Nevertheless, the 56.4% of respondents using AI/ML suggest weve passed the half-way point in the race to digitize alternative investment processes.
A breakdown of the respondents employers shows that just over half work for CTAs, and six out of ten of them use AI/ML. Thats hardly a surprise, given that automated CTAs are the predominant managed-futures model (the Barclay Fund Flow Indicator, for instance, estimates that systematic CTAs hold eight times more assets than their discretionary counterparts).AI/ML as trading ideas generator and portfolio optimization
AI and machine learning let traders and portfolio managers scan immense data sets and act on data that informs a desired outcome. The algorithms optimize positive outcomes and downgrade adverse results automatically, essentially teaching themselves what works and what doesnt.
The hedge fund pros which were surveyed are not turning everything over to the algorithms. Instead, theyre using them to formulate investment ideas and build portfolios informed by data analysis that the human brain could never hope to accomplish. Just over a quarter use AI/ML to execute trades, suggesting a general reluctance to let the machines pull the trigger.
The survey respondents are giving the machines their due. More than one-half say AI/ML guides 20-60% of their decision-making, and just under one-fifth say it accounts for 80-100%. The 43% average reinforces the proposition that the people are still running things, but they rely deeply on advanced automation and data-analysis tools.Smaller funds enjoy more AI/ML applications
Hedge fund pros have always enjoyed access to the most advanced analytical tools. Indeed, more than half of our respondents have been using AI three years or more, and a plurality have used them for more than five years.
These results pose an intriguing question: Have the hedge fund quants who have used AI/ML for years become set in their waysand thus vulnerable to disruptive fin-tech newcomers?
More than two-thirds of our survey respondents use AI/ML on assets of less than $50 million, suggesting a disinclination to make big bets on the bots. Another possibility is that smaller funds are better able to execute their trades without giving away their techniques and strategies to competing funds.