Research has shown that active managers produce persistent excess returns on their best ideas — but they squander these returns by relying on overly diversified portfolios. Now, a new start-up is purporting to have a solution to this long-running problem.
Melius Investments, founded by finance veteran Tim Mullaney, is betting that active management can be saved by using well-known artificial intelligence techniques that have improved results in everything from vaccine development to hurricane prediction.
Melius, a registered investment adviser, has developed a set of model investment portfolios and is pitching them to institutional investors, including pensions, and wealth managers as an alternative to both existing active and passive managers. The portfolios leverage a couple of recent innovations, including recent research published by the CFA Institute.
That research found that in active portfolios, alpha-generating investments get swamped by the mediocre ideas, called beta anchors, that are included by managers to reduce risk. Active managers are essentially using a blunt tool — diversification — for risk management, and sacrificing their edge by relying on it.
The start-up is also relying on machine learning techniques, called ensemble methods, that use the top insights of multiple portfolio managers, as opposed to the single-expert model that has long been used by active managers. Ensemble methods combine multiple predictive algorithms into one that is potentially more accurate.
The techniques preserve individual stock or bond pickers’ alpha and then protect investors from concentration risk by using multiple managers, Mullaney said, describing it as an alternative way to diversify portfolios.
Mullaney said the outdated risk tool now used by managers is causing over-diversification and poor returns. As a result, the industry is being restructured as investors abandon active and flood passive funds.
“You’re asking boxers to fight with one arm tied behind their backs when you have half of managers’ portfolios in assets that aren’t designed to generate alpha,” said Mullaney, who held executive roles at BNY Mellon and Post Advisory Group before going out on his own.
[II Deep Dive: It’s Active Management’s Time to Shine. Has It?]
Melius has used technology, called Ensemble Active Management from Turing Technology Associates, to develop multiple portfolios with different investment objectives. Mullaney has been running the portfolios for about a year to generate track records. Each Ensemble portfolio, which includes up to 50 stocks, uses the highest-conviction picks of 12 managers.
Although the time period is short, results have been positive. From May of 2019 until April 30, 2020, a large-cap blend model portfolio returned 10.7 percent net of fees, which is a net excess return over the benchmark of 6.5 percent. A small-cap blend portfolio returned 5.6 percent from June 10, 2019 to April 30, 2020 versus the benchmark’s return of -12.3 percent, amounting to an excess return of 17.9 percent for the ensemble portfolio.
The time period measured included one of the toughest periods ever for small-cap stocks. An Ensemble high-dividend portfolio lost 6.9 percent from June 24, 2019 to April 30, 2020 — but that was still better than the benchmark, which lost 8.8 percent.
“Big institutions hire several managers in each category, so they are familiar with value of using multiple experts. But each expert also brings the beta anchor with them,” Mullaney said. “If they asked each of them to bring their best 20 or 30 names, they would have done better. Some of these portfolios have 200 names in them. You can’t follow that many stocks.”
Investors have been increasingly exploring using concentrated managers, which are more likely to generate excess returns. But when a manager’s style is out of favor, investors can suffer from extreme price moves.
“Take the Janus 20 fund,” he said. “It did well for a while, but when that manager’s approach goes out of style, even if short term, what happens? The investors gets hurt, the consultant gets embarrassed, and the manager gets fired. So that’s why you don’t see highly concentrated portfolios.”
Even though traditional active asset managers have been losing out to passive for a decade or more, few firms are willing to restructure their portfolio management teams — paving the way for disrupters based on new technology.
“At the end of the day, no one would argue that the current generation of active management is failing in its mandate,” said Mullaney. “I would ask the doubters: then what do you do?”