Quant Winter
According to this recent article in FT just 15% of quant funds beat the US stockmarket index last year. Cliff Asness, head of AQR, which employs over 80 phDs and manages $186bn, has called it a “quant winter”. However, the “quant winter” has now lasted a couple of years.
Old fashioned “active” portfolio managers have been losing assets to quant funds and index trackers over a decade now. Quant funds were supposed to be the one area of growth – in part because it was easier to market “back-tested strategies with factors” than a “star fund manager” who met companies and asked questions and made qualitative judgements.
Winter was entirely predictable
I think this was entirely predictable. And actually it was predicted by someone who used to run a quant fund: Doyne Farmer. Farmer’s background is physics, and Santa Fe Institute. At one point (like Claude Shannon) he built a computer to beat the roulette wheel in casinos. Like Shannon, he also got interested in investing. He didn’t go down the “buy and hold” route that Shannon did, but instead like Ed Thorp started a quant fund (I think in part because he got frustrated with economists telling him his ideas were wrong, and markets were efficient etc).
You can see his talk on YouTube here:
The gist is that the intellectual framework from ecology can be imported into finance. And that the market is an ecosystem with various players feeding off each other – often the players are optimised to different time scales. Thus High Frequency Traders (HFT) and market makers attempt to make many short term trades and generate profits. Trend followers and value investing strategies (which are almost diametrically opposed) seek to optimise profit over longer periods of time. At any specific time each strategy is preying off the other ones.
Lotka Voltera
Farmer uses a Lotka Voltera example of which involves rabbits and foxes: If there is an overabundance of rabbits the fox population increases, but it tends to overshoot, causing the population of rabbits to plummet until foxes begin to starve, which in turn allows the population of rabbits to rise again, completing the cycle. Farmer’s point is that in markets the population of a species is replaced by the capital deployed in a trading activity.
So it was entirely obvious that as more money flowed into quant trading strategies, and away from traditional “buy and hold” portfolio managers, then those quant strategies would struggle. Too many foxes and not enough rabbits. Quants have become the “dumb money”. Lotka Voltera being a mathematical model, most quants will get this. It also suggests there’s not much quant funds can do about the situation, until they start suffering outflows, competition is likely to remain tough.
Beating the quants and the index trackers
It’s an oft quoted statistic that 90% of active fund managers underperform their index over a 10 year period. So I’m feeling pretty smug that not only do I seem to be able to pick long term winners like Games Workshop. I can also pick one of the very few active fund managers to outperform. TM CRUX UK Core Fund has generated 157.72% over ten years on the accumulation shares against 118.32% for the All Share total return, the annualised equivalents are 9.92% and 8.12%. The fund was started by Patrick Barton, who I used to cover UK banks as a sell side equity research analyst with many years ago at Credit Suisse. Since Patrick founded the fund it is up 159.7% and ASXTR is 97.54% which are 8.1% and 5.71% annualised respectively. The fund is now managed by Jamie Ward, who has continued to do well (the regulator would probably like me to point out past performance doesn’t reflect future performance and all that). NB this is not selection bias, this is the only active fund manager I have backed in the last 20 years.
I should have bet against Buffett!
So when, in 2007, Warren Buffett announced his $1m dollar 10 year index tracker bet
… I should have taken him on! Sadly, I was confident enough that Patrick Barton’s fund would outperform over a 10 year period that I backed him with tens of thousands of my own money. Unfortunately I wasn’t quite confident enough in Patrick’s fund to bet a million dollars against Warren Buffett !?!? Also, Warren being the old shrewdie that he is, might also have objected that Patrick’s low turnover fund wasn’t really a hedge fund or specified that he needed to beat the S&P index, rather than the FTSE All Share benchmark.
Another former UK banks sell side analyst
Interestingly Terry Smith who is also a former UK banks sell side research analyst has outperformed his index too. His Fund (T class accumulation shares) is up +364% since inception and remains the No.1 performer since its inception in the Investment Association Global sector by a cumulative margin of 233 percentage points above the average for the sector which has delivered +132% over the same timeframe.
Although Patrick & Jamie’s fund invests in different companies to the Fundsmith one – they have both adopted similar low turnover “buy and hold” styles. Aside from the fact they are both former UK banks analysts there is another similarity: they both buy quality companies and hold them over many years. There is a certain irony that the “active” fund managers who have outperformed have done so by doing very little. Back testing? Perhaps some of the quant funds should have tried back testing “buying concentrated positions in quality companies and passing the time by writing very readable fund newsletters to their loyal investors”…
Thinking about thinking
I wonder if this is relevant for the debate about automation and machine learning algorithms taking our jobs? Algorithms are just hard wired rules. Human beings are on the other hand are better than algorithms at thinking about thinking (and people like Doyne Farmer, Patrick Barton and Terry Smith are better than most).
You can tell that I don’t think that it is a coincidence that these former UK banks’ analysts have done well. UK banks is the most competitive sector to be an analyst in – portfolio managers generally have a better understanding of banking (after all, they work in finance!) than sectors like insurance, oil companies, utilities or drug companies. I’m happy to admit I’m not on the same level as Terry or Patrick, but covering UK banks for many years did teach me some valuable lessons about what works and what doesn’t; how just because you hire lots of PhDs doesn’t always make your decisions better (HSBC used to proclaim that Household International hired lots of PhDs!). It was also stressful, one of my competitors at Teather and Greenwood died of a heart attack in his early 30s.
Investing where uncertainty is greatest
Now, rather than large UK banks, I’m focused on the much smaller end of the market, where uncertainty (not the same as risk) is greater, bid offer spreads are terrible but returns can also be greater. Bid offer spreads are also so wide that of necessity investors have to trade infrequently. The professional fund managers can only invest in large companies because they fear liquidity risk, so effectively they are betting that great tennis players like Roger Federer and Rafa Nadal will continue to be great tennis players. They are! But even Roger will have to retire someday.
My approach is trying to identify a 16 year old, who has the potential to make it into the tennis top 50. I think this approach could be aided with Natural Language Processing to analyse the voluntary disclosure: how the management tell their own story. But at the moment I still read a lot of Annual Reports with my own eyes.
Low frequency, but important, decisions
To me, it is also interesting that computers seem to be very good at making lots of frequent decisions that are just about more often right than wrong (say 55% of the time) that make small marginal gains. Computers don’t seem to be good at analysing lots of data and making a few important decisions (perhaps one per year). Casino’s and insurance companies know that they can use probability and they will come out ahead over a year. But many human decisions we only get one shot at, and although the brain is not perfect, some of the biases (avoid risk of ruin, even if it is a low probability event) which at first appear irrational, actually make sense.
No one seems to even think it is possible to build a neural net that is good at thinking about thinking. For that reason, perhaps some of the most valuable skills of the future might not be economics (which doesn’t seem to teach much thinking) engineering or programming. But perhaps more “useless” degrees, thinking about thinking, understanding ambiguity whether Philosophy, Classics, History (both Terry Smith and Patrick Barton studied history at university) or even Theology are good for teaching skills that could turn out to be the most valuable. Perhaps employers, who are paying big sums of money to hire “data analysts” (is there any other kind of analyst?) with similar statistical inference skills to the quant funds, are the ones being dumb?
UPDATE: Marc Rubinstein, another banks analyst, turned investor, turned blogger that I used to work with has sent me this link from the FT about how active fund managers are trying to use big data. So called quantamental investing. I don’t think that “more data” is the answer to beating computers though. In my experience, it is the ability to look at the same information and draw better conclusions which marks out long term performance. Both Patrick and Terry Smith were well known for publishing SELL notes explaining why a simple P/E ratio was the wrong way to compare banks. Terry Smith famously published a negative research note on Barclays, when he was employed by them at BZW. Patrick warned of increased risks at Royal Bank and Northern Rock before the credit crunch, because of the hidden risks in their leveraged business models. If you contact me on @brucepackard on twitter I can send you through the notes Patrick published.
Photo by Thomas Lipke on Unsplash
Disclosure: I’ve invested in Patrick Barton / Jamie Ward’s fund since inception. No one has paid me to write this. Performance data provided by the fund managers to end December 2019. Article is not intended as financial advice, instead: Do Your Own Thinking. Economists’ theories can be ignored most of the time, but even a stopped clock tells the right time twice a day.
About Post Author
