In a conversation between Jack Forehand and Matt Zeigler about factor investing within the realm of financial planning, an important insight emerged: while quantitative investing models are hailed for their systematic and mechanical nature, it’s important to recognize that they are crafted by humans. Each model development process involves a series of human-based decisions that ultimately shape the strategies investors rely on.
Here is how Jack explains it (press the play button in the video below):
Although this article can’t delve into every decision, it aims to shed light on the discretion exercised in constructing these models and formulating systematic investment strategies.
Quantitative investment models often originate from methodologies documented in books or academic papers. For instance, here at Validea, we base the majority of our strategies on the methods of great investors and other proven strategies that are outlined in published writings. However, the interpretation of these methodologies can vary. Different minds might emphasize specific criteria or apply nuanced judgments, resulting in variations of the same quantitative strategy. Even a simple factor like value can be measured in many different ways, including a variety of individual metrics or a composite of them. Thus, even at the inception of a model, the human factor becomes evident.
The Underlying Dataset
Selecting the right dataset forms another decision point. While Validea uses a specific data source, other quantitative managers might prefer Bloomberg, S&P Compustat, or a series of other fundamental or alternative providers. This seemingly technical choice involves human judgment, as the accuracy, coverage, and relevance of data can vary. Such decisions about data sources directly influence the robustness and reliability of a quant strategy. We spent considerable time and effort with quality assurance checks on the data, but these are checks there were developed over time and by a human and are implemented in an automated way day in and day out.
Choosing the Investment Universe
Deciding which stocks to include in the investment universe introduces yet another element of human discretion. The quant professional needs to determine whether to focus on Large, Medium, or Small Caps. This choice can impact risk exposure, potential returns, and overall strategy performance, reflecting the judgment calls that underpin systematic investing. For instance, at Validea, our investable universe is roughly 3,000 names, and size and market liquidity are key factors in determining that universe. If we were a $5 billion active stock manager, our universe would be very, very different and smaller.
The size of the portfolio is not a mere technicality. Some quant strategies involve purchasing a large basket of stocks, while others (like our models) opt for a more concentrated approach. The human quant professional must decide whether to diversify broadly or concentrate on a select few, considering factors such as risk tolerance, resource allocation, and investment objectives.
Determining how to weight individual stocks in a portfolio is another important decision. Choices range from market capitalization-based weighting to equal weighting or even fundamentally driven weights. This decision shapes the portfolio’s exposure to different factors, sectors, and risk levels. So at the outset this decision needs to be determined. Of course, as a quant you could test the optimal weighting scheme, but that testing process involves human judgement. For instance, if you market cap weight your portfolio you are likely to be tilting the portfolio toward expensive stocks whereas an equally weighted portfolio naturally avoids this bias in the portfolio.
Start With the Qualitative First
The world of quantitative investing, while lauded for its systematic approach, is underpinned by a multitude of human decisions. From interpreting methodologies and selecting datasets to managing risk, the human quant professional exercises judgment at every step and all along the investment process. While quant portfolios certainly involve significantly less decision making than discretionary ones, they are far from free of human input. Next time you hear about a quantitative strategy, you might be best served by asking about the qualitative side of it first – who built it, what are some of the decisions that go into it – just to make sure those building it are the ones you want managing your portfolio over the long-term.