Do quants dream of electric sheep?

Despite a long history and general familiarity, a surprising number of investors still imagine quantitative investing as mysterious or a “black box,” driven by computers with little human oversight. We at MDT Advisers take every opportunity to rebut this view as we think that the better quant investing is understood, the more its potential advantages become clear. 

What’s in a name?

Yes, quant investing is data-driven, as the name suggests. And, yes, by relying on empirical data instead of subjective judgment, we aim to eliminate emotional biases and harness the power of predictive analytics. But it doesn’t mean computers make all the decisions. In reality, our approach makes significant use of human judgment.

We’ve always prioritized transparency and accountability in our process. Because the process was designed internally, we control the input data and choose how to analyze it. We have a clear understanding of the stock selection and portfolio construction process — not a black box.

For example, our use of “decision trees” is a technique rooted in regression tree models that are now standard tools in industries, such as insurance and the physical sciences. We apply them to evaluate stocks using a series of yes/no questions based on fundamental criteria (like financial statements and valuations) and sentiment indicators (such as earnings momentum and price trends). This helps us generate return forecasts for each stock in a structured, data-informed way.

One of the key benefits of our quantitative approach is its ability to identify a broad range of companies for potential investment while maintaining a high level of transparency. We strive to make it easy for investors to understand why certain companies score well and others do not.

New models, new data

Another common misconception about quant investing is that models are static. That is not true at MDT. Ours are dynamic and can be updated regularly with new data to reflect the latest market developments. For instance, we recently enhanced our stock return forecasting model by incorporating the analysis of economic moats — advantages that bolster a company’s ability to combat competitive forces — across US industries. These qualities, such as strong brand identity or robust patent portfolios, can give a company a sustainable competitive advantage.

Our research showed that incorporating an analysis of economic moats helps to identify buying opportunities among companies that, despite solid fundamentals, have fallen out of favor with investors. These barriers to entry can help companies fend off competition, potentially leading to a healthy rebound in operating performance and share price. Incorporating this analysis may also help avoid companies whose stock price might continue to fall.

Man and machine, not man versus machine

To us, quant investing isn’t just about algorithms — it involves a significant human element. We closely monitor every aspect of our investment process. Our system generates a list of trades designed to optimize each portfolio daily, but we don’t execute these trades mindlessly. We conduct a thorough review and override trade recommendations when we believe the models aren’t capturing all relevant factors. This pre-trade vetting process helps us to understand the rationale behind every trade.

Quant investing is part of a broader movement toward data-driven decision-making in business and finance. We believe the best outcomes come from a hybrid approach, in which potent technology is paired with thoughtful human oversight and continuous refinement.

Learn more about MDT.