Join our community of smart investors

When good strategies fail

Both momentum and defensive investing have done badly so far this year. This might be just bad luck
When good strategies fail

A bad market can hurt even good strategies. This has been the message of the performance of our momentum portfolio so far this year. In the first quarter, it fell over 10 per cent, underperforming the FTSE 350.

This wasn’t because of especially big falls by any particular stock or sector. Six of the 20 constituents lost more than 20 per cent, including stocks as different as Sophos (SOPH), Phoenix Global (PGR) and Purplebricks (PURP).

This is consistent with a theory proposed by Victoria Dobrynskaya at the National Research University in Moscow. She says that momentum strategies tend to have high betas in falling markets and low betas in rising ones. This makes momentum investing dangerous for fund managers who worry about their performance relative to the market and other managers. To compensate for this danger, momentum strategies must offer a risk premium in more normal markets.

In recent years, they have done just this. Even after this year’s 10 per cent drop, our momentum portfolio (which comprises simply the 20 best-performing shares with a market capitalisation of over £500m in the previous 12 months) has doubled over the past five years. That means it has beaten all but five of the all companies unit trusts in Trustnet’s database.

This, however, is puzzling simply because the risk premium is too big. Over the past five years, our momentum portfolio has outperformed the FTSE 350 by eight percentage points: a 10.8 per cent annual average rise compared with a 2.8 per cent average rise on the market. I find it implausible that such huge outperformance is justified by risk alone; very few risk premia are that big.


Benchmark portfolio performance    
 in Q1Last 12mLast 3yLast 5yLast 10y
Negative momentum-10.8-3.00.0-34.8na
High beta-
Low risk-8.0-1.922.524.064.3
Mega caps-8.8-
FTSE 350-7.9-3.05.814.331.9
Price performance only: excludes dividends and dealing costs 

More likely, at least some of momentum’s good performance is due to a behavioural factor: momentum stocks tend to be underpriced perhaps because investors underreact to good news. For investors, though, this raises a danger. Maybe investors wised up to this mispricing last year and thus eliminated it. If so, we can expect lacklustre returns on momentum in the future. I know, I’ve worried about this for years and the danger hasn’t materialised yet. But it might do so – and in an efficient market, it should.

In one sense, however, momentum is still working. Our positive momentum strategy did outperform negative momentum in the first quarter (Q1), thanks in large part to the halving in value of two shares: AA (AA.) and Dignity (DTY). This continues a long tendency for past losers to do badly. The old saying, 'never try to catch a falling knife' is true. It’s dangerous to bet on battered shares recovering.

The same thing is true of value stocks. Our value portfolio – the 20 highest yielders – lost almost 10 per cent in Q1 thanks in part to a huge loss on Capita (CPI).

This reminds us that the phrase 'income stock' is misleading. Our value portfolio (and income stocks generally) comprised three different types in Q1. Some were troubled stocks where investors were hoping for a turnaround. In the case of Capita and Provident Financial (PFG), such hopes were mistaken, although they were valid for Centrica (CNA), which rose in the quarter. Another type is shares that carry cyclical risk. This sometimes pays off nicely, as owners of housebuilders will tell you. In Q1, however, it didn’t for our portfolio: Kier (KIE) and Galliford Try (GFRD), among others, lost. And a third type of income stock is the one that’s on a high yield because investors believe it has gone ex-growth. Our portfolio held GlaxoSmithKline (GSK) and Go-Ahead (GOG), and both actually rose in Q1.

These three types have little in common. Those of you who want to invest for income should be very wary indeed of the first type, cautious about the second, but more enthusiastic about the third.

You might think that in a falling market a low-risk equity portfolio would do relatively well. Ours, however, did not in Q1. It fell slightly more than the market, with losses on many holdings.

As with momentum stocks, this reminds us that market risk matters. If the general market falls a lot it will drag down pretty much any portfolio – in fact, all 248 funds in Trustnet’s all UK companies sector fell in the first quarter.

It also reminds us what the case for defensive stocks is. It is not that they outperform a falling market; sometimes they do, sometimes not simply because past betas aren’t necessarily a guide to future ones especially over short periods. 

Instead, the case is that defensive stocks carry the risk of underperforming a rising market. This makes them unattractive to those fund managers who worry about such underperformance, which means defensives are on average underpriced. This means they should do well over the long term. And they have. Over the past 10 years, our low-risk portfolio has risen twice as much as the FTSE 350.


Which brings me to a surprise. We’d expect that in a falling market high-beta stocks would do especially badly. But they haven’t. In fact, our high-beta portfolio actually beat the market thanks to rises in Melrose Industries (MRO), Smurfit Kappa (SKG) and Asos (ASC).

My Bayesian prior is that this is just luck. Two big pieces of evidence tell me this. One comes from economists at AQR Capital Management. They’ve shown that for years it has paid off to bet against beta not just in equity markets but in bonds and commodities, too. The other is the performance of my high-beta portfolio, which has been poor over the long run.

There’s a reason for this. Some investors cannot borrow as much as they’d like to. When they are bullish, therefore, they express this not by borrowing more as conventional financial theory says, but by buying assets that are in effect geared plays on the general market – that is, high-beta ones. This, however, means that high-beta shares are overpriced relative to others, which means they should on average underperform.

For me, this seems a plausible theory that fits most of the evidence. We would never expect it to fit all the evidence simply because share prices are so volatile. Such volatility means that there will be occasions when even bad ideas pay off and good ones don’t. My hunch is that this is what’s happened so far this year.

Of course, my hunch is as worthless as anybody else’s. The only way to test it is to continue to monitor my no-thought portfolios.

All portfolios are drawn from UK shares excluding investment trusts with a market capitalisation of over £500m.