Silicon Fen’s Hedge Fund

Cantab Capital engineers returns with systematic macro approach


Often the holiday island daydream, for someone who enjoys their job, is to carry on with much of what they are doing but remove the bits they don’t like – the politics, bureaucracy and lack of nimbleness and the like. People who set up hedge funds often look to replicate the culture of a successful trading desk, or investment culture, of a successful asset management business as they start off in their own firm. Dr Ewan Kirk, of systematic investment firm Cantab Capital Partners (CCP), has done something like that in taking some of the strengths of Goldman Sachs into his billion dollar business, whilst taking full advantage of the management company’s location.

Sometimes the name of a hedge fund firm is some combination of the names of the founders (Marshall Wace or Brevan Howard), or reflects an aspect of the trading strategy such as Vega, or Sector Asset Management. Another method of nomenclature is to utilise a relevant place name, like Pershing Square Capital Management, or Standard Pacific (of San Francisco). In the case of CCP, where Cantab is an abbreviation of the latin word meaning “of Cambridge”, the name is very significant, and in a way which goes beyond an archetype of geography.

Cambridge in England is not a million miles in spirit away from Cambridge Massachusetts. The University of Cambridge in the English fens has a cousin in Harvard in Cambridge Massachusetts, and just as Harvard is now known as a hotbed of technology-focused bright young things (“The Social Network” etc.), so Cambridge in England is a focus of high technology industries. Indeed by some estimates 6% of Europe’s spend on R&D is made in the Cambridge area, and the city is home to clusters of technology firms including Microsoft, Autonomy and ARM, the chip company.


Progressing at Goldman
So it was natural that when Ewan Kirk finished his Ph.D in Astrophysics from Cambridge he worked for a few years in high tech companies in the area before being approached by Goldman Sachs. It was indicative that Goldman approached him, and significant that he was recruited into the J.Aaron operation – the commodities part of the business.

“I started in the commodities business in oil,” says Kirk, “then moved to metals and grains. I was made a Managing Director in 1998, and took responsibility for the currencies part of the Strategies Group. I was made a partner at Goldman in 2000, and ultimately took responsibility for all the Strategies Group staff in Europe.” These 150 people in Europe for whom Kirk was boss worked in currencies, rates, fixed income, equities, commodities and credit. The Strategies Group were embedded specialists in technology and derivatives who could work with sales staff (and traders and research where necessary) in each asset area to create solutions for clients. This might involve adding a knock-in option as part of a structure to better express a client view or cheapen the cost. The Strategies Group staff were creative facilitators within the asset classes but had distinct reporting lines and responsibilities from the asset class-focused specialists.

The point from the perspective of a potential investor in Cantab Capital is that Kirk had to deal with market-related issues, helping clients make money or protect capital at a high level, and in different asset classes. This involved growing and managing a group of bright, capable technologists and modellers. Kirk was not taking principal risk – the market risks were taken by the end clients or by those responsible for the hedging or running of books at Goldman.

Market events during the period when Kirk worked in the Strategies Group presented some great learning opportunities. “Although the credit crunch of 2008/9 was an incredible time to be in markets (and to be invested in markets),” points out the CCP CEO, “it felt equally cataclysmic in 1998 when Russia defaulted, and the Fed needed to arrange a rescue for LTCM.” And indeed the TMT Bubble shortly followed. So although the 2008 credit crunch was extraordinary for markets, CCP trusted their models and they performed as expected.

Alpha models and volatility
CCP describes itself as a systematic global macro hedge fund manager, though inevitably it is compared to AHL, Winton, and Aspect, and like them the management of CCP accept that the term systematic CTA is also going to be applied to them. The firm was founded in 2006 by Kirk, Erich Schlaikjer (also ex-Goldman Sachs) and Chris Pugh (ex-KBC and D.E. Shaw). The three founders remain the managing partners of the business. There are 26 full time employees of whom 17 are dedicated to research and development. The average age at CCP is in the low 30s.

Typifying the structured, prepared approach Kirk says that “we undertook nine months of model and system development before the investment of any capital.” The CCP Quantitative Fund was launched in March 2007, and now contains five share classes offering investors a choice of volatility targets and liquidity terms. To be clear, each class is invested in the same Cayman Islands master fund. Investors that need a lower volatility of return would tend to gravitate towards the Babbage class which runs with a target of 10% volatility. However, as is usual where there is such a choice, most of the capital is invested in the Aristarchus class (and managed account equivalent) that targets 20% volatility on a gross basis, and which should produce the higher return.

Alpha generation at CCP comes from as many as 230 models that operate over a range of timescales applied to 80 markets. Within that there are up to twenty main models. The risk allocation between the various asset classes such as interest rates and equity futures is systematically determined and can vary considerably as opportunities present themselves in each class. Allocations to a single asset class can vary from 10% to 40%. So fixed income might be a large minority of the risk allocations at one point and three months later it could be just 10-15%. The signal generating models are clustered into three broad baskets as part of the structural effort to diversify the sources of alpha – medium-term momentum, a value orientated and carry focused basket, and short-term trading signals.

One of the reasons why investors in hedge funds pay their premium fees and are able to sleep at night is that the high end of the industry not only has better sources of alpha, but has a much greater focus on risk management than long-only managers. Across all strategies this tends to mean that risk measurement is much more actively used in hedge fund businesses and therefore meaningful to the investment process, but in the large systematic CTAs this is augmented by very thorough portfolio construction approaches that typically define and build the portfolios around forecasts of key metrics.

So it is at Cantab Capital Partners – risk management is at the heart of the investment philosophy. Multiple complementary and interlocking systems dynamically reallocate risk across the entire portfolio to ensure that a finely calibrated constant risk profile is targeted. Target positions, risk, and profit and loss are calculated and monitored in real time for the entire portfolio. In terms of product differentiation there is an emphasis on forecast volatility, but a key operational measure is ‘expected shortfall’.

Expected shortfall is a measure of tail risk and is often known as conditional value at risk. Whilst traditional value at risk determines the minimum loss with some confidence, expected shortfall is the average loss when one has a large loss. This takes into account the “shape” of the tail which can hide many horrors. Investment managers running hedge funds periodically take risk management related decisions which they know are potentially stay-in-business decisions. Maintaining risk levels or halving them is a key decision in a crisis. If expected shortfall is used in a feedback loop with aggregate portfolio exposure (and other risk measures such as forecasted volatility or the running P&L) the manager is implicitly running the capital with a wary eye on his business risk.

In order to meet institutional investor requirements CCP aimed high from the off on the hygiene factors of operational efficiency and compliance checks. Trades are reconciled and confirmed in real time, and senior members of the firm meet daily to review the performance of the strategies and the systems. There is a weekly risk committee meeting which evaluates and approves new strategies and technologies.

People, technology and data
“When I was at Goldman Sachs we spent a lot of money on technology and a lot of money on smart people,” Kirk says. “Leadership in technology is often said to be one of the key success factors for Goldman Sachs. For example, when I was in the commodities group we employed 24 quants, and at that time our closest competitor was Morgan Stanley who at that point employed two quants in commodities in Europe. And we utilised great technology.”

Just about all mature funds in quantitative finance will say the same: technology and people are the two most important elements of the fund management effort. CCP is not going to outspend D.E. Shaw or Renaissance Technologies, but has established an architecture and approach to enable rapid research testing – which it proudly demonstrates to potential clients.

This is a key attribute in this investment strategy. Robert Litterman of Goldman Sachs Asset Management told the Quant Invest 2009 conference: “You (quants) have to adapt your process. What we’re going to have to do to be successful is to be more ‘dynamic’ and more opportunistic and focus especially on more proprietary forecasting signals … and exploit shorter-term opportunistic and event-driven types of phenomenon.” This change of emphasis is something like what has happened in pharmaceutical research – a shift from narrow testing on a family of high potential candidate compounds to assessing a much broader range of materials on a faster basis. The research has a much larger ‘process’ element, and is less reliant on a spark of genius to produce results than historically. Research productivity can be a competitive advantage, though hopefully there is some spark of insight and academic curiosity too.

Bob Litterman also observed at the same conference that “we’re putting together data that’s not machine-readable, and finding databases that haven’t been explored nearly as well as others.” This has also been put into practice by CCP. Access to different data sources can be an edge to a quant, just like access to unusual information sources (proprietary research) can provide an edge to a fundamentally driven discretionary equity manager. CCP has a proprietary market database of unusual depth and integrity, which itself may be an edge for the firm.

Taking the lesson from Goldman about employing smart people, and utilising its location CCP takes advantage of the fact that Cambridge produces more than 600 quantitative Ph.D s a year. “We interview six people a week here,” says Kirk, “and last year we hired four of them.” Interviewees eventually meet every member of the firm – thorough and wide interviewing is another investment bank method. “We like our candidates to be super smart and with the ability to get things done. We hire junior raw talent that we can help to grow in the way that we think appropriate,” he says. The thinking is that it is easier to give a smart person market understanding, than to take a person with market experience and make them expert in academic areas like machine-learning and Monte Carlo techniques, or mission critical high frequency software.

Having created a good work environment for programmers, statisticians and mathematicians (think of Google Friday or an internet start-up, with a foosball table and free food ordered in on Friday) Kirk has identified that one of his skills may be in creating something from the talent brought in. The difficult thing in putting the firm together he says, is melding that talent pool into “buttoned-up, tightly controlled, verifiable tested systems which are surrounded by a rock solid operational infrastructure.”

For example CCP tracks the correspondence between the true models (the ones running live), to the back-test carried forward. So if you take your back-test and run it for one more day, the P&L of the model should be the same as the P&L in the live version. Having calculated it, over time the difference between those two versions has been tiny.

Aside from ready access to raw Ph.Ds CCP’s location give it another advantage. The firm has close links with Prof. Chris Rogers, a top British statistician who heads the Cambridge University Statistics Laboratory. He is described by Kirk as one of the brightest people on the planet, and as Research Director of CCP he helps to guide the overall academic thrust. Rogers is particularly well-suited to this role as he has extensive experience of working in the financial industry and acts as a key conduit for financial technology transfer between academia and the markets. “He is not there for decoration,” affirms Kirk, “but he is also not there day-to-day. So he is not an alpha generating head of research, rather his job is to ensure we have (academic) rigour in what we are doing.” Like all good senior researchers Rogers can be creative as well as thoroughly logical.

UCITS hedge fund

CCP is launching a UCITS version of its offshore fund. The Cantab Quantitative UCITS Fund will open in 2011, launched on the Matrix Group platform. Ewan Kirk sees some advantages to UCITS. “To some, UCITS hedge funds are just the next 130/30,” he says. “That is, a fabulous way of introducing retail investors to higher fees,” he says with a deadpan tone. “For some forms of investment the risk framework of the regulations, like the use of Value at Risk, is a very useful discipline, but not for us and other funds with systematic approaches. We understand our risk to such an extent that we don’t find complying with the regulations in any way onerous or an overhead.”

According to the CCP view, the emergence of UCITS hedge funds does bring a degree of risk oversight and regulation to an area that has had light regulation. Kirk outlines a plausible vision of the future: “It is conceivable, though maybe not likely, that in five years’ time every investor who could be in UCITS will have moved all their hedge fund investments into that format. Why not?”

“What we have is the commoditisation of hedge fund strategies. What we might see is a bifurcation in the hedge fund industry – those funds that can provide daily or weekly liquidity (fund strategies invested in liquid futures markets or large cap equities) will be onshore and everything else should be offshore. The non-liquid (offshore) strategies should employ a private-equity-like revenue model – then all investors will know that if you invest in distressed debt you can’t expect to get your money back next month.”

Clearly, there is a constituency of investors who will never invest in offshore funds for regulatory reasons. So for Cantab Capital the UCITS format gives access to such investors. Even for some existing clients there may be utility in moving their investments because of regulatory change they are experiencing. Another rationale for existing investors is that some want to launch funds of UCITS hedge funds (about 40% of the investor base of CCP is funds of funds and 40% institutional).

In good company
Cantab Capital has made great progress in the three and three quarter years that the CCP Quantitative Fund has been running. In some ways the firm has overachieved. For example, there are few performance differences for CCP versus a more established second tier (by size) CTA like Rivoli International, yet the Cambridge-based management team manage more capital. Other firms tap into smart raw talent, use advanced statistical approaches and all of them describe a similar portfolio risk control approach. Why should Cantab Capital Partners have got commercial traction more quickly?

cantabtable4First, there were no legacy performance hiccoughs to overcome in building the track record since 2007. The infrastructure of the firm and style of operating comes across as similar to the best of breed amongst first-tier CTAs (meeting the institutional imperatives). That CCP ticks the boxes is illustrated by a recent win though this wasn’t mentioned in our interview. CCP is one of six firms appointed by the UK’s Pension Protection Fund (PPF) to run Global Tactical Asset Allocation mandates – the other firms are Winton Capital, Aspect Capital, BlueCrest Capital, DB Advisors and Neuberger Berman. Only CCP and Winton were given immediate appointments, meaning that the firm received capital ahead of Aspect and BlueCrest with the other funds on deferred appointment. Good company for CCP and obviously a source of satisfaction.

CCP has at least two additional features which may impress investors. One is the halo effect of former Goldman Sachs partners encapsulating some of the authority of that institution – something which is common enough with hedge funds. The second feature is a distinctive Cambridge flavour, a local flavour if you like. The academic rigour, the technology platform and software and programming skills are all part of it, but the fact that the CEO cycles 500 yards to work is so of Cambridge, so Cantab.