The new head of machine learning at AQR Capital Management was just recognized as the 2019 Quant of the Year by The Journal of Portfolio Management. A famed researcher, author and adjunct professor, Marcos López de Prado founded Guggenheim Partners’ quantitative investment strategies business and headed up Tudor Investment’s global quant research team before joining AQR in August. He also has two PhDs and completed his post-doctoral research at Harvard and Cornell. Needless to say, the award seems more than justified. But de Prado has also published some interesting theories on the proper role of quants at machine learning funds that may or may not endear him to investment-focused engineers or old-school traders.
‘Machine learning does not fail, researchers fail’
In a recent paper partly based on his book, “Advances in Financial Machine Learning,” de Prado argues that success in quantitative finance is rather rare due to funds making the same fundamental mistakes over and over again. They put quants in a position to fail by squeezing them into a siloed approach designed for traditional portfolio managers, where trading desks develop individual strategies for the sake of diversification, he wrote, all while throwing some shade at PMs. “Because nobody fully understands the logic behind their bets, they can hardly work as a team and develop deeper insights beyond the initial intuition.”
However, this siloed, me-focused approach typically backfires with quants due to the sheer complexity of developing algorithmic trading strategies, similar to that of building a car from scratch. “One week you need to be a master welder, another week an electrician, another week a mechanical engineer, another week a painter, ... try, fail and circle back to welding. It is a futile endeavor,” said de Prado. The end result is usually a “frantic” search for investment opportunities, with quants eventually settling for false positives that only appear strong or “overcrowded” strategies with shared and underwhelming outcomes.
So what’s the answer for hedge funds? To make their quants cogs in a systematic machine. “Your firm must set up a research factory – where tasks of the assembly line are clearly divided into subtasks, where the role of each quant is to specialize in a particular subtask,” he wrote.
If de Prado is correct – and his reputation suggests he likely may be – the idea of “front-office” quant traders with the autonomy to work in virtual isolation, like traditional PMs, may in fact be a fallacy. Despite a team-focused mantra, this reality would arguably put a cap on what quants can do and how much money they can earn – something many PhDs already grumble about.
One former quant recently argued in an op-ed that systematic traders “have to toil in the trenches for relatively small pickings” because, as engineers by trade, they are still “considered dispensable and easy to replace,” no matter their ability to generate revenue. "A quant can prove that their trading strategy is adding value, whilst there is increasing evidence that the human trader who uses gut-instinct has just been lucky," according to lifelong quant Robert Carver.
De Prado actually seems to agree with both points, though his opinion that quants should act as “assembly line” workers in a “research factory” may not help their case as individuals who are looking to strike it rich after years of schooling.
Have a confidential story, tip, or comment you’d like to share? Contact: firstname.lastname@example.org Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by actual human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. Eventually it will – unless it’s offensive or libelous (in which case it won’t).