Benchmark construction, market capitalisation, and portfolio risk can all influence a fund’s active share, according to research by Hermes Investment Management—and all three factors can reduce the reliability of the measure.
“Without the presence of skill, a manager that looks like a lion based on their active share might actually manage like a lamb.” —Eoin Murray, Hermes Investment ManagementThe research was published in a paper titled “Looks like a lion, manages like a lamb” by Eoin Murray, head of Hermes’ investment office. Murray warned that active share—a measure designed to highlight by how much a manager’s portfolio deviates from its benchmark—could be heavily influenced by outside factors and was no indicator of skill.
He argued that “without the presence of skill, a manager that looks like a lion based on their active share might actually manage like a lamb.”
“It is instructive to think about the source of a manager’s active share, and whether there is consistency between how it is created and the manager’s investment philosophy and mandate,” Murray added.
The research showed that active share was heavily influenced by the market cap of companies selected by managers. This meant that small cap funds “tend to have a disproportionately high active share”, which skews broader samples of funds. Murray suggested that small cap bias could explain the outperformance of high active share managers in the 2009 study that introduced the measure.
Benchmark construction was also a potential factor, Murray argued. Indices dominated by a smaller number of large stocks would require managers to take a big position in the top companies to be neutral, and so would reduce the ability to take a significant relative active bet.
“This makes comparison of active share across funds a more subtle exercise,” he said.
“There is a danger, caused perhaps by our deeply ingrained behavioural biases that drive us to believe we are more capable of predicting the future than we actually are, that this single metric could be considered to be the measure of expected future performance,” Murray wrote. “That would be a gross oversimplification.”
Murray said a “more complete picture” of skill—and therefore a more accurate indicator of likely future outperformance—could be given by incorporating tracking error.
“We can think of active share as being more indicative of a manager’s propensity to pick stocks whereas tracking error could be a better proxy for their factor exposures,” he wrote.
Murray’s paper backed up research by fellow fund management group AQR. In April, authors Andrea Frazzini, Jacques Friedman, and Lukasz Pomorski claimed a 2009 study by Martijn Cremers and Antti Petajisto was “subject to misinterpretation”.
The academics hit back, telling CIO their research indicated that high active share funds showed outperformance regardless of style.
Related Content: The Problem with Active Share