It's that time of year when investors all tend to reflect on our relative performance to figure out whether we should throw up our hands and just invest in a (boring) tracker. And what a year it's been in the equity market! From the depths of despair in the early summer when the Euro crisis looked set to swallow us all... to the fiscal cliff relief relay of recent days. When all of that was said and done, in the year to January 3rd 2012*, the FTSE 100 gained 6.6% while the FTSE All Share was up 9.0%. As it's our mission to highlight what works when in the markets, let's take a look at how effectively our model portfolios of the Investing Masters have been performing versus these benchmarks.
Guru Models Keep On Slaying The Market....
To recap, since December 15th 2011, we've been tracking 60 long only strategies and 5 short strategies, spanning from the investing greats (the likes of Graham) to classic academic finance papers (such as Piotroski). You can read more about how we run and rebalance the strategies here.
Of the long only strategies, an astonishing 90% (54/60) have beaten the FTSE 100 with gains of between 9% and 76%. Versus the FTSE All Share, 82% of the Strategies (49/60) are beating the Index. The average return of these long models was 22.4%.
On the short side, all but one of the short selling strategies are in negative territory (as they should be), and even that one, the Altman Z-Score screen, is well below the market return (having eked out just 0.22%). The worst (or rather best) performing short screen is down 24.4% in the last year, showing how important it is to avoid fundamentally weak, near bankrupt stocks with potentially dodgy earnings - no surprises there!
Which Strategies have done best?
As you can see clearly from the following chart, this year has been a story of Value and Income, which have outperformed other investing styles.
At the top of the tables it's been a story of value, dividends and contrarianism trumping growth and momentum. The 15 top performing screens have been:
|Screen Name||Style||% Return|
|Bill Miller Contrarian Value Screen||Value||76.7|
|David Dreman High Dividend Screen||Value||46.0|
|Charles Kirkpatrick Value Screen||Value||45.9|
|John Templeton Screen||Value||41.3|
|Benjamin Graham Net Nets Screen||Bargain||38.2|
|Piotroski F-Score Price to Book Value Screen||Value||37.8|
|Dreman Low Price to Cash Flow Screen||Value||37.8|
|David Dreman Low PE Screen||Value||37.6|
|Benjamin Graham NCAV Bargain Stock Screen||Bargain||34.8|
|Richard Driehaus Screen||Momentum||34.6|
|Negative Enterprise Value Screen||Bargain||34.3|
|Winning Growth amp; Income||Income||32.4|
|Naked Trader-esque Screen||Growth||31.7|
|Free Cash Flow Cows Screen||Bargain||30.5|
|Warren Buffett - Hagstrom Screen||Quality||29.0|
Top of the table is Stockopedia's model of the investment style of Bill Miller, former manager of the Legg Mason Value Trust. It highlights unloved stocks with strong cash earnings and has generated an incredible one year return of almost 77.0%. This is a very concentrated screen, though, so it will be interesting to see whether this performance can be sustained.
In second place is a High Dividend model based on the writings of famous contrarian investor, David Dreman, of Dreman Value Management. Focusing on out-of-favour dividend-paying stocks, it has returned almost 46.0% to investors over the past year. Impressively, all three of Stockopedia's Dreman models are in the Top 15 models table over one year.
Just behind that is the combined Value and Momentum Screen of Charles Kirkpatrick, up almost 46%. It combines quantitative filters for relative price strength and relative reported earnings growth, with a value criterion - using relative price-to-sales percentiles. Again, it's very concentrated.
Ben Graham's Net Net and NCAVs screen have also done very well, up 38.6% and 34.8% respectively. So has The F-Score Price to Book Value model of Josef Piotroski, associate professor of accounting at the Stanford University Graduate School of Business. This is in sixth position, returning 37.8% over the past year. The F-Score aims to identify deep bargain-bucket stocks that are in recovery.
Did we expect so many models to outperform?
Not really. Some of the models we're tracking are ones that we favoured, others were included for the sake of completeness. As a result, we kind of expected the distribution of returns to be a bell-curve around the market mean, albeit one shifted to the right, i.e. some screens would be doing worse and others (hopefully more) doing better than the average. But, in fact, very few screens have underperfomed the market.
So what's going on here? Well, there are a number of important caveats - the biggest one is that it's just a year's performance of course but, at least, what we're seeing here is based on live portfolio tracking (not just wishful backtesting). Secondly, part of this outperformance is explained by a small-cap effect. The FTSE Smallcap Index is up 26% this year and, because we have such a comprehensive stock database, many screens do pick up the more ignored small caps. Notwithstanding the impact of that, the performance is still just astonishing!
Get Your Quant On
We keep stressing how important it is to use quantitative investing tools in your investing process. We wrote a post recently about Nate Silver's Moneyballing of the American election and how simple algorithms can be better decision-makers than instinctual human judgement.
This is especially likely to be true when emotions like fear and greed get involved. Instead of listening to brokers, tipsters and rumours and being swayed by the Siren song of the stories they sell, we would argue that using a dispassionate common sense stock screening process (whichever one that may be) can help improve the gullible, emotional and easily influenced human decision maker at most of our cores.
So see you this time next year - when we may well be eating a vast amount of humble pie on the back of some nasty mean-reversion! But, for now, we're going to enjoy this moment in the sun.
*For this comparison, we've taken the year to January 3rd - this is not because it biases the results in any way but simply because some of us were nursing New Year hangovers and it took a while to run the numbers. Thinking about it, December 31st was a rather peculiar time because of the fiscal cliff uncertainty on that day so perhaps this (entirely arbitrary) choice of dates has some logic.