[It is interesting how innovation can emerge from the simple act of mixing disparate elements together in a novel way. It is in the blending of the best parts of different ideas and perspectives that true alchemy can happen to create a brand-new solution. This deliberate act of challenging convention, of seeing problems from other points of view, and mixing and matching elements in creative ways, has been an important part of the process leading to meaningful change no matter what industry or area of endeavor you may be focused on—even asset management.
To explore this innovation process and how it is being creatively applied to investment management by progressive asset managers, we were introduced to Brin Cunningham, CEO of ECM Strategies and manager of the Evolution Multi-Strategy Portfolio – a short-term trading strategy built around human nature and algorithms. We were interested in asking questions to better understand the mindset and process behind the creation of new, more dynamic investment approaches that are trying to evolve the asset management industry from traditional investment strategies.]
Bill Hortz: Can you please explain your investment philosophy and why the word “evolution” seems to be a key component of your thinking?
Brin Cunningham: Our investment philosophy comes from the blending of behavioral finance with statistical modeling and machine learning.
We believe that deep-seated behavioral biases are not only prevalent but predictable, especially as they pertain to financial markets. Markets would be highly efficient if investors were rational when it comes to maximizing their wealth and well-being. However, investors often behave irrationally which leads to anomalies or deviations from what conventional theory would suggest. Certain anomalies are easy to identify and explain but hard for the typical investor to predict with regularity.
Investors are tasked with deciphering an endless stream of financial data. Unfortunately, humans are not adept at quickly discerning the optimal action given a complex set of facts and, consequently in these situations, they fall back on more innate instincts like fear and greed. Statistical analysis coupled with disciplined rules provide us a framework to exploit this irrational human behavior.
We further acknowledge that every individual strategy has market environments when they perform well or poorly. Therefore, a multi-strategy portfolio is the best way to minimize cyclicality and thus provide more consistent returns.
We also believe that the evolution of an investment strategy is key to long-term success. Evolution requires continuous learning and the more we know, the more opportunities we have for success. Learning also requires the ability to ask the hard questions both of current and future strategies. Market environments shift and change so it is important to keep evolving our solutions.
Hortz: Were there any key experiences you had that influenced the development of your strategy?
Cunningham: I have spent time both as an allocator to managers and managing money directly. This has given me the unique perspective of what it takes to be a successful investment manager and the best practices utilized by the top firms. A common thread among the best firms is a keen understanding of why the strategy should work. In other words, a cause and effect. Not only why the strategy has worked but why it will work in the future. What factors influence results both positively and negatively. It sounds simple but it is not as common as you might think.
Hortz: What specific behavioral finance tenets are built into your portfolio construction?
Cunningham: Since we run a multi-strategy portfolio, there are several behaviors that we attempt to exploit. These include things like fear and greed, herding, overreaction, and recall bias. Recall bias, for example, causes investors to overestimate the likelihood of events they can easily recall, and underestimate events that are more obscure, complex, or unforeseeable. In the context of our portfolio, the recall bias makes it very difficult for investors to price portfolio insurance correctly. They either overprice it when they can easily recall risky events, or they underprice it when they can easily recall positive events.
Research in behavioral economics provides a theoretical foundation for why market participants may overreact. In his book “Thinking, Fast and Slow,” Daniel Kahneman describes how rapid decision making, causes people to resort to “fast thinking.” The fast thinking system is instinctive and emotional, while the slow thinking system is more deliberate and logical. During turbulent periods, people make “fast” decisions, relying on biased heuristics and miss-calibrated probabilities. This leads to blatant mistakes in logic.
We utilize behavioral biases as the foundation for our underlying strategies. Investor biases lead to supply and demand imbalances, which are discovered from the statistical signature of a specific index or security. This allows us to create algorithms that predict when a bias is occurring.
Hortz: On the investment analytics side of the equation, what is your stance on how many parameters or factors should be used in your models and how do you determine the right ones to focus on?
Cunningham: Too many parameters result in too many degrees of freedom and leads to overfitting. Overfitting is typically the cause of models “breaking.” Models do not break per se, but there are developers who do not do a good job upfront in creating the model and utilizing flexible parameters. As a result, the historical numbers look great but there is a low probability they will be repeatable.
We choose parameters and build models based on the behavioral bias we are trying to exploit with our experience using various algorithms. That does not mean we arrive at the correct conclusion right away. In fact, there is often a great deal of trial and error both with the algorithms and the parameters. The key is we want the model to work with a wide range within each parameter to avoid overfitting.