The measure of a man has been variously defined as what he does with power or where he stands at times of challenge, but the measure of a manager is undoubtedly whether he or she can juggle.
Specifically, can fund managers run different portfolios successfully at the same time? This is the subject of a new paper by Ilhan Demiralp and Chitru Fernando of the Price College of Business at the University of Oklahoma, which importantly focuses on managers rather than funds.
‘Using the manager instead of the fund as the unit of observation also allows us to avoid two potential measurement errors associated with the latter approach,’ they explained.
First, looking only at funds can neglect to allow for manager changes. And second, funds in isolation offer only a small sample size – one that can be abused by marketing types.
‘If a fund generates positive abnormal returns and those returns persist, then one concludes that the fund manager has skill,’ Demiralp and Fernando said. ‘But what if this manager simultaneously manages other funds, one or more of which generate negative abnormal returns? Under this scenario, as our study illustrates, it is erroneous to argue for the existence of managerial skill, because the observed outperformance in one fund may be due to luck rather than ability.’
Balls in the air
Drawing on the Center for Research in Security Prices Survivor-Bias-Free US Mutual Fund Database from January 1992 to June 2014 – eliminating index trackers, funds with less than $15 million of net assets, and team-managed products to concentrate only on individual performance – the academics identified 8,394 single-manager strategies run by 4,719 named persons.
First, Demiralp and Fernando established that portfolios headed by the same manager tended to offer similar relative performance within each fund’s own peer group. That is, if a manager had one top-decile fund, their other products were skewed to the higher deciles too rather than randomly distributed. ‘This implies that ex post performances of managers are not entirely due to luck, but managers possess some skill or some managers are skilled while others are not,’ they summarized.
Furthermore, the researchers demonstrated that this was not solely attributable to managers running facsimile portfolios in their various funds: the pattern of performance consistency held when accounting for portfolio similarity and even for strategies in different asset classes and sectors.
So far, this indicated that managers steered their funds with at least a baseline of skill, whether good or bad; the performance of their various strategies appeared not to be determined by luck alone. The next step was to enquire whether managers who ran suites of highly performing funds were likely to maintain that achievement.
Go for the juggler
Demiralp and Fernando sorted managers by the dispersion of their funds’ performance, and traced whether it remained constant or varied from one quarter to 24 quarters thereafter. The standard deviations of the managers’ performance ranking increased ‘almost monotonically’ in the decile scores. In other words, managers with funds clustered in the lower-performance deciles stayed there for the next six years, while higher performers were persistently strong too.
In total, 88.3% of managers with high-decile performance and low variance across their different funds’ decile performance in a given quarter exhibited those same characteristics in the subsequent quarter; 72.5% kept up that profile of consistently high-decile performance over the following four quarters. Likewise, 88.7% of the managers with low performance and high dispersion across their portfolios were stuck in the same situation in the next quarter.
‘On average, good or poor manager performance is not due simply to chance or idiosyncratic events but rather is caused by factors that tend to persist, such as managerial skill,’ Demiralp and Fernando concluded.
So if a manager can run several funds to a high level at the same time, that is a sign of skill rather than luck. And, more importantly, that skill is likely to persist.