The world is changing for quantitative researchers in finance. Not long ago, they actually needed to know something about financial services, these days they may just need to be...quants.
So said Marcos Lopez de Prado, former head of machine learning at hedge fund AQR Capital Management and currently a professor of machine learning at Cornell University. Speaking at this week's AI and Data Science in Trading Conference in London, Lopez de Prado said the nature of quantitative research in finance. While some teams still employ quants with knowledge in finance, there's a movement afoot to start crowdsourcing quant research needs instead.
Under the "old paradigm" investment managers each hired their own siloed teams of researchers with specific finance knowledge, said Lopez de Prado. However, these in-house teams were hard to staff - very few quants combined the"domain knowledge" and quant expertise necessary to tackle finance data sets. “Consequently many complex investment opportunities were not arbitraged or exploited,” said Lopez de Prado, pointing out that only 0.65% of articles in key economic journals contain terms related to artificial intelligence.
Faced with a dearth of finance quants, Lopez de Prado said finance firms have innovated. Instead of hiring their own quant finance teams who understand the market, buy-side firms are increasingly inviting quants of all backgrounds to compete in tournaments aimed at deriving forecasts from large generic-seeming data sets.
"The [tournament] data is obfuscated," said Lopez de Prado. In other words, crowd-sourced quants compete to find patterns in financial data even though they have no idea what the data actually refers to: "You don't know that the data refers to IBM or that a column is the PE ratio." It's all about looking for patterns and the data could just as easily refer to migratory birds as stock tickers. Under this approach, prior finance knowledge is irrelevant.
If this sounds like bad news for quant researchers in the financial services industry, it might well be. Lopez de Prado didn't say so, but the implication is that steady jobs for finance-specific quantitative researchers who commanded a premium by virtue of their domain knowledge are being eroded by an army of cheap generalist quants sitting in their bedrooms and looking for patterns in generic data.
Instead of drawing a salary and receiving a bonus, Lopez de Prado said the new 'bedroom quants' competing in tournaments are driven by the prospect of winning individual prizes for accurate forecasts. "Their reward is the payout for the week," he said, using the example of Numerai, the San Francisco-based crowd sourced hedge fund which works on the basis that 40,000 crowd-sourced data scientists know best. Numerai offers a regular 'bonanza' for competitors with the most accurate predictions.
Funds using the tournament method save many times over, said Lopez de Prado. They don't cut spending on employment costs, but on the cost of hardware and office space too. "More and more people will realize the [old] silo approach is not scalable and that the advantages [of crowdsourcing] are offset by the disadvantages. We are going to see crowdsourcing more prevalent in quantitative research," he predicted. That sounds ominous if you currently have a quant job on the buy-side.
Lopez de Prado said there are three options for quant research (Silos, Platforms and Tournaments) and that one - tournaments - does not presume prior you knowledge. You can see his presentation here.
Photo by Andrew Neel on Unsplash
Have a confidential story, tip, or comment you’d like to share? Contact: firstname.lastname@example.org in the first instance. Whatsapp/Signal/Telegram also available.
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by 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.)