Cornell University: Data on fired managers’ performance may improve investments

You would never drive your car using only your rear-view mirror, but according to Scott Stewart, MBA ’83, Ph.D. ’85, that’s exactly how many professional investors make key investment decisions.

“In my experience as a professional money manager, one of the classic retail mistakes was buying stocks after they went up, then selling them after they went down,” said Stewart, clinical professor of finance and accounting in the Samuel Curtis Johnson Graduate School of Management. “A lot of professional investors make the same mistake: ‘buying’ managers after they outperform the stock market, and selling them after they underperform.”

Stewart is the author of “Performance, Perception and Manager Selection,” published April 1 in The Journal of Portfolio Management. The gist: Professional investors ignore the performance of terminated fund managers – the “non-decisions,” Stewart calls it – when developing confidence in their management strategies.

Stewart proposes that investment plans can improve their selection processes by collecting and studying the performance data of terminated managers, in addition to retained managers, to get a fuller picture of performance.

Before entering academia in 2001 as an assistant professor at the Boston University School of Management, Stewart spent 16 years as a portfolio manager, including 14 at Fidelity Investments, where he founded its Structured Investments Group, with managed assets of $45 billion. He saw many of his corporate clients managing funds “based on looking through the rear-view mirror,” he said.

Research he published in 2009 on this same topic determined that institutional fund managers lost an estimated $170 billion due to poor manager selection decisions between 1985 and 2006.

“In that research project, we wanted to look at two issues,” said Stewart, who returned to Cornell in 2014. “One was, why do institutional money managers – people who run foundations, endowments and pension plans – move money from one manager to another? And two, does this add value?”

The faculty co-director of Cornell’s Parker Center for Investment Research, Stewart said the mistake many managers make is failing to factor in the performance of fired money managers when making investment decisions.

To test this theory, Stewart examined the results of three surveys of investors regarding pension plan decision-making, manager selection activity and plan officers’ perception of success. The first survey was conducted in 2004, and led to his 2009 published research; he used more recent survey results (2014 and 2016) to confirm the validity of the older work.

And so as not to reveal the true intent of the survey (whether they study the performance of terminated managers), respondents were asked how closely they studied performance results in general.

In all three surveys, results indicated that professional investors had high confidence in their manager selections, and that performance was very good once they were hired. But based on Stewart’s analysis of industry asset flows, account changes and managers’ returns, this confidence is misplaced.

Stewart sees two reasons for this disconnect.

“The first is way that information is reported,” he said. “When an endowment, for example, reviews the performance of their managers, they only look at the performance of their current management system. So if a manager goes through a bad period, and you fire them soon after, they’re not on the list in the next quarterly report. The bad information disappears, but there also is good information in that manager’ subsequent record.

“And then there’s the psychology – how people make decisions, and how they look for information that reinforces what they believe,” he said. “The term ‘non-decision’ means we don’t look at the alternatives. In life, we really don’t know what that non-decision is: I earned my Ph.D. at Cornell, but what would have happened if I didn’t do that? Well, who knows? But in this business, you can know because the return information is available.”

The key is making the effort to get that information, he said. As one way to gain a fuller picture, Stewart suggests benchmarking new managers against the performance of terminated managers. He also said that, due to the rise-and-fall nature of financial markets in general, “past studies suggest that deciding to hire managers with poor recent performance, although uncomfortable, may be profitable.”

His final recommendation: “Pause and think before firing a ‘bad’ manager. You may be doing it just at the wrong time.”