Funds labelled as ‘closet trackers’ because their low active shares may not be anything of the sort, according to a study by Fidelity International, which found that index concentration can skew results.
A white paper by Fidelity’s Barney Rowe, Ran Wang, Paras Anand and Tony Gibb considered whether active share was an appropriate tool for measuring how active a fund is.
The popularity of passive strategies in recent years has soared as investors have sought to partake in the extraordinary post-crisis bull run.
The availability of cheap index-tracking funds has put active strategies and their managers under increasing pressure to justify their fees and show the alpha they are able to deliver in rising market conditions.
Source: Investment Association
As such, any fund or manager deemed not offering a substantial level of activeness has been labelled a ‘closet tracker’.
It has also caught the attention of regulators, with the UK Financial Conduct Authority claiming that £100bn was held in ‘partly active funds’ in its Asset Management Market Study of 2016.
However, the use of the active share measure was criticised by some asset managers and the regulator itself conceded that it should not be used as the sole basis to identify partly active funds.
Nevertheless, active share has become well-established and used by investors and fund pickers to make investment decisions based on the value provided by active funds over passive strategies.
The measure was popularised by a 2009 academic paper ‘How active is your manager’, authored by Martijn Cremers and Antti Petajisto.
Cremers and Petajisto quantified activeness by measuring the difference between a fund’s holdings and its benchmark constituents, with a score of 100 per cent suggesting a fund has no holdings in common and one of 0 per cent exactly matching the benchmark.
A ‘closet tracker’ is classified by Cremers and Petajisto as one whose active share is below 60 per cent. Below such a level, it is argued, it become harder to outperform the benchmark net of charges.
However, the Fidelity team said while the 60 per cent cut-off has become a ‘rule of thumb’ for fund pickers and regulators, such a definition might not be appropriate in every circumstance.
As such, the Fidelity team considered whether number of benchmark constituents and index concentration might also be important to consider.
To assess this the analysts built discretionary portfolios to see whether the 60 per cent figure made sense in all situations.
Two characteristics that the team applied to its simulations were benchmark-agnostic managers and the Ucits 5/10/40 rule compliance to avoid simulating closet trackers.
The first characteristic aimed to show whether benchmark-agnostic mangers were to showing low active share. Meanwhile, the 5/10/40 rule limits the maximum holding in securities from a single issuer to 10 per cent of the portfolio and the total of all holdings with single issuer allocations greater than 5 per cent to a no more than 40 per cent of the portfolio.
The Fidelity analysts found typically lower active share figures in more concentrated indices with fewer constituents, which was also true for its benchmark-agnostic portfolios and indeed below the 60 per cent threshold.
The findings also showed that the threshold was frequently approached or crossed in simulations for benchmark indices with fewer than 200 stocks.
In addition, smaller and more concentrated portfolios in its simulation had a greater chance of achieving a higher active share than larger portfolios.
As such, the Fidelity team found that the 60 per cent threshold may be too stringent for smaller, more concentrated benchmarks. The team concluded that the threshold was “somewhat arbitrary” when applied regardless of index structure.
“For portfolio managers that are operating in markets with relatively few constituents, their ability to achieve high active share is impeded, while it is relatively straightforward to achieve high active share in highly diversified markets,” they said.
“We would argue that the threshold should be deflated for more concentrated indices.
“Conversely, for well diversified indices, the bar should be set higher than 60 per cent to reflect the much broader investable universe.” 
They concluded: “We advocate evaluating active share in the context of what is realistically achievable for a given index, while taking relevant additional constraints such as the Ucits 5/10/40 diversification rule into account.”
Jason Hollands (pictured), managing director at Bestinvest, said that index concentration can have a distorting effect on active share and that it is something that investors should consider before making a decision based on one measure.
“Certain indices are incredibly concentrated and for risk reasons you would expect managers to have a very different positioning, there is something in that,” he said. “But any of these metrics have limitations.”