The All-Share index has fallen below its 10-month moving average, posing the question: should we now sell equities?
The case for doing so was first made by Mebane Faber of Cambria Investment Management in 2006. He showed that a simple rule of buying US shares when the S&P was above its 10-month average and selling when it was below it would have given investors higher returns and less volatility than simply holding shares since 1901.
My table shows how the same rule would have performed in the UK recently. I’m assuming that one applied the rule at the end of each month.
| Performance of 10-month rule | ||||
| Since 1985 | Since Dec 1999 | |||
| 10M rule | Buy & hold | 10M rule | Buy & hold | |
| Annual return | 9.2 | 9.5 | 7.4 | 4.5 |
| Std deviation | 11.2 | 15.2 | 8.8 | 13.8 |
| Sharpe ratio | 0.82 | 0.63 | 0.84 | 0.33 |
| Worst 12 months | -22.6 | -34.4 | -8.1 | -34.4 |
| Based on total returns on All-Share index | ||||
Since 1985, the rule would have given slightly lower returns than a buy-and-hold strategy but with lower volatility, thus giving us a better Sharpe ratio.
Since 2000, however, the rule has done far better than buy-and-hold, delivering higher returns and less volatility.
The difference between the two periods tells us when the rule works and when it doesn’t.
Imagine the market were to follow a saw-tooth pattern, falling (say) 5 per cent one month and rising (say) 6 per cent the next. In such a market, the 10-month rule would be a disaster. It would get us out before those 6 per cent rises and in before the falls. We’d be wiped out.
When 'buy on dips' works, the 10-month rule fails.
The rule works well, however, when falls lead to further falls.
This explains the pattern in my table. In the 1980s and 1990s, it paid to buy on dips so the 10-month rule failed.
In the 2000s, however, the rule worked wonderfully well. It got us out of equities for most of the 2000-03 tech crash, and got us out from November 2007 to May 2009, thus avoiding most of the financial crisis.
Whether we should apply the rule now depends therefore upon whether this is a dip or the start of a protracted bear market.
My hunch is that it is a dip. For one thing, UK equity valuations are reasonable: the dividend yield on the All-Share index is actually above its long-term average. And prospects for global growth are still good, which should encourage demand for risky assets: for the last 20 years, there’s been a strong correlation between annual equity returns and global industrial production growth. And fears of rising US interest rates are partly self-refuting: if stock markets look jittery, the Fed won’t raise rates much.
Hunches, however, are a pathetically weak basis of an investment strategy.
There’s a big question here. Is the fact that the 10-month rule has worked well in the 2000s simply due to the bad luck of us having had two long bear markets? Or is it because stock markets have become more momentum-driven – because there are fewer value investors now and more trend-followers? If it’s the former, we can ignore the 10-month rule. If it’s the latter, we cannot. Personally, I don’t think we can be wholly confident either way (or at least, not rationally confident).
We can, though, say something for sure. The 10-month rule does protect us from the risk of large protracted falls: the biggest annual loss under the rule is much smaller than that under a buy-and-hold strategy. Investors who are especially loss-averse, therefore – that is, those willing to sacrifice returns to avoid large losses – might want to consider trimming their equity exposure in light of the message sent by the 10-month rule.
