When momentum doesn't work

In the second quarter, our no-thought momentum portfolio (which comprises the 20 large stocks with the highest returns in the year to March 2015) fell by 5.7 per cent - an even worse performance than the FTSE 350. It's not just the positive leg of momentum investing that failed. Our negative momentum portfolio - the 20 biggest losers in the year to March - actually rose by 7.1 per cent in the quarter. The 'sell losers' strategy thus failed as well as the 'buy winners'.

However, this only partly reverses what we saw in the first quarter, when positive momentum did well and negative momentum stocks fell. So far this year, both legs of momentum investing have worked. More importantly, both have done so in the long run. In the past three years, our positive momentum portfolio has risen by 52.4 per cent, easily beating the FTSE 350's 22.7 per cent rise, while the negative momentum portfolio has fallen by 22.1 per cent. This continues the long-term pattern we've seen around the world - for momentum strategies in shares (and for that matter in other assets too) to pay off well.

Which poses the question: why did momentum fail in the second quarter (Q2)? The nasty possibility is that, earlier this year, investors wised up to the virtues of momentum investing. And, in doing so, bought momentum shares and so bid up their prices to such a high level that subsequent returns have been poor.

Two things, though, make me doubt this. One is that we've seen several falls in momentum before, only for the strategy to bounce back: momentum fell in the middle of last year but recovered nicely. Secondly, momentum's long-run success might not be due solely to investors being irrational and underreacting to good or bad news. Victoria Dobrynskaya at the LSE suggests that momentum carries benchmark risk - the danger of underperforming the market in both very bad and very good times. This makes momentum unattractive to fund managers for whom relative performance matters, which implies that the strategy must do well in average times to compensate them for this risk.

Perhaps, therefore, momentum's fall is just one of those things: even good strategies fail sometimes.

However, not all stockpicking methods failed in the second quarter. Our no-thought value and low-risk portfolios both beat the market in Q2 - in both cases extending longer-run gains. This raises a question: could it be that the risks to momentum investing can be diversified away by holding defensive or value stocks?

If we look at absolute returns, the answer is no. If we take quarterly returns since the start of 2006 we see strong positive correlations between momentum and either value or defensives - of 0.67 and 0.6 respectively. This is because all baskets of stocks carry market risk and so tend to rise and fall as the general market rises or falls: even most defensive stocks have a positive beta.

However, if we look at returns relative to the FTSE 350, there is diversification potential. The correlation between defensives and momentum has been slightly negative, at minus 0.12, implying that defensives have a slightly better than 50-50 chance of beating the market if momentum underperforms. In this sense, Q2's pattern is typical.

There is, however, a positive correlation (of 0.34) between relative returns on value and momentum. This could tell us that momentum stocks carry some cyclical risk, tending to do badly when near-term growth expectations fall, which is when value stocks do badly.

Nevertheless, this hints at an encouraging possibility. Insofar as momentum and defensives have a low correlation, then we might be able to spread the risks surrounding both strategies by holding both. Doing so won’t of course remove market risk. But it could reduce behavioural risk - the danger that one of the strategies will fail, if only temporarily. The fact that both strategies have tended to do well in the long run increases the attractiveness of this strategy.

While defensive stocks have outperformed, however, their obverse - high-beta stocks - have underperformed. Granted, the gap was small in the second quarter. But over the past five years it has been significant. This is odd: in theory, high-beta shares should beat a rising market while defensives should underperform, but in fact the opposite has happened.

One theory as to why this might be is that some investors face borrowing constraints. To this extent, if they are optimistic about the market they cannot borrow and buy stocks generally but instead they seek leverage on the market by buying higher-beta shares. This causes such shares to be overpriced and so subsequently underperform. As researchers at AQR Capital Management have shown, strategies that 'bet against beta' - such as defensive shares - thus do well on average.

However, in recent years most stockpicking strategies should have beaten the market. This is because the biggest stocks have underperformed: an equal-weighted basket of the 20 biggest UK stocks fell 4.4 per cent in Q2, continuing years of underperformance. This isn't just because falling commodity prices have hurt miners: AstraZeneca and Glaxo have also done badly.

This has an important mathematical effect. Because indices are weighted by market capitalisation, poor performance by a few big stocks has meant that most shares have beaten the index. This means that if you have picked stocks with no skill at all you'll have beaten the index merely if you have average luck. This favourable wind has also helped our no-thought strategies.

However, I'm not sure this can continue. There's no long-run tendency for big stocks to underperform; if there were, there'd be a tendency for companies to become equally-sized - which hasn't happened and probably won't. There is, however, a tendency for relatively expensive shares to underperform - and small-caps are now expensive relative to big ones. I suspect, therefore, that mega-caps might do a little better over the next few years. If so, then it will be harder for stockpicking strategies to beat the market.

The latest portfolios

Momentum (the 20 biggest risers in the past 12 months): Greggs, Betfair, Dart, Onesavings Bank, Marshalls, Hutchison China Meditech. JD Sports, New Europe Property, Howden Joinery, Barratt Devs, Taylor Wimpey, Crest Nicholson, Redrow, Ted Baker, NMC Health, Workspace, Card Factory, Persimmon, Moneysupermarket, Rightmove.

Negative momentum (the 20 biggest fallers in the past 12 months): Anglo American, AO World, BHP Billiton, De La Rue, Drax, Genel Energy, Hunting, Lonmin, Ophir Energy, Oxford Instruments, Premier Oil, Pure Circle, Quindell, Serco, Soco, Telecom Plus, Tesco, Tullow, Vedanta, Weir.

Value (the 20 highest yielders): Amlin, Anglo American, Ashmore, Berkeley, BHP Billiton, BP, Electrocomponents, Esure, Glaxo, Infinis Energy, Ladbrokes, Phoenix Gp, Premier Farnell, Redefine, Soco, SSE, Standard Chartered, Tate & Lyle, Vedanta, Wm Morrison.

High beta (the 20 highest beta stocks): Antofagasta, Asos, Bodycote, Centamin, Hays, Henderson, Inchcape, Int Cons Airlines, IPF, John Wood, Jupiter, Keller, Lamprell, Lonmin, Northgate, Polymetal, Regus, Thomas Cook, Travis Perkins, Vedanta.

Low risk (the 20 lowest beta stocks): Betfair, Breedon Aggregates, Clarkson, Dechra, Dignity, Domino's Pizza, Entertainment One, Euromoney, First Group, Homeserve, Imperial Innovations, James Fisher, James Halstead, National Grid, Quindell, Synergy Health, Telecity, Telecom Plus, Vectura, Wm Morrison.

Mega caps (the 20 largest stocks): AstraZeneca, Barclays, BAT, BG, BHP Billiton, BP, BT, Diageo, Glaxo, Glencore, HSBC, Lloyds Banking, Prudential, RBS, Reckitt Benckiser, Rio Tinto, Royal Dutch, SABMiller, Unilever, Vodafone.

All portfolios are equal-weighted and drawn from UK shares excluding investment trusts with a market capitalisation of over £500m, taken from the IC’s stock screen.