November 16, 2016

Assessing the Importance of Active Share

Widespread 'closet indexing' creates opportunities for truly active managers?
Mike Gallagher CFA, CAIA, Head of Distribution

In 1999, the equity return expectations of major corporate and public defined benefit pension plans reflected the tendency of investors to extrapolate a prevailing trend. Double-digit returns in perpetuity could seem reasonable in the heady era of ‘Dow 36,000’.

More realistic assumptions prevail today. Recent research by the State of Illinois showed that the percentage of US State and Public retirement investment plans assuming total annual returns in excess of 8% fell from 83% in 2001, to 35% in 2014.

Each unit of excess return is more valuable in a lower return environment, so how should investors adjust what they demand from managers hired to help meet their financial goals? A first step is to assess whether ‘active’ managers are living up to their name.

One measure of the extent of a manager’s active bets is ‘active share’, popularised by a 2009 academic paper by Cremers and Petajisto. Calculated as the sum of the over- and under-weight bets taken in a portfolio relative to its benchmark index, divided by two (giving a maximum ‘score’ of 100%), active share has been widely adopted because it:

  • is an intuitive measure that is simple to calculate;
  • provides a useful complement to traditional measures of risk such as tracking error; and
  • helps identify managers or firms who may be managing their career risk, rather than their portfolio, by ‘closet indexing’.

Intriguingly, Cremers and Petajisto’s work also presents evidence that managers who maintain high levels of active share have delivered returns superior to those achieved by their less active peers. Furthermore, a subsequent paper by Cremers, written in conjunction with the economist Ankur Pareek, indicates that the superior performance of high active share managers is amplified for managers who hold positions for long (more than two year) periods.

Clearly, active share is not a ‘cure-all’. An unskilled manager taking large bets can do more damage than an unskilled index-hugger. And the freedom of a manager to differ from a benchmark is proportionate to its breadth. For example, an actively managed portfolio of UK large cap companies benchmarked against the FTSE 100 will typically have lower active share than a US all-cap portfolio benchmarked against the Wilshire 5000, simply due to the fact that the UK index offers fewer companies to choose from. Finally, some critics note the ‘snapshot’ nature of the view provided by active share, which, unlike tracking error, contains no information in respect of the historic behaviour of the portfolio.

However, it remains the case that multiple studies have either extended or, while critically assessing the active share concept, replicated the findings of the initial work by Petajisto et al. that detected a significant pattern of outperformance among high active share managers. Given its additional usefulness as a simple yardstick with which to assess whether a manager is truly ‘active’, Intermede believes the measure earns its place in the toolkit of anyone assessing active managers, whether individually, or as a class.