QFS Asset Management - taking a unique view on the business of managing money

e-Forex talks with two new members of QFS Asset Management, Karlheinz Muhr, CEO and Jim Conklin, Ph.D., Head of Research. QFS is a leading asset and risk management firm based in Greenwich, Connecticut, USA.

First Published: e-Forex Magazine 48 / Tradertalk / July, 2012

How long have you been in the asset management business and what attracted you to the industry? 

Karlheinz: I have been in the investment business for over 25 years. I joined QFS last year, and became the CEO as a result of the acquisition of Cenario Capital Management. I was the Chairman and co-founder of Cenario after leaving Credit Suisse in 2009. Cenario provided investors with volatility risk advisory and management services. Additionally, we developed a proprietary Risk Index (the CRIX) – a leading indicator of heightened risk aversion and market turbulence – which is still available on Bloomberg Terminal. I was fortunate to have been one of the people in the industry to have pioneered the concept of volatility as an asset class. In 1999, I formed Volaris with my partners, which was sold to Credit Suisse in 2003. I stayed at Credit Suisse and served as a Managing Director of Credit Suisse in the Asset Management Division and was head of Credit Suisse Volaris. I was also a member of both the Chairman’s Board and the Management Council of Credit Suisse. 

Jim Conklin and Karlheinz Muhr

Jim Conklin and Karlheinz Muhr

My academic background is in economics and finance. I got into the investment management business by starting on the fixed income trading desks of First Boston Corp. in 1985. I went on to hold a variety of positions at UBS Warburg and Credit Suisse First Boston, spanning from head of Global Credit Fixed Income to head of Derivative Mortgage Trading. I think this helped shape the way I think about risk in the context of asset management. At QFS, we don’t believe there is such a thing as a free lunch. I believe prudent and successful investment management comes from a manager’s ability to truly understand and manage risk. The investment decision process should not start with “what return do I need or target?” But, “what is my risk budget?”

Jim: I joined QFS last year as well, as the Director of Research. I have over 14 years of investment experience, working as a quant-trader and strategist in FX volatility trading, currencies, commodities, fixed income relative value, and equity long-short at various institutions. Before joining QFS, I was Head of Investment Research at FX Concepts and a member of the Management Committee there. And before that, I was a Portfolio Manager at Fortress Investments in the Liquid Markets group. Like Dr. Grossman, I came to finance from academia, having held faculty positions at the University of Texas in Austin and the University of Madrid Carlos III. I switched careers because I wanted to apply the skills and knowledge I learned in academics in a fast-moving environment that rewards decision-making.  

My time at QFS has been especially gratifying because I share convictions with my colleagues about the investment research process and risk management. My doctoral thesis in economics incorporated game theory, dynamic modelling and numeric computation to solve models of sovereign debt repayment. Those models amount to understanding why sovereigns choose to pay or default.  In the mid-90s my work was beginning to seem a little ‘academic’ due to a dearth of defaults; however the last decade has provided a rich environment for application of my area of specialization. Coincidentally, Dr. Grossman was a professor of economics at Princeton while I attended as an undergraduate student. Later I got my PhD in economics from Stanford, which gave me the opportunity to study under Robert Hall, Thomas Sargent and Joe Stiglitz. It was an exciting time to work with scholars whose research focused on the interaction of uncertainty, imperfect information and government policy, and how these factors affect asset prices, business cycles and macroeconomic growth.

When was QFS Asset Management (QFS) founded and what were the original objectives for the company when it started? 

In 1988, Sanford Grossman, PhD founded Quantitative Financial Strategies, Inc. as a financial research firm to develop investment models using his pioneering work in quantitative finance. The QFS Currency Program is based on research in international financial economics that explores the transmission of information through asset prices in the presence of shifting global capital flows. 

One of the most significant publications that relates to the origins of the firm is The Informational Role of Prices (MIT Press, 1989).  The idea of being able to make inferences from asset prices to make economic forecasts forms the fundamental underpinnings of the firm’s approach to investing. The firm’s original objectives remain the same as they are today: to provide investors with a favourably risk-adjusted investment returns in liquid markets that have low correlation to other components of their portfolio, while using a rigorous and systematic method of risk reduction and optimal position sizing.

Who are the key people involved in the firm and what are their main day to day responsibilities? 

Karlheinz: We have made changes at QFS to give the firm a more institutional structure. We have operating committees with Dr. Grossman and me as the two members of the Board. I chair the Management Committee and he chairs the Investment Committee. Naturally, we have many of the other expected committees, such as compliance and operational risk. While we are a relatively small firm, we have found that instilling these processes helps us achieve our goal of industry best practices. 

Jim Conklin is the Director of Research and manages a team that includes five PhDs.  Steve Van Besien, a co-founder of Volaris and Cenario Capital Management, chairs the Origination Committee. Steve has a background of over 20 years advising clients on risk. Robert Shustak is our COO and CFO and chairs Operational Risk, the committee that oversees operational and accounting functions. David Zimmerman is our General Counsel. Jonathan Silber is our Head of Trading and George Holt is our head of Technology.

I would like to mention that a fundamental change that has taken place at QFS is my assumption of the CEO position. This gives my partner and the founder, Dr. Grossman, the opportunity to be fully engaged in the research process. We are already seeing the benefits of a more invigorated and accelerated research process. Naturally we are fortunate to have Jim Conklin managing the process and contributing his ideas and skills. 

Jim: We are pursuing a business plan to build up the research team slowly and steadily over the coming years.  We enjoy a marvellous legacy of a rigorous research tradition. We have an extensive library of proprietary research the firm has done over the years, and a team of senior researchers trained in and committed to our research process.  As we expand the team, our goal is to both maintain our research tradition and enhance our pace of innovation. Consequently, our hiring process is quite selective and we plan to grow the team slowly over time, interviewing doctoral level researchers from both finance and non-finance backgrounds.  We hire for both strong skills and training and for new ideas and perspectives. 

Jim Conklin

Jim Conklin

What asset classes does QFS currently trade and what percentage of your business lines are focused on currencies? 

The QFS Currency Program trades currencies through OTC forwards and options.  The QFS Fixed Income Program trades sovereign debt bond futures and futures on short-term interest rates (such as Eurodollar and Euribor rates).  In addition to currency forwards and bond futures, the QFS Global Macro Program also trades equity index futures and commodity futures. We trade currencies directly in the QFS Currency and Global Macro Programs, but QFS’s research on all asset classes benefits all of QFS’s investment strategies, even if we don’t explicitly trade a specific asset class in a particular strategy.

How does Risk Management influence and shape your overall investment philosophy? 

Karlheinz: There is an old axiom in our industry. There are only two things you can really control: Risk and Fees. So, you better get those right. We believe we do a good job on execution costs for our trading and we really focus on risk. Most money managers build their methodologies from the question “what return do I want?” We turn that upside down and start by asking how much risk we want to have in the portfolio. After we’ve established a risk budget, the market will tell us what the expected return is at any given point in time. Everything we do after that is to manage the expected return: our trades embody our processes’ reaction to shifts in expected return opportunities within a determined risk budget. Jim will discuss our Drawdown and Scaling Technology later in the interview.

What currency programs does QFS offer and how have they recently been performing? 

Karlheinz: QFS offers one standalone currency product, the QFS Currency Program. It can be accessed through various structures or one of many platforms. Performance has been strong as we have moved through the period of Great Deleveraging (July 2007 – April 2012).  Over this time frame the QFS Currency Program has an annualized return of 9.14% on a composite net basis.  This has been a very difficult period for traditional beta exposures, which is why the QFS Currency Program acts as a tremendous portfolio diversifier. Additionally, we offer traditional and alpha currency overlay and bespoke client structured mandates.

What are the main factors that drive your firm’s currency investment decisions? 

Jim: Two major themes drive our approach to investing in countries via currency markets:

The first macroeconomic thesis is that divergences in growth from expectation drive capital flows to their uses of highest marginal productivity across the globe.  By identifying such shifts at the business cycle forecast horizon, we endeavour to anticipate future currency movements.  That is, as our signals detect fundamental macroeconomic divergence, where one economy is projected to accelerate from its prior expected growth path, our forecast anticipates an appreciation in that economy’s currency.  On the other hand, our forecast falls when we detect deceleration relative to a prior expected growth path.

The second thesis embedded in our currency forecasts is that at a long-run horizon the relative soundness of monetary and fiscal policy is a key determinant of the movement of exchange rates.  In similar fashion to the divergence thesis above, we derive sufficient statistics that form a future projection of the credibility of a nation’s commitment to sound money.  To illustrate the concept of a sufficient statistic that proxies for broader, long-term monetary soundness, consider the 5-year - 5-year forward break-even inflation rate in the U.S.  As the rate rises, it signals that participants in both inflation-linked and nominal U.S. government securities markets believe that the inflation rate 5 years into the future will rise, whether by lack of institutional commitment on the part of the Fed or by circumstances imposed on the Fed by the U.S. Congress.

Karlheinz Muhr

Karlheinz Muhr

In what fundamental ways does your investment approach differ from many other money managers? 

Jim: Principally, we are an investment firm dedicated to trading fundamental macroeconomic ideas.  As an integral part of our investment process, we employ the scientific method to vet and refine our ideas.  Specifically, our research process always begins with a fundamental thesis or idea; for example, “international capital flows are an important determinant of movements in foreign exchange rates.”  Next we derive a falsifiable, quantitative expression of that idea and we look to test it against data.  In fact, the vast majority of the research documents that we generate describe empirical implementations of ideas that we are able to falsify and consequently do not trade.  In this sense we differ from the majority of “quant shops” or CTAs, in that they are typically more agnostic (Bayesian) in their use of signals and factors.  Often a “quant” firm emphasizes the systematic nature of their investment process, and typically uses factors or transformations of historical data that have forecasted returns successfully in the past.  The essence of our process is our focus on fundamental causes in the research process from the outset. 

Karlheinz: QFS’s competitive advantage rests on the fundamentally-based investment processes developed by QFS’s research team over the past two decades.  Some of the hallmarks of QFS’s approach to investment process development include mastery of academic research in economics, finance and quantitative fields (our research team’s five PhDs have 80+ years of experience in the academic and private sectors); the centrality of fundamental rationales behind our investment theses, whether macroeconomic, micro-structural or behavioural; systemization, automation and diligence in the daily operation of our investment processes; strong emphasis and resource dedication to risk management and draw-down control; and finally, ongoing research and development as the core competence and activity at the firm.  The firm is widely considered a leader in the field of global macro trading, quantitative risk management and advanced portfolio optimization techniques.  The firm’s research culture is dedicated to scientific investigation and reliance on detailed, repeatable methods to govern investment-related activities. We have found that our Currency Program has a low correlation with other hedge funds or trading advisors.

Do you have any particular preference for specific trading time horizons, such as medium to long term? 

QFS’s investment Programs take a unique approach to producing alpha based on macroeconomic fundamentals that play out over a business cycle time horizon.  QFS utilizes econometric models, theories of the informational roles of prices and a robust, systematic, disciplined portfolio construction and risk management approach to build risk-controlled portfolios that capture alpha over a 6-24 month time horizon.

We all know that FX is a deep and liquid market. How important is that to the development of your trading models and strategies and what relevance does it have for your draw-down control methodology? 

Jim: Keynes was attributed as saying, “the market can remain irrational longer than you can remain solvent.”  Like the quote says, we recognize that non-fundamental factors can move asset prices in the short run. Drawdown and Scaling Technology is essential to our investment process because it imposes risk discipline required to stick with fundamental themes over the long-term. Drawdown scaling is embedded in our optimal portfolio construction methodology and operates within a real-time computing environment.  To most effectively deploy the technology, we trade assets with deep liquidity and maintain a 24-hour-a-day execution desk.

The Drawdown Control methodology is based on academic research by Dr. Grossman in the early 1990s. The method solves the problem of maximizing an investor’s lifetime wealth of a portfolio, subject to a maximum loss or drawdown limit. The solution to the problem outlines both an accelerating schedule for reducing risk as the portfolio approaches its drawdown limit, and a symmetric schedule for ramping up risk as the portfolio appreciates once again. A significant benefit of this approach is our ability to attempt to financially engineer the mitigation of the left tail in our return distribution. And, in fact, our Currency Program has exhibited characteristics of positive skew. 

Karlheinz: Our approach to forming expected returns is bottom up and fundamental, implemented through a sophisticated process and methodology that we’ve developed through many years. That approach is designed to allow us to move in and out of markets rapidly to capture the changing global opportunity set. Of course, there are times when markets aren’t driven solely by fundamental information. Our drawdown control methodology drives our process to seek to exit losing positions in a fast and expedient way. The depth and liquidity of futures and FX forwards markets, and option markets on them, typically allow us to do this.

In the past you have talked about countries acting like corporations when it comes to currency trading. How does this impact on your investment thinking? 

In currency trading we look at countries like corporations. The central bank is equivalent to a company’s treasury department, the dividend is like the short term yield and the currency is like a company’s stock. Many of the factors you would use to analyse a company, we use to analyse a country. We believe that over time money will flow from countries with low marginal productivity of capital to countries with high marginal productivity of capital just as it does with corporations. Sometimes this coincides with GDP growth (slow to high), but not necessarily. Money will flow from countries that are imprudent with their currency management to countries that are prudent with their currency management. You can think of it as exploiting the dispersion of global business and monetary cycles: We “borrow” (sell) from countries where growth prospects are poor, short-term interest rates are low, and monetary policy is or will be loose. At the same time, we “lend” (buy) to countries where growth prospects are good, short-term interest rates are relatively high, and monetary policy is or will be tight. We use deep economic fundamental analysis to make a judgment call about countries and markets we want to be in. All our positions are long/short; long one currency and short another.

QFS takes a systematic approach to investing. How do you apply this to your model development? 

Jim: Our model development process tries to accomplish two principal objectives.  First, we specialize in understanding the investment opportunities created by macroeconomic, financial and policy events.  So we start with a fundamental hypothesis.  Elaborating investment hypotheses as measurable, falsifiable statistical equations is the next step, and coding and statistical testing of the idea completes the process.  We reject the large majority of the ideas we test, but for those that hold up, we implement them as expected return equations.

The second objective we try to accomplish is to implement our conclusions about fundamental investing in as consistent, disciplined and objective way possible.  Armed with explicit equations expressing our views about expected asset returns, we use technology to automate the implementation of our investment ideas. Each Investment Program at QFS uses a fully integrated system consisting of several computing modules (subsystems).  The major component systems focus on expected return calculations, risk assessments (variance/co-variance matrix), a dynamic portfolio optimization approach (find the trading plan that maximizes expected return subject to risk levels) and a drawdown control system designed to reduce risk (positions) subject to the level of drawdown.  All elements – expected return forecasts, correlations/volatilities, portfolio optimization and drawdown control – operate in real time and incorporate live data feeds and continually readjust the positions suggested by the model as influenced by changing conditions.

QFS Asset Management - taking a unique view on the business of managing money

How do you go about back-testing your proprietary currency trading systems and signals to confirm that your strategies will perform as required and will remain relevant to the long term trading goals and performance criteria of the firm? 

Jim: Once a quarter on every model, the QFS research team runs statistical tests to review the models’ statistical consistency.  

Karlheinz: QFS’s trading systems constitute an integrated network of programs for research, statistical analysis, testing and simulation, data collection, portfolio optimization, risk management and accounting. QFS’s trading systems operate in real time, integrating information from real-time data feeds, performing portfolio valuations, optimizations and associated trades.

In this edition of e-Forex we have examined the development of Machine Readable News and how the next generation of FX algorithms are now being designed to more effectively capture the power of the news. How does QFS go about exploiting real-time insights that are embedded within market and economic data? 

Jim: The area of machine-readable news is fascinating and is something we are following.  Historically our focus has principally been on modelling macroeconomic fundamentals with a fairly long time horizon, again, with models based on fundamental hypotheses.  Nevertheless, our real-time trading system’s optimization design naturally lends itself to aggregation with higher frequency signals and there are theories of market micro-structure that might combine well at the frequencies machine-readable news work for. 

To what extent have you developed research agendas and analytical programmes to help improve the design of new trading strategies and to maintain the performance of your existing investment processes? 

The Investment Committee oversees the operation and performance of the trading systems.  The Committee draws on information from markets directly, analysis from the research group, and feedback from the execution desk.  At least quarterly, the Committee makes decisions to modify parameters governing the trading system (e.g. tolerance to portfolio risk), makes decisions governing non-trading system related factors (e.g., counterparty selection), or manages contingencies unforeseen that prevent the trading system from generating executable signals (e.g., emergency closure of markets).

To support a consistent, disciplined application of each trading strategy, investment decisions are made by the models.  Members of management and trading monitor the accounts during the trading day to ensure that exposures and risk guidelines are being maintained.  Ongoing improvements to the models are made under the direction of our research team of five PhDs (including Dr. Grossman and Dr. Conklin). In addition, a four-person technology team assists in the implementation of the research effort.  A staff of five traders, led by a Director of Trading, executes the trades that are indicated by the model (trading occurs 24 hours a day, five days a week).  The trade execution group has authority for overseeing the implementation of the trades and has limited discretion related to the timing of execution

How much reliance does QFS place on the latest automated trading tool-sets and execution algorithms to help you better manage risk, optimize trade execution pathways and meet your investment objectives? 

Jim: Our proprietary trading system incorporates dynamic programming techniques to optimally enter/exit trades while minimizing long-run transaction costs in real-time.  Once signaled by the model, trades are executed by the trading desk on a best-execution basis, and generally are completed within a given trading session.  The trading desk re-evaluates the trading models at frequent intervals during the day.

What electronic trading platforms does QFS find most appropriate to use and what factors influenced your choice? 

We chose to develop a proprietary trading system in-house nearly two decades ago.  Our approach to risk control, specifically our implementation of draw-down control in real time, requires a custom approach.

How did you go about building your trading desk IT infrastructure and was the trading software and connectivity technology provided by third party vendors? 

Our trading system infrastructure has naturally evolved over the last two decades.  Our proprietary C++ framework and programs were all developed in house by the firm’s technology and research staff in consultation with our traders.  This ensures a deep understanding of our trading system’s internals, which we view as one of our core competencies.  We have traditionally not relied on 3rd party vendors in this area unless source code was part of the delivery, and even then only for things like user interface tools. 

In order to maintain an object-oriented infrastructure that can adapt to change over the long term, we have constantly reevaluated the architecture and determined when code reuse, refactoring, or reengineering various modules was the right choice for the problem at hand.  Some elements of the code date back nearly twenty years, but the overall system is constantly growing and changing, and is now nearly a million lines of C++ code just for the production code base.  Data and data management are obviously a critical part of our technology infrastructure as well, and our programmers are adept with the APIs of the leading market data vendors, including Bloomberg and Thomson Reuters.

Technology infrastructure also helps to manage operational risk.  We endeavor to design high availability into all of our critical business processes, and hardware and software solutions play a critical role there.  Technology can eliminate single points of failure, and facilitates real time replication to our backup site.  We maintain and regularly test a business continuity plan that is robust and very detailed, and we have backup sites in Stamford, CT and Wappingers Falls, NY which together can quickly take on the primary site role.

What new strategies and products has QFS been exploring as part of your continuing efforts to widen the range of currency investment solutions and customised services you make available to clients? 

For the past few quarters our primary emphasis has been in developing FX volatility trading models.  Beyond that, modifications to some currency forecasts incorporating new data sources and the addition of a few new currency pairs are items that are also important initiatives.  Again, because our expected returns are based on economy-specific implementations of broad fundamental trading theses, adding new assets does take considerable time and care.

Looking ahead, where do you see the main challenges facing QFS as you seek new ways for capturing and exploiting investment opportunities with currencies whilst maintaining the key strengths and operational advantages of your investment and risk management services? 

Jim: The firm has an academic environment, meaning researchers have plenty of time and quiet to focus on investment problems.  Our ideas come from our own model and our quarterly statistical reviews of our models, changes in markets and shifts in global policies, and reading papers and through internal discussions. We think about the global economic environment. We study the world economies and determine what is working and what isn’t. Are we missing any risks or new opportunities out there? 

A constant challenge is to balance very different mentalities required to manage quality research.  One mentality must focus on details and maintain a precise, accurate process for insuring the accuracy of new research and the error-free production of models currently invested.  Another mentality requires the researcher to follow current events, think creatively about new ways to quantify and express market and economic processes in abstract ways, and think of how to combine data to test new approaches.  The kind of stimulus and environment needed to derive mathematical formulae, check data, review computer code, and keep abreast of markets and events can be very different.  Finding individuals who have the talent, training and enthusiasm for such a broad range of endeavours is especially challenging.