Momentum succeeds again

Our no-thought momentum portfolio - which comprises simply an equal-weighted basket of the 20 best-performing shares over the previous 12 months with a market capitalisation of over £500m - rose by 9.8 per cent in the first quarter, easily beating the FTSE 350. This partially reverses a mediocre performance last year and means that, over longish periods, momentum investing has paid off very well. In the last ten years, our momentum portfolio has more than tripled while the FTSE 350 index has risen less than 50 per cent.

This performance is especially spectacular when contrasted to its mirror image portfolio, negative momentum, the 20 worst performers over the previous 12 months. This portfolio lost 11.6 per cent in Q1, which continues a trend. If in the last three years you'd gone long of our positive momentum portfolio and short of the negative momentum portfolio, you'd have made over 80 per cent.

Such big gains are not a quirk of the recent UK market. Momentum investing has worked for years, on average, in US equities, in commodities and bonds, in both developed and emerging market indices, and even in the 19th century.

But why? Part of the explanation is that investors under-react. We are Bayesian conservatives: we stick too strongly to our prior beliefs in the face of information which undermines them. (This behaviour is, of course, not confined to investing as the general election campaign is demonstrating.) Once we've decided that a stock is a poor investment, we stick to that belief even though new developments suggest it isn't. This causes the share price to under-react to good news with the result that it gradually drifts upwards weeks after the news, as other investors gradually cotton onto its merits.

On the short leg of momentum, this process is exacerbated by a reluctance or inability to sell after bad news. This might be due to short sales constraints: some fund managers are unable to short sell stocks even if they reasonably believe them to be over-priced. Or it might be due to the disposition effect; investors cling onto losing stocks in the hope they'll break even, perhaps because selling such stocks means having to acknowledge our error in buying them - and we hate admitting we are wrong, even to ourselves.

All this poses the question: why haven't smart investors cottoned onto the enormous evidence that momentum investing works? Had they done so, they would have bought momentum stocks, thus pushing up their prices to levels from which subsequent returns were only average. But this hasn't happened. There's a good reason why. It's because momentum investing is risky, in three different ways:

■ Benchmark risk. Victoria Dobrynskaya at the London School of Economics has shown that momentum portfolios have the wrong sort of beta: they tend to underperform both sharply falling and sharply rising markets, but do well in more normal times. Our positive momentum portfolio, for example, underperformed during the worst of the 2008 financial crisis. This makes them risky for professional investors who are judged on relative performance.

■ Growth risk. A common reason for a stock to do well is that its growth options have improved. But it is risky for a firm to exploit such options. Doing so often requires it to invest more. And there's a lot that can go wrong with such expansion projects.

■ Behavioural risk. Despite the above, there is always the danger that investors have indeed wised up to momentum, which makes a big bet upon it risky.

Risks such as these mean that investors haven't bid away the momentum anomaly. And perhaps they won't.

However, momentum isn't the only strategy that's had a good start to the year. Value investing has also done well. Our value portfolio, which comprises the 20 highest dividend yields, has also beaten the FTSE 350.

I suspect there's a simple reason for this. Many value stocks are cyclical. The improving economy in the past few months has therefore caused them to do well. The longer-term pattern is consistent with this. Value has done well since 2010 as the economy has recovered, but it did atrociously during the recession of 2008-09: our value portfolio back then contained mortgage lenders such as Bradford and Bingley and some housebuilders, which did catastrophically badly.

There has, though, been a safer source of out-performance this year. Our defensive portfolio (the 20 lowest beta stocks) has also slightly beaten the FTSE 350. And this is part of a pattern. Defensives have beaten the index over the past five years, even though in theory they should have underperformed a rising market. That they have not done so is consistent with the idea that investors under-rate the virtues of dull stocks, causing them to be under-priced. And again, this anomaly is not bid away by the smart money because of benchmark risk; defensives should underperform a sharply rising market, which makes them risky for fund managers worried about underperforming their peers.

This does not, however, mean that defensives always outperform. There have been periods when they haven't. One reason for this is simply noise: even defensive shares are volatile and so will sometimes do badly through dumb bad luck.

Another reason, though, has been pointed out by MIT's Andrew Lo. Markets, he says, aren't so much efficient or inefficient as adaptive: they evolve. You can think of investment strategies chasing profits as being like species competing for food. If a food source becomes abundant, the species will multiply which will cause the food source to become depleted. That in turn will cause the species to decline, until the food source grows back. In this way, just as there are population cycles in nature, so investment strategies wax and wane.

Two of our strategies have waned recently. Megacaps and high beta have both underperformed.

In part, there's a common reason for this: the fall in commodity prices has hit mining stocks, which have featured in both portfolios.

Also, during the upturn megacaps have under-performed simply because they are less cyclical than other stocks: their poor performance since 2010 is the counterpart of good rises in the FTSE 250 and small-cap indices.

I suspect, though, that this is only temporary. If we look at the past 10 years, which includes the recession, megacaps have done about as well as the broader market. This is consistent with Gibrat's law, which says that growth is independent of starting size.

The longer-term poor performance of high beta stocks is more puzzling. Here's a theory. High beta shares are often "growth" stocks. However, because fears of secular stagnation have increased in recent years, hopes for good long-term growth have waned. This has hurt stocks offering such growth more than it has damaged cyclical or momentum ones. The upshot has been a de-rating of high beta. Perhaps, then, the poor performance of high beta is the counterpart of the good performance of bonds: both reflect stagnation risk.

Now, I'm not sure whether this is the case. But it suggests a reason why we should care about these no-thought portfolios. They don't just work as measures of which investment strategies work and don't, but also as a barometer of macroeconomic conditions generally. Just as value stocks are an indicator of cyclical conditions, so high beta stocks might be an indicator of sentiment about longer-term growth expectations.

The new portfolios

Momentum (the 20 best performers in the last 12 months): Assura, Betfair, Carnival, Cineworld, Dechra, Domino's Pizza, Evraz, Greggs, GW Pharma, Int Cons Airlines, Man, Marshalls, Moneysupermarket, New Europe Property, Optimal Payments, Shire, Songbird, Spirit Pub Co, Synergy Health, Workspace

Negative momentum (the 20 worst performers in the past 12 months): Anglo American, Brown (N), Bwin, Drax, First Gp, Foxtons, Genel Energy, Hunting, Lonmin, Ophir Energy, Petrofac, Premier Oil, Quindell, Serco, Soco, Supergroup, Telecom Plus, Tullow, Vedanta, Weir.

Value (the 20 highest yielders): Ashmore, Balfour Beatty, Berkeley, BP, Brit, Carillion, Catlin, De La Rue, Esure, Infinis, J.Sainsbury, Ladbrokes, Morrison, Phoenix, Premier Farnell, Redefine, Renewables Infrastructure, Serco, SSE, Vedanta

High beta (the 20 highest beta stocks, based on five years of monthly returns): Ashtead, Barclays, Bodycote, BP, Hays, Henderson, Inchcape, International Consolidated Airlines Airlines, John Wood, Jupiter, Keller, Lonmin, Mondi, Northgate, Polymetal, RBS, SIG, St James Place, Travis Perkins, Vesuvius.

Low risk (the 20 lowest beta stocks, subject to no more than three from any one sector): Acacia Mining, AstraZeneca, Clarkson, De La Rue, Dechra, Diploma, Drax, Genus, Green Dragon, Greggs, IG Group, Imperial Innovation, James Fisher, Lancashire, National Grid, New Europe Property, Qinetiq, Randgold, Severn Trent, Tate & Lyle.

Megacaps (the 20 biggest stocks): Astrazeneca, Barclays, BAT, BHP Billiton, BP, BT, Diageo, Glaxo, Glencore, HSBC, Lloyds Banking, National Grid, Prudential, RBS, Reckitt Benckiser, Rio Tinto, Royal Dutch Shell, SABMiller, Unilever, Vodafone.

All portfolios are drawn from the IC stockscreen for shares with a market capitalisation of over £500m.