Join our community of smart investors

Being right (and wrong)

What I got right in 2018, and what I got wrong (and why it matters)
December 18, 2018

Investing, like all our dealings with the real world, should be a learning experience: we must ask what we got wrong, what right, and what we can learn. The end of the year is as good a time as any for a round-up of these. So what did I get right and wrong this year?

The most important thing I got right was the direction of the market. In November last year I wrote that a few leading indicators “point to the market falling over the next 12 months, by around 4 per cent”. In fact, the All-Share index fell 6 per cent. I also warned that low ratios of the global money stock to global share prices, and big foreign buying of US equities, were bearish signals. This was correct.

This is a good win. Most forecasters (and background probabilities) said the market would rise this year. That I didn’t do so suggests I got something right.

Another win was market timing. The 'buy on Halloween, sell on May Day' rule worked well for much of this year. The All-Share index gave a total return of 2.2. per cent from Halloween 2017 to this May Day, and lost 3.5 per cent from May Day to Halloween. I shouldn’t rejoice too much, though: shares have fallen further since Halloween although it is of course too soon to say that this invalidates our rule.

Another good call was on China. In late 2017 I said that slower Chinese monetary growth “warns us of possible double losses – on miners and emerging markets” although “there might be a little more juice to be squeezed out of them”. This worked: mining stocks rose from then until June, but have fallen more than 10 per cent since then. And these losses have been accompanied by falls in emerging markets, confirming my view that the two are, for practical purposes, very similar assets. With China’s monetary growth having slowed further this year, this warning still holds.

So, what did I get wrong?

I’m not happy with my support for the 10-month rule – that we should hold shares if their price is above their 10-month moving average and sell when they are below them. This rule worked okay for mining stocks this year, but not so much for the All-Share index. Applying it at the end of each month would have got us out of the market in February and back in at the end of April when prices were higher. And it would have kept us in the market until late October, losing us 3.5 per cent.

This, though, only reminds us what we knew – that the rule works badly if the market is prone to small dips and rallies. Its virtue is that it protects us from the occasional protracted bear market. It is of course too soon to say whether, in keeping us out of the market now, this virtue will come into play.

There’s another failure. Back in January I wrote that 2018 should be a good year for stockpickers.

Wrong. In the last 12 months, only 62 of 253 unit trusts in the all companies sector have beaten Scottish Widows All-Share tracker fund, and several of these are FTSE 100 trackers. Most stockpickers have underperformed tracker funds

I thought smaller stocks would beat the FTSE 100, meaning that most shares would beat the index (which, remember, is weighted by market capitalisation), which would in turn have meant that most stockpickers would have done so. The logic was good. But the premise was wrong. In fact, the FTSE 100 outperformed the FTSE Small Cap index this year. With a few mega-caps such as AstraZeneca, GlaxoSmithKline and BP holding up okay, most shares have underperformed the index, which means that most stockpickers have done so.

You might think all of us have an excuse here. Brexit negotiations have gone badly. That has weakened the pound, which has caused a few large companies with lots of overseas earnings to do relatively well.

For me, this is no excuse. Quite the opposite. It highlights a weakness in my approach.

All these calls have something in common. They do not rely upon my judgment: the world doesn’t need another middle-aged man with an opinion. Instead, I use simple statistics to unearth leading indicators. In doing this, I’m following the psychologists Paul Meehl and Robyn Dawes who have shown that simple algorithms can sometimes beat professional judgement. My ego is irrelevant: I am merely the bearer of an intellectual tradition.

This approach works well insofar as it helps us avoid the countless errors of judgement to which we are all prone. In this context, these errors include letting our opinions of the market be influenced by others and by the seasons.

But it has a flaw. It cannot account for new or one-off developments that disrupt past statistical relationships. And Brexit has been one of these, in the context of the FTSE 100’s performance relative to small-caps if not other contexts.

That said, I don’t know if I would have done better had I taken account of this possibility. I don’t know how I could have forecast the course of Brexit negotiations. And the fact that so many fund managers underperformed the market tells us that they didn’t do so either.

For me, therefore, a big lesson (or perhaps reminder) of 2018 has been that statistical relationships can be broken – perhaps temporarily, perhaps permanently – by events. On the other hand, though, the fact that some of these relationships remain successful suggests we should not entirely abandon this approach, either.