Batman vs. Superman: Can Active Asset Managers Fight Back Against the Quants?


The following is a guest article written by Fabrice Bouland, CEO and co-founder of Paris-based research portal Alphametry.

Active managers are feeling the heat.  Risk premia strategies, robot-advisors, systematic funds and alternative data usage are disturbing the quiet territory of discretionary investing.

And things might get even uglier. Artificial Intelligence and its sub-components including Machine Learning, Deep Learning, Natural Language Processing are threatening to eradicate any form of human decision-making in investment. It is only a matter of time.

The almighty portfolio manager is dead. Vive les machines!

The critical importance of “process”

Historically financial information was scarce and more difficult to access, creating asymmetry between economical agents. Price discrepancies were numerous and could be spotted by a human brain. Computers arrived and the 21st century has provided a technology revolution, starting in the ‘80s. Since then we’ve seen the Bloomberg terminal and the Internet as the information advantage started to collapse.

This has changed the way we all work and use information. No matter the industry, whether medicine,  aviation or military operations, the amount of complexity is just too overwhelming for a single individual’s memory or attention span. Process must exist, ranging from simple checklists to structured analysis where information is constantly weighted to feed and update different scenarii[1].

MiFID II: where are the emperor’s clothes?

With billions of assets under management, one might expect discretionary asset managers to have a similar sophistication for its investments. Is it the case? Despite rising expertise among active portfolio managers, underperformance in active management is increasingly in the spotlight.  MiFID II has provided a vivid illustration of the discretionary asset management process, such as it is.

Besides freezing the European research market — hopefully temporarily — the first few months of MiFID II unbundling have been informative about one thing: how inputs to the investment decision-making process are really used.   In the wake of having to absorb research costs, European investment firms conducted last minute surveys among portfolio managers to collect their opinions about which research providers were needed.  At best, a consensus was established and the research franchise perimeter left mostly untouched, largely helped by plummeting written research prices pushed by global investment banks.

In less favorable cases for the portfolio manager, cheap bank portals were contracted, and most of the higher priced independent research cut off. In other more extreme cases, smaller investment firms are currently undertaking the courageous exercise of doing without research at all, to find out if it was necessary in the first place!

Regulation has revealed that the discretionary asset management industry does not understand where the value comes from to deliver its business proposition.  If the industry doesn’t know what inputs are essential for the investment process, it has no metrics for measuring the investment decision-making path.   Repeating and scaling good investment ideas is difficult, if not impossible.

Relying on opinions and assumptions coupled with the lack of process limits the industry’s ability to leverage new capabilities.    Without a structured analytical approach, discretionary managers will not benefit from new alpha-generating research like alternative data. Two worlds are now colliding, qualitative and quantitative. Is this a lost battle?

The big opportunity

Advances in technology might suggest that we are approaching a shift where mankind will be totally disconnected from machines intelligence. Reality is subtler.

“All models are wrong, but some are useful” observes the statistician George Box, one of the great statistical minds of the 20th century. Spurious correlations are everywhere, and failing to find a meaningful causal effect is frequent in financial research. Gary Chropuvka, a partner at Goldman Sachs Asset Management’s Quantitative Investment Strategies, explains: “overfitting is the kryptonite of our industry”[2].

This is where our human brain becomes critical, deciding what correlation between a data set and an asset price would make sense to study. IBM’s Watson’s chief engineer David Ferrucci explains that machines may get better at “mimicking human meaning,” and thereby better at predicting human behavior, but “there’s a difference between mimicking and reflecting meaning, and originating meaning”[3].

In a series of insightful articles, Estimize CEO Leigh Drogen lays a practical vision of discretionary managers’ future: to survive, quantify. This means performance, like underperformance, should be reviewed analytically rather than through the lens of an opinionated conversation between portfolio managers and analysts. The future of investing requires a framework where qualitative inputs can be easily quantified and married with any other numerical inputs on a timescale.

But new tools or process remain a delicate and constant trade between the portfolio manager and the C-levels, where the former must produce and the latter measures. Easy to buy, but how easy to implement?

Organizational structure at the centre

Although a threat for management boards of the past century, information digitization is bringing opportunity to leaders of our time.  In the media industry, for example, the decline of print has led to the birth of integrated digital newsroom. Journalists had to reset their working habits from four-night weeks to web-focused day jobs, and to start collaborating with many other very skilled and capable employees.

To lead their investment firms into the future, management teams face similar challenges. They must draft new organizations which fit within a digital world. An organization where the portfolio manager plays center field and no longer libero, and has access to the best possible yet user-friendly software. A multi-talented team to leverage new alternative data and new markets. An investing team where decisions are no longer hierarchical but consensual. A team where financial incentives are aligned between the fundamental analysts, quants, data scientists, computer engineers, risk and portfolio managers.

The super-hero’s investor

Discretionary investing is certainly not dead. Technology is ready to support and greatly enhance the everyday life of the portfolio manager, torn between information consumption and analysis. But management has a much higher challenge for the forthcoming year: that of engineering a definable process for investing.

At the end of the movie in the title of this article, Superman fights alongside Batman. Wonder Woman offers her last-minute support, perhaps as a stark reminder that women are still underutilized minority in investment management.  My advice: if you are serious about adapting your investing process, hire DC Comics as your next consultant.

[1] Richards J. Heuer Jr., Randolph H. Pherson, Structured Analytic Techniques for Intelligence Analysis,

[2] Financial Times, Spurious correlations are kryptonite of Wall St’s AI rush, March 14, 2018

[3] Tetlock, Philip., Superforecasting: The Art and Science of Prediction


About Author

Fabrice Bouland is the CEO and co-founder of Alphametry. Prior to launching Alphametry, he founded and managed for 10 years the institutional broker OTCex-HPC, a global group of 200 employees operating on all asset classes. His first venture in 2000 was the first European web-based marketplace for OTC derivatives, still in used today by the world's largest banks. He previously spent 10 years at BNP Paribas as an equity derivatives trader in France and Germany.

Leave A Reply