When sell means buy

Flying in the face of the stock market's brightest minds may sound like a perverse - and risky - strategy. After all, equity analysts spend 12 hours a day researching the firms they follow, and get paid a fortune for doing so. You wouldn't intuitively expect to make money by doing the opposite of what the experts collectively recommend.

In practice, that's exactly what seems to happen, though. In 2001, three US academics - Brad Barber, Reuven Lehavey and Maureen McNicolls - studied the performance of US analysts' most and least favoured stocks in the prior year.

Their findings were striking: those shares with the most strongly positive recommendations subsequently fell by an average of 31.2 per cent, while those with the most strongly negative recommendations rose by an average of 48.7 per cent.

Inspired by this approach of buying hated shares, we've compiled a contrarian portfolio for the UK market over the past three years. And it's beaten the market every single time.

Our most recent selection - made in February 2006 - rose 30.1 per cent over the following 12 months. By contrast, the FTSE All-Share index gained just 12.9 per cent during that time. See the box for a discussion of this performance.

How It Works

The construction of our portfolio couldn't be simpler. We simply go through the FTSE 350 index and pick out the shares with the worst ratings from analysts.

In order to qualify for our 'hate squad,' a share must have a following of at least 10 analysts. This is to ensure a reasonable depth of opinion about a share.

At least two of the analysts must have outright sell recommendations on the share, and no more than a quarter of the group can have outright buy recommendations.

Our final selection is simply the 10 most unpopular shares in the market, based on these criteria.

Why It Works

While the rules for assembling the portfolio are clear and straightforward, the reasons for its success are not as obvious.

Overreaction is one possible explanation. When a company does badly and its shares tumble, analysts respond by issuing sell recommendations. Often, though, this is a case of slamming the stable-door after the horse has already bolted.

If all the analysts are screaming "sell," there's a good chance that a share will be forced down further than is actually justified. This is especially true when negative sentiment persists for a long time.

Eventually, investors wise up to the fact that the share is cheaper than it ought to be, even given the bad news about the company. The share then recovers strongly to make up for the overdone fall.

Although analysts probably realise there's a good chance of this happening, they don't generally like to run against the herd. It's much safer to do whatever everyone else is doing.

John Maynard Keynes famously said that "it is better to fail conventionally than to succeed unconventionally." Analysts fear being the lone voice and looking silly if they turn out to be wrong. As well as losing face, they might end up losing their job if they screw up like this too often.

As a result, analysts stick to cheering on popular shares that have gone up because they assume they will continue to do so. At the same time, they shun losers, because of the risk that they will continue to do so.

Momentum matters in professional investment. Fund managers - who are the main consumers of analyst research - are frequently judged on their performance over as little as three months. They often don't have time to sit around waiting for an oversold share to recover.

Another possible reason for our strategy's success is because it is risky. Shares that have fallen a lot offer higher expected returns to investors to compensate them for the big risks they are taking on.

As we'll show, however, not all the shares that analysts sell have already gone down. They do sometimes lay into those that have gone up a lot.

Source: Datastream, Reuters

This Year's Selection

In the accompanying table, we present the top-10 shares that the analyst community currently loves to hate.

Six of the 10 have underperformed the wider market significantly over the past year. The four that haven't show that analysts don't merely kick dogs when they're down. They are capable of negative feelings towards recent winners too.

We've added some extra information about these companies this year. As well as looking at those with the most sell recommendations, we also present data on overall negativity in recommendations. This includes not only outright sell advice but also 'underperform,' 'underweight,' 'reduce,' and various other milder negative recommendations.

Another new angle is subtracting total negative recommendations from total positive recommendations to give a 'net' recommendation figure. We also look at the mean score of recommendations provided by Reuters. This is a simple sliding scale that assigns each recommendation a numeric score. These extend from 'one' for a buy to 'five' for a sell. The result for each share is then divided by the number of recommendations to give an average score.

As a rule, we don't like to think too much about fundamental reasons why this selection might do well or badly. The portfolio is based purely on recommendations. If we started analysing the firms in detail, we'd probably lose our nerve and end up agreeing with the analysts.

Nevertheless, a quick glance down the list of names sheds some light on why these shares are so unloved. HMV has long been a disaster area, thanks to the rise of internet shopping. The smoking ban and the fading popularity of bingo have kept Rank in the headlines for all the wrong reasons. And Alliance & Leicester is a victim of the credit crunch and fears of a housing crash in Britain.

Just to be absolutely clear, we don't care about any of this. The fact that these 10 shares are the most hated ones around is all that we need to know.

Will It Work This Time?

Although our portfolio has generated good results every year we've run it, there's no guarantee that it will continue to do so. After all, three years isn't much of a sample period. Our performance could be due to nothing more than luck.

As in previous years, FTSE 250 shares are well represented in this year's selection. This has hitherto counted in favour of our portfolios. But it may well undermine it in the coming year. The tide has turned decisively against mid-caps in recent months, with large-caps doing much better.

Anticipation of an economic slowdown could be the reason behind this. Firms in the FTSE 250 tend to be more UK-focused and possibly more sensitive to the economic cycle than some of their peers in the FTSE 100. If the domestic economy slows sharply in 2008, this could hurt our portfolio.

In such an environment, we may not get any help from takeover activity. Matalan was bought late last year, while Somerfield and DFS were both prominent acquisitions in the previous years. It's not just the deals themselves we might miss, though. The anticipation of deals that never materialise also helps buoy up shares.

2006 Portfolio Assessed

The 'sell means buy' portfolio we assembled in February last year was our most successful one yet. The gain of 30.1 per cent was well ahead of the 20.4 and 16.5 per cent performances in the two prior years.

Often the success of contrarian portfolios is attributable to giant wins by a few individual shares, which more than offsets losses or average showings by the rest. Not so with our one, though. Eight out of 10 increased in value, and seven of those beat the wider market.

Only one of the companies here - Matalan - was actually taken over during the period. However, Hanson was subsequently bought out, showing that while analysts may not be able to see the value in a company, business rivals certainly can.

The presence of FTSE 250 firms in the list is worth noting. The period in question was a good time for mid-capitalisation firms, with the FTSE 250 index rising 23.5 per cent compared with just 11.9 per cent for the large-cap FTSE 100.

Ditch Their Darlings

If buying the shares that analysts all hate works so well, how about selling the ones that they all love? The academic research that inspired our portfolio suggests this may be worthwhile. And in previous years, we have enjoyed some success with it. For example, the 2005 portfolio of City favourites managed only a 6.1 per cent gain, which was less than what the wider market registered.

Selling the analysts' pet favourites wouldn't have worked so nicely last year, however. The shares with the most buy recommendations went up by 15.9 per cent, compared to the 12.9 per cent gain in the FTSE All-Share index. While that's nowhere near as good as the portfolio of sells, it's still a respectable enough performance. Seven out of 10 beat the market, which means the analysts achieved one of the main objectives of their fund-manager clients.

With this in mind, we turn to their favourite shares of today. Interestingly, six of the shares have underperformed the wider market over the past 12 months. This casts some doubt on the idea that analysts are simply 'glory hunters' who cheer loudly for those shares that have already done well.

FTSE 250 companies are even more strongly represented here than they are among analysts' top-10 sells. If the recent underperformance of the mid-cap index continues, this selection might struggle to fulfil the analysts' hopes.

The presence of three support services companies is interesting. This could indicate that the analysts covering that sector are a particularly excitable bunch. Once again, we're not really too bothered why the analysts like these companies. The fact that they do is enough to turn us off them.

The 'S' Word

'Sell' is still a dirty word in the stock market, although it seems to be more common that it once was. Especially during the technology bubble, big institutions with investment banking operations often used to lean on their analysts to prevent them from issuing sell recommendations on companies for which they were doing corporate finance work - or for which they hoped to do corporate finance work.

Following changes to the rules in 2004, analysts have had to declare whether their research meets the Financial Services Authority's (FSA) objectivity standards. According to FSA research carried out the following year, this did produce a shift in behaviour.

For brokerage firms that didn't have a business relationship with the companies they covered, negative recommendations issued went up by 2 per cent, while positive ones fell by the same amount.

For brokerage firms that did have a business relationship with the companies they covered, the percentage of outright sells remained static. However, slightly negative or neutral recommendations went up 8 per cent at the expense of positive recommendations.

Overall, negative recommendations remained much less common than positive or neutral ones. For brokerages with no business links to the firms they covered, negative calls amounted to 18 per cent of the total issued, while the figure was just 11 per cent for those with ties.

Interpreting individual analyst recommendations still takes a fair degree of judgement, more than it probably ought to require. As well as buy, hold, and sell, City scribblers employ a rich lexicon of synonyms - and euphemisms.

In the positive camp, 'add,' 'accumulate,' 'outperform,' 'overweight,' are among the most common ways of expressing appreciation. Besides 'hold,' there is 'equal weight,' 'market perform,' 'in line,' 'neutral.'

Still, no one has yet bettered the creativity of Terry Smith's legendary advice to shareholders of the Mirror Group: "Can't recommend a purchase." This refreshingly honest acronym enraged Robert Maxwell and cost Mr Smith his job.