Cantab Capital On Debunking The Myths

Cantab Capital On Debunking The Myths

Cantab Capital – DEBUNKING THE MYTHS, original article here 


Dr Matthew Killeya dispels a few of the myths surrounding systematic investment.

Out on the road and back home in Cambridge we meet a lot of people and speak to them about systematic trading. The conversations are often interesting and overall it has been a joy to witness people respond positively to the systematic space over the years. Yet, in spite of our best efforts, we have seen a few myths spring up around our industry. So here goes my attempt to deal with some of the more persistent ones.


This Credit And Equity Fund Saw Sizable Contributions From Its Stocks In Q3

Arena Investors Chilton Capital Management Schonfeld Strategic Advisors Robert Atchinson Phillip Gross favorite hedge fundsThe DG Value Funds were up 2.7% for the third quarter, with individual fund classes ranging from 2.54% to 2.84%. The HFRI Distressed/ Restructuring Index was up 0.21%, while the HFRI Event-Driven Index declined 0.21%. The Credit Suisse High-Yield Index returned 0.91%, and the Russell 2000 fell 4.36%, while the S&P 500 returned 0.58% for Read More

The first thing to say is that it’s certainly true that the systematic or CTA space and computer-based trading has expanded significantly in recent times. Perhaps this is not surprising given the similar expansion in computing power in that time. The amount of data and available computing is increasing all the time yet the fraction of the world’s capital invested in systematic firms from large institutions(1) is still small.
Systematic firms generally trade the largest, deepest markets. These are liquidly-traded futures and forwards markets, large cap stocks(2), and vanilla options markets. The key is not how large positions are, but how large you are as a fraction of the market volume. Systematic firms are very focused on transaction costs and liquidity in the assets they trade and allocations to these assets are scaled accordingly. Their focus on data and technology allows them to accurately estimate trading costs on any asset for their own trading and estimate — and control — costs of liquidating their own portfolios.(3)
There’s a lot more we could say about this one, but instead we direct you to our research paper on this topic. We’ll move on to our second myth to be debunked which is often conflated with the first.


The genealogy of our industry has seen several large funds spawn children.(4) Whilst this might give rise to concerns about similarities, splits often emerge because the founders of the new firms have different visions to those of their previous employers and these again are a force for differentiation over time.(5)
The best systematic firms are extremely active in research and development. They employ teams of scientists who research new strategies and analyse data constantly. Thus our products are constantly evolving and there is reason to expect further divergence over time. For example, satellite imaging data has only recently become generally available. This offers the potential to measure the amount of sunshine or ‘good growing days’ for various crops around the world. Augmented with crop production figures per geographic location, researchers can build models to predict crop yields and ultimately to profit on price moves.
Systematic firms are focused on the similarity issue themselves. They can estimate their similarity to other funds and quantify this risk. By building factors such as ‘momentum’ and ‘value’ firms can regress the returns of their peers against these factors to understand to first order what they are doing.(6) Neutralising or orthogonalising your own returns to limit your exposure to these factors is also natural. Statistics offers a large number of sophisticated and subtle tools to do this.(7)


In fact, systematic firms start from the same point as discretionary or fundamental trading: with a hypothesis. The hypothesis could be that interest rates of two countries predict the exchange rate between the two. It could be that the slow release of news or herding behaviour of markets — which after all are simply collections of people — leads to trends or momentum in markets. The key difference is that, given a hypothesis, systematic funds take the next natural step as scientists: they use the wealth of available historical data and available statistical techniques to learn about, test, validate and refine the hypothesis. And once this process is complete — and only then — they use it to invest.
One limitation is that hypotheses are only useful if they can be legitimately tested by data. If one cannot hope to test the validity of a hypothesis, then it is discarded. This is no bad thing.


People are comfortable flying in airplanes, relying on satellite navigation and using smart phones. All these use complex mathematics and technology that most of us cannot hope to fully understand. However, most of us are prepared to risk our lives in an airplane that we cannot fully understand.(8) Paradoxically, people can baulk at systematic investment that has a degree of complexity.(9)
Einstein (10) said “everything should be as simple as possible and not simpler”. This sentiment holds true with systematic investment. We apply a principled hypothesis-based and disciplined approach and carefully construct a process to manage billions of dollars.
That said, we do not over-indulge in complexity. We follow the statistical doctrine of searching for the simplest explanations and models to explain the data. The reason is simple: simpler models have a lower likelihood of falling foul of ‘over-fitting’ and thus a higher likelihood of realising profit in the future.(11)
Conversely a discretionary firm could be viewed as largely a ‘black box’(12) from the perspective of an investor. It is a collection of portfolio managers investing based on the most complex process we know – the human brain. At least systematic trading has the advantage that its process is defined exactly in a set of rigorously-tested computer code and is fully reproducible over history.


Systematic firms are unique in their ability to precisely target and control their level of volatility. Many of us therefore offer a constant level of volatility in our funds. Moreover this level is completely tuneable so that literally any level of volatility can be offered to investors.(13) Though headline volatility levels are often large(14), many can and do offer lower volatility share classes. This volatility tuning is also offered to managed accounts and lower volatility levels can be tailored to investor needs. The assets we trade allow us to rescale positions without constraint and the technology we deploy comes into its own in allowing us to run multiple copies of our products, and different volatility levels, with close to zero tracking error between them.
The constant level of volatility on offer is also a great strength of systematic firms. An investor in an equity index tracker would have seen volatility vary wildly over the last few years — from below 10% to over 50% at the time of the credit crisis. Compare this to many systematic firms that maintained a constant level of volatility through the largest financial crisis in living memory.


There is no real evidence to support this. Volatility is the key quantity in determining drawdown levels(15). The absolute size of expected drawdowns of an investment is linearly related to its volatility level.(16) So halve the volatility level and you halve the expected drawdown. As we already discussed, the volatility level of many systematic funds is constant and completely tuneable. Thus expected drawdown is also similarly tuneable.
One key advantage that systematic firms have is that they run their models back through decades of market history. This enables them to build up a deep understanding of the detailed statistical characteristics of their investment programmes.(17) They can do this because they build repeatable algorithmic processes. It makes sense to run these processes through history and this can be done in a robust and reliable fashion. Discretionary traders have no such luxury, because — even if they did have access to the decades of high quality data and analysis — they don’t have a repeatable statistical process. Therefore they can’t sensibly define what their hypothetical positions would have been in the past.


Like all investments, systematic funds are exposed to Central Bank intervention, freak hurricanes and geopolitical events.(18) Whilst this is true, the statement they are moreexposed than traditional macro or fundamental traders is a myth. In fact there are good reasons to think the opposite. An event such as the abandoning of a currency peg is a very difficult one to predict.(19) In fact the world is full of events which are very difficult — or even impossible — to predict. Our edge is recognising this difficulty and confronting it head on. We are well versed at recognising the limits of quantitative modelling and its ability to predict. Thus rather than professing expertise in specific assets, we are generalists and invest in a great many things. We are diversified. That way we are generally robust to localised events in any one asset: a Canadian rate cut, a Brazilian drought, or a surprise UK election result. An emerging market currency fund may have a little extra information about Brazil or China but they are typically much more exposed to idiosyncratic events which they can’t hope to predict reliably.


Systematic funds have a wealth of historical data and tools to analyse it. This is very powerful but also potentially dangerous since there is a danger of over-fitting to history.(20) The issues with historical data analysis are subtle and nuanced — and will no doubt be the topic of future blog entries — but this is why we employ highly skilled statisticians. All data contains information, but there is skill in extracting this information.(21)
Some of the origins of systematic trading can be traced back to academic researchers. There, researchers historically sought to use market data to construct rules for ‘systematic trading’. Lack of available data within the academic environment(22) has perhaps led to its over-analysis and some over-fitting. Systematic firms by contrast have their research tested in the markets every single day.(23)
There is now a large number of systematic firms with long and expanding track records of profitable performance. Viewed one way, this is a collection of scientists amassing evidence in support of their form of investment.


A human trader is — whether they admit it or not — an empirical scientist. He or she is making observations and learning about the markets each time they execute and order and interact with a market order book. Systematic firms operate on a similar premise albeit with a more rigorous application of science.
Systematic firms value scientific skills highly. We need them to analyse the wealth of data at our disposal. At the same time we maintain a healthy scepticism about our model. We are focused on events that invalidate or are not captured in our datasets. A historically low rates environment, for example, begs questions of risk management of bonds positions. There may be concern about risk being skewed against long positions, with rates rises more likely than falls. Thus, in this example we develop ways to measure skew and optimally feed it into all bets, not just rates — a general solution to a general problem.
A focus on technology and transparency means that many people have eyes on the investment strategies. Similar to the assets traded by the strategies, they observe their investment process too and feed this information back into long-term research.


Finally remember that systematic or discretionary, no investment firm has all information available to them. Nobody has a crystal ball and nobody can predict the future with certainty. Decision making under conditions of uncertainty are the bread and butter of professional statisticians and systematic investment, where acknowledging and quantifying uncertainty is key.

Linkback to article here

Updated on

No posts to display