-1.20.8-2.0
Aug0.30.20.91.80.11.5-4.1
Sept-1.8-1.6-2.4-2.8-1.90.67.2
Oct2.12.31.2-0.2-1.20.33.2
Nov0.90.91.40.81.91.00.8
Dec2.12.02.51.62.20.2-6.3
Source: Thomson Datastream

It’s difficult to blame this poor performance upon economic fluctuations. Yes, the economy is still strongly seasonal, but its seasonality is very different from the seasonal pattern in shares. Industrial production tends to rise a lot in September as we return from our holidays - headline figures obscure this fact because the ONS adjusts them for seasonal variations - but this is exactly when smaller stocks do badly.

A likelier explanation for September’s troubles lies with investors’ sentiment. As the nights draw in, we become depressed by the approach of winter, and this causes us to avoid equities. (The same process in reverse explains why April is so good a month for shares: the lighter nights improve our mood and willingness to take risks.)

Two things corroborate this theory. One is that economists at York University in Toronto have found that shares in September are especially sensitive to the weather, with the market being more likely to fall if it rains. This is consistent with bad weather hurting us by reminding us of the approach of winter, whereas it doesn’t have such a depressing effect in other months.

Secondly, although shares tend to do badly in September, gilts do well; since 1998 September has been one of the better months for the market, albeit not statistically significantly so. This is consistent with investors seeking safer assets in the month as appetite for risk falls.

There is, though, one piece of evidence that isn’t quite so confirming. If sentiment were responsible for the September effect, we’d expect Aim shares to do especially badly in the month, as these are (generally speaking) more sensitive to sentiment. But this isn’t entirely the case. Yes, September has been the worst month for Aim stocks, but not by very much; on average the sector has only slightly underperformed the FTSE 100 in the month.

The message here is not that we should sell stocks now and buy again in October; there are costs to doing so. It is instead that if the market does fall in the next few weeks, we should not read much into it. The pundits won’t admit to this, but such a drop might be due not to fundamental economic forces but instead simply due to the time of year. And insofar as history is any guide, such weakness might well prove to be a buying opportunity.