NEW YORK - Stock and bond pickers have been losing ground for years to low-cost funds that simply track indexes like the S&P 500 or factors such as low volatility. Now they face a new threat: robots. In 2019, logic favors a bigger role for the machines.

Technology has been creeping into investing for decades, with quantitative hedge funds like James Simons’ Renaissance Technologies, D.E. Shaw and Two Sigma leading the way. By recruiting hundreds of math and physics Ph.D.s and employing powerful computers and high-speed data lines, they exploit anomalies between individual securities or follow market momentum to deliver consistent returns.

Now they and new entrants have taken things up a notch by combining machine learning, a technology used in self-driving vehicles in which computers improve their ability to recognize patterns, with alternative data sources. These include satellite imagery and natural language processing of company earnings calls. The aim is to go beyond fleeting intraday arbitrage opportunities to profitable trades than can last longer.

BlackRock BLK.N gave this strategy a big endorsement in 2017 when it announced it would lay off a handful of portfolio managers in its active equity business and use AI to help make investment decisions. Hedge fund AQR recently hired prominent data scientist Marcos Lopez de Prado to spearhead its machine-learning efforts. Fidelity Investments has quietly built up a team of nearly 140 data geeks.

The transition isn’t easy or foolproof, and no one expects computers to fully replace portfolio managers. Jeff Shen, BlackRock’s head of systematic active equities, says it’s difficult to detect market signals among the noise of more than 200 data sources the firm tracks. Lopez de Prado has written that many firms mistakenly force their information analysts to compete with one another rather than collaborate, or find false patterns with algorithms.

Humans already struggle to prove their worth picking big stocks. That’s one reason why investors pulled $160 billion out of actively managed U.S. equity funds in the 12 months to the end of October while pouring over $170 billion into index-linked and exchange-traded strategies, according to Morningstar. It’s time to welcome the AI-powered investment overlords.

On Twitter https://twitter.com/tombuerkle

- This is a Breakingviews prediction for 2019. To see more of our predictions, click https://reut.rs/2R6H5pG

- For previous columns by the author, Reuters customers can click on BUERKLE/

- SIGN UP FOR BREAKINGVIEWS EMAIL ALERTS http://bit.ly/BVsubscribe

 

(Editing by Jeffrey Goldfarb and Martin Langfield) ((thomas.buerkle@thomsonreuters.com; Reuters Messaging: thomas.buerkle.thomsonreuters.com@reuters.net))