By: DeFred G. Folts III, Chief Investment Strategist

A Different Perspective on Quantitative Research in Investment Management

Given the evolution of AI and Machine Learning, the ascendency of computer driven, quantitative and black box approaches to investing is growing exponentially in the investment industry.  This phenomenon is often considered in terms of man vs. machine with the question being, which is better?  However, at 3EDGE Asset Management we see this conversation differently.  We do not believe that the argument should be focused on man vs. machine, but rather we believe that the better approach to utilizing quantitative, computer driven analysis comes from the combination of man plus machine.

3EDGE Asset Management, LP (“3EDGE”) is a multi-asset investment firm focused on top down portfolio allocation across asset classes and geographies to create globally diversified portfolios. Our goal is to generate attractive, risk-adjusted investment returns over full market cycles.

We believe in the inter-connectedness of markets, and that the global capital markets constitute a complex, nonlinear, dynamic system. Such systems exist throughout science, engineering, social systems, nature and evolution.  Inherent in these systems are feedback loops and time delays which cause cycles and, at times, extreme behavior.  System behavior is oftentimes uncertain, counterintuitive and sensitive to small changes in inputs and these systems typically fluctuate across various states of disequilibrium.  In market terms, this means that the global capital markets tend to fluctuate between states of overvaluation and undervaluation and are very rarely in equilibrium.

Man + Machine
In our opinion, the best way to analyze a complex system is to integrate quantitative analysis and human judgment.  Our research and investment process does rely on quantitative computer models; however, our approach is not purely systematic and should not be considered a black box.  Though quantitative analysis plays an important role, all model output is further analyzed and confirmed by our investment committee before any investment decisions are undertaken.  This is important since, in a complex system with so many inherent uncertainties, it simply is not possible to include a priori every single variable that could impact a solution.

Therefore, our investment process seeks to combine systematic data analysis with human expertise and judgment -- in other words, man plus machine. The advantage of employing quantitative models and robust computing power in the investment research process is that computer models can provide constant, data-driven insights across a much larger universe of information and it is an analytical tool that is not encumbered by the many well-known human biases which can compromise investment decision making.  However, we also believe in complementing our proprietary quantitative research by applying our experience, expertise and judgement.  Human intervention is seen to be additive to our research and investment process particularly when considering more subtle shifts, and more granular insights into the behavior of the global capital markets than may be possible with only a computer data driven process without human intervention.

“Allowing the human to coach a highly-automated system produces results up to 50 percent better than if the automation were left to its own devices.  Collaboration between humans and computers, particularly in knowledge-based domains where complementary strengths can be leveraged, holds much future potential.” Mary Cummings, IEEE Computer Society, 2014

Inductive + Deductive
Another fundamental difference between our approach to quantitative research and many other traditional quantitative investment models is that our approach is both deductive and inductive.  The fundamental idea behind our approach is to first arrive at the essential theories that best explain the true cause and effect relationships that drive the global capital markets (deductive), and then to test these theories through quantitative modeling (inductive).  However, effectively employing such an approach is only possible after decades of study of the behavior of the global markets across asset classes and market cycles.  More traditional quantitative analysis and black box approaches are oftentimes almost entirely data driven and therefore inductive in nature. Large amounts of historical data are collected and analyzed using regression based modeling techniques and then the model output is evaluated using back-tested results. We believe that this approach can be fraught with risks and is oftentimes unreliable for analyzing the behavior of a complex, nonlinear system like the global markets.  Therefore, 3EDGE has undertaken what we would consider to be the more difficult and time consuming approach of establishing the theory first, and then testing those theories through our proprietary quantitative research models.

At 3EDGE Asset Management, in seeking to generate attractive, risk-adjusted investment returns over full market cycles we combine a systematic process of data analysis with our human experience, expertise and judgement.  This combination is designed to take advantage of the strengths of both man and machine in our investment process.  We do not believe that the argument in the investment industry needs to be focused on man vs. machine, but rather we believe that the better approach to utilizing computer driven quantitative investment analysis comes from the combination of man plus machine. 


DISCLOSURES: The opinions expressed in this commentary are those of DeFred G. Folts III, Chief Investment Strategist of 3EDGE Asset Management, LP.  The information provided in this summary includes information from sources that 3EDGE believes to be reliable, but the accuracy of such information cannot be guaranteed. This commentary does not constitute an offer to buy or sell any security. Investments in securities, including common stocks, fixed income, commodities and ETFs, involve the risk of loss that investors should be prepared to bear. Past performance may not be indicative of future results.