Stock Screen Review of 2015

LESSONS FROM THE MASTER

Mr Slater disagreed with how I had adapted the process he set out in The Zulu Principle for the "Small Cap Slater" screen I run in the magazine. Facing a dearth of results from the screen I decided to allow stocks to fail one of the quality tests, based on the principles laid out in Mr Slater's book, while insisting they had a very low price-to-earnings-growth (PEG) ratio (a very useful two-factor measure of value that Mr Slater helped popularise with UK investors). Mr Slater suggested I should be taking a different tack by loosening the valuation criteria, but "doubling down" on the quality criteria.

Some of the torrid performances we've seen from smaller companies this year give added credence to Mr Slater's critique. Indeed, one stock in particular, Globo (GBO), picked by my relaxed version of the Zulu screen epitomises the importance of keeping a close eye on quality. Globo is a company that was structured to show strong earnings growth, however, as it was achieving this by dubious means, it did not generate cash. And here's the rub, Globo qualified for my "relaxed" Zulu Principle screen because I allowed it to fail the cash conversion test.

DON'T EVEN GLOBO THERE

I would like to think my write-up of Globo which accompanied my Slater screen was sufficiently clear about the cash conundrum faced by any would-be investor. Chinese clothing company CamKids was another company where questions had to be raised about cash after it appeared in my small-cap special situation screen. CamKids later crashed and burned.

Mr Slater's remarks on the application of his investment approach touched on a challenge I've faced with several screens, which is that as share prices and valuations have risen over recent years, value-focused screens have found it increasingly hard to generate many results. The key question this raises is whether a screen can be relaxed while maintaining the essence of the investment process that has made it successful. A purist's response to this would simply be to stick by the criteria and note that no stocks fitted it. I personally am in Mr Slater's camp in thinking that one should at least try to adapt to meet the circumstances, although the way one adapts needs to be chosen wisely.

A broader conclusion to be drawn from the experience is how important it is to consider the question of cash generation when making an investment, especially in exciting young growth companies. This is not an easy analysis to carry out because, by their nature, most young high-growth companies are very cash hungry and investments are often made with the explicit understanding that cash will only really start to be thrown off later on when the business matures and its capital requirements reduce.

While in the case of the Zulu Principle screen, I feel Mr Slater offers a superior approach to the question of how to relax the screen's criteria than my own, some of my other Globo-loving screens (it was selected no fewer than five times in 2015) would have lost their raison d'être were they tightened up to catch such villains. It's the risk of Globo-style 'bum steers' that means it often makes sense to regard screens as a kicking-off point for further research rather than an end in themselves.

ALL COMES OUT IN THE WASH

Despite several of my screens facing 100 per cent losses on Globo and a 100 per cent loss for the special situations screen on CamKids (this has contributed to a truly terrible performance from the small-cap special situations screen since April) on the whole the screens are once again showing outperformance of the aggregate in the "year to date" (see table). The table is ordered by the level of outperformance that represents the return relative to that which could have been expected from the index (ie if the index falls 50 per cent a flat return from the screen would represent a 100 per cent outperformance, whereas were the index to rise 50 per cent while the screen rose 100 per cent, outperformance would be one-third).

Screen performances

Source: Thomson Datastream

My use of the term 'year to date' is rather ambiguous. As I run a stock screen most weeks, the year to date for the individual screens can range from almost a full year to as little as a week. On average, though, the so-called year to date covers about six months and the outperformance over that period stands at 4.3 per cent. This outperformance, though, comes against the backdrop of poor market conditions so the actual average total return from the screens is a negligible 0.6 per cent. Meanwhile, just over three-quarters of the screens are outperforming. Outperformance is measured against the index/indices the screen selects shares from and is based on total returns (share price movements plus dividends paid and reinvested).

I think it's worth noting that the screens currently having a particularly disappointing run are mainly those that have a tendency to focus on smaller companies to try to outperform by focusing on higher-risk situations. Many of these screens have been strong performers (such as the Piotroski screen and Greenblatt Magic Formula), and it is important to see a screen's success or failure in light of its long-term performance as all approaches to stockpicking can be expected to go through bad patches.

LONG-TERM THINKING

From the longer-term perspective, the aggregate performance of the screens continues to look impressive. Indeed, the average life of my screens currently stands at three-and-a-quarter years, the average index total return stands at 23.5 per cent and the average screen total return stands at 82.7 per cent. What's more, only one of the screens, Global Value, is underperforming since inception.

LONG-TERM INCOME SCREENS

Source: Thomson Datastream

The performance figures in the main table above include all my regular screens, but a number of thematic screens that I've run in the past and dropped when the theme has no longer seemed relevant are not included. The list above also includes two different versions of my Best of British and Contrarian Value screens as I focus on both tracking the top five picks from these screens as well as a more diversified portfolio. I've included my long-term income screens in a separate table as these screen are explicitly based on five-year holding periods unlike the other screens, which I assess on annual reshuffling - although, some look better when assessed on a long-term buy-and-hold basis.

Looking at the screens on an annualised basis provides a different way of drawing comparisons between the screens, although the successes and failings of the strategies can depend very much on what's going on in the wider market when they are run. The annualised figures listed in the table below are based only on complete 12-month periods for each screen - ie annualised performance data for a screen with a three-and-a-half-year history will only cover the first three years of the period.

ANNUALISED TABLE

Source: Thomson Datastream

CRYSTAL-BALL GAZING

My stock screens have an exciting year coming up because a number are turning five years old. As far as the investment industry is concerned, strategies with five-year records provide a track record from which some conclusions can be drawn. Personally, I'm somewhat sceptical about attaching significance to arbitrary time periods, nevertheless there are undeniable virtues in doing something and sticking with it for the long haul, not least because it provides a very good way of demonstrating what works, when it works and why it works.

There is one clear risk to the potential for the screens to outperform next year, and that comes from the resources sector. Due to the weightings used to construct indices, the performance of this sector has a considerable influence on the FTSE 100, 350 and All-Share. These indices are all important sources of stocks for the screens. The screens are monitored based on equal-weighted portfolios, which means they probably owe some of their historic outperformance to putting less weight on their resources picks. Therefore if the fortunes of the resources sector reverse, which could be both sudden and pronounced, the relative lack of resources exposure could leave the screens floundering.

For now, though, the screens continue to look good on the whole.