In the long run, robot-analysts appear to make more profitable stock recommendations than human research analysts, according to a study from the University of Indiana.
The study’s researchers examined 76,568 reports issued by seven robot-analyst companies between 2003 and 2018 and published their preliminary findings back in January.
They found that portfolios based on the buy recommendations of robot-analysts outperformed those of human analysts, suggesting robot-analysts’ buy recommendations are more profitable.
What’s more, they found that robot-analysts collectively produce a more balanced distribution of buy, hold, and sell recommendations than do human analysts, and are less likely to rely on companies’ periodic earnings reports when carrying out their analysis.
In contrast to human analysts, robot-analysts typically work by poring over the reams of data released in firms’ annual reports.
Robo-analysts should also be distinguished from their close relatives, robot-advisers.
Financial analysts worried about losing their jobs should resist alarmism, though.
For one thing, robot-analysts don’t appear to outperform human analysts when it comes to the other side of the equation: sell recommendations. The Indiana study found “no evidence” to indicate that robot-analysts’ sell recommendations are incrementally more profitable than those of human analysts.
More pertinently, it’s debatable how seriously investors take recommendations made by robots – even if, as the Indiana study suggests, they shouldn’t be dismissed out of hand.
To its credit, the study also examined how far investors actually respond to robot-analysts – and found that investors do not appear to incorporate and trade on the signals provided by their recommendation revisions.
What’s more, it’s still far from clear that robot-advisers – robot-analysts’ cousins – will become a dominant force on Wall Street.