A report detailing the methodology, however, says the automated process had difficulty distinguishing between gold, silver and bronze ratings, compared to live analysts.
But the system seems to make the important calls correctly, the report said. “The differences between Negative and Neutral, or Neutral and 'Recommended' [a rating of gold, silver or bronze] appear to be quite readily captured by the model,” the report says.
Morningstar’s back-testing found that funds rated negative by analysts are only recommended 4.4 percent of the time by the quant system, while funds recommended by analysts are rated negative by the quant rating just 0.9 percent of the time.
Investors can improve their results if they use quantitative ratings together with the “star” system, Strauts said. For example, sticking with funds recommended by analysts (and now a computer) that also have four or five stars for past performance.
However, a rare sweet spot can be found if a recommended fund has just one star. These funds have produced superior alpha for the next 36 months (based on results from 2003 through 2016.)
Such unusual gems are otherwise “very good funds that have run through a period of very terrible performance,” Strauts said.