In short, not much. Like any attempt to reduce a complex process to a digit, it fails as an investment assessment tool.
“Active Share” is the number you get when you sum the difference between the weight of a stock in a portfolio and the weight of that stock in an appropriate benchmark for all stocks in the benchmark. The maximum score on this basis is 200%. We can simply divide by two or agree to only sum the overweight positions to arrive at a figure where the maximum score is 100%.
The score is an attempt to measure the risk that fund is taking – but note that we are talking about relative risk versus an index rather than absolute risk (the probability of a permanent loss of capital). By using current holdings, the score is not a rear view mirror measure of relative risk in the way that tracking error (the standard deviation of relative returns) is.
It is not a new concept. This analyst well remembers being admonished by his CIO for calculating such a “one dimensional and potentially misleading metric” in 2003. The measure was given academic credentials in 2006 by Martijn Cremers and Antti Petajisto of the Yale School of Management.
Active share or “active money” has received much press of late – a google search gives us 288 million hits. Some in the manager research community have proposed investors should only look at actively managed funds with a high active share score. The assertion is that funds with a low score are effectively index trackers charging active management fees. They cling to the observation that the average excess return from high active share managers is positive and significant while the average for lower active score managers is close to benchmark.
The problem with this analysis is that it takes no account of manager skill. A high active share score means that there is potential for the portfolio to deviate significantly from the path of the benchmark over a short period of time – nothing more. Who would you back: Fangio in a Ford or Bob the Builder in a Bugatti? Surely it is right that if unskilled journeymen are to be allowed to manage a portfolio he or she should be constrained to limit the damage? “Skill” was renamed by Messrs Grinold and Kahn as the information coefficient. If a manager has a positive information coefficient, then Active Share score will simply provide guidance on portfolio construction. The Active Share score cannot tell us whether or not skill exists – this can only be achieved through rigorous, detailed and consistently applied manager research of the qualitative variety.
A succinct and well observed assessment of Active Share has been penned by Rob Harris, CEO of Majedie Asset Management. He reminds us that Active Share is determined by the nature of the benchmark selected – is it concentrated in a small number of large companies for example? If the answer is yes and those companies are fundamentally attractive, would you want your manager to avoid them simply to increase their Active Share score? He draws on his experience as part of a team which has delivered an annualised excess return of 4.2% against the FTSE All-Share Index since July 2003 with an Active Share that has varied between 50%-60%. This level of active money would see a fund placed in the “index huggers to avoid” bucket by myopic manager researchers. This analyst can confirm that Majedie have been consistent as their UK strategy was part of the calculation described as “one dimensional and myopic” in 2003.
The investment group at Cerno Capital is principally concerned with assessing the return potential of investments. When we do select an active manager, our qualitative analysis has determined that they are truly active, regardless of any Active Share score.