Active Share Efficiency: A Measure Beyond Active Share

Transcription

1 Active Share Efficiency: A Measure Beyond Active Share Introduction Active Share measures the proportion of holdings in an equity portfolio that differ from the strategy s benchmark. When initially researched, Active Share was found to be correlated with managers outperformance, but its application has been expanded, and it is now widely used by investors and managers alike to evaluate and compare active management across strategies. We propose a new measure called Active Share Efficiency, building on Active Share by taking into account managers excess returns. Active Share Efficiency is thus able to assess the effectiveness of managers active positions and provide a standardized comparison across managers at all levels of Active Share. We also propose Peer Share Efficiency. This new statistic extends evestment s Peer Share analytics, calculated for 49 peer groups of equity managers in our database. In an environment where investors are migrating away from manager selection based on traditional style boxes to strategies that are concentrated and unconstrained, we believe Active Share Efficiency and Peer Share Efficiency are important additions to the analytical toolbox for choosing and monitoring equity managers. Definition of Active Share The notion of Active Share was introduced by Yale University academics Martijn Cremers and Antti Petajisto in a 2006 working paper, as a novel measure of the degree of active management in equity portfolios. 1 Cremers and Petajisto posited Active Share as a predictor of excess return, as did William Sharpe in his initial publication on the Sharpe ratio. 2 Although similar statistics have been in use for years by investment managers and consultants, Cremers and Petajisto s 2006 publication formalized the calculation of Active Share, and has made a constructive addition to the investment industry debate on the value of active portfolio management. Active Share compares the weights of individual securities in a portfolio to those in its benchmark (its calculation is illustrated in the Appendix). When the portfolio holds the same securities as the benchmark, at identical weights, there would be no differences in positions and Active Share would equal zero. Similarly, a portfolio that held none of the securities in its benchmark would have an Active Share of 100. In practice, for most actively managed portfolios Active Share falls between 60 and 90, depending on the manager s process and benchmark. Exhibit 1 illustrates two simplified hypothetical portfolios: On the left, the portfolio s positions (shown by the light blue bars) are held at weights similar to 1 K.J. Martijn Cremers and Antti Petajisto, How Active Is Your Portfolio Manager? A New Measure That Predicts Performance, Available at papers.ssrn.com/sol3/papers.cfm?abstract_id=891719, accessed Aug. 12, William F. Sharpe, Mutual Fund Performance, The Journal of Business Vol. 39 (1, Part 2: Supplement on Security Prices), Jan. 1966, pp Available at accessed Aug. 11,

2 those in its benchmark (the dark blue bars), resulting in a low Active Share. In the example on the right, positions vary widely from their benchmark weights; additionally, this portfolio holds a large proportion of assets in non-benchmark securities and cash, resulting in a high Active Share. Exhibit 1: Illustrations of Portfolio Active Share For years, the most widely used measure of a portfolio s degree of active management has been tracking error, which compares the volatility of a portfolio s active returns (or historical returns relative to its benchmark). Active Share looks at portfolio management from a different angle, by assessing the manager s construction of the portfolio rather than the returns it has generated. A portfolio with low Active Share may be destined to perform in line with its benchmark, while a high Active Share portfolio has the potential for generating excess returns through many opportunities different from the benchmark. However, actual returns are dependent on the manager s skill in selecting securities (or his good fortune), and not on high or low Active Share alone. As a tool for performance measurement, Active Share is not meant to replace tracking error. In fact, Cremers and Petajisto recommend bringing both statistics to bear, for a more comprehensive assessment of the sources of portfolio return. In their view, tracking error offers a better view of a manager s use of factor timing through the choice of industries and sectors, while Active Share is a better gauge of the value added through stock selection. Still, Active Share has been warmly embraced by many in the investment community as a concept that highlights the strengths of the unconstrained and concentrated investing strategies that have lately gained currency. It is an intuitive measure that is fairly steady over time for a given manager, and describes the causes of excess return stock selection, 2

3 which is the essence of investment managers work. Tracking error, and its companion measure, information ratio, are more about the effects of portfolio management; moreover, tracking error and information ratio can be volatile and are often influenced by market factors outside managers control, making them harder to interpret. Some in the industry are skeptical about the value of Active Share as a predictor, and a number of research notes and papers have been published offering cautions on the interpretation of Active Share. 3 This paper does not join the debate over the forecasting power of Active Share, but instead considers its value as a tool for analyzing portfolio construction, past performance and benchmark selection. Characteristics of Active Share Although Active Share calls for a simple calculation, its interpretation can be complex. The context of market segment, portfolio content, and benchmark choice are essential considerations for understanding how a manager arrives at a particular Active Share measure, and whether a manager can translate higher Active Share into greater excess return. 1. Level of Active Share: There are no fixed boundaries on what constitutes high or low Active Share, as these definitions vary for different segments of the equity market. Exhibit 2 and 3 compare the ranges of Active Share for four evestment manager universes. Large cap core U.S. equity managers tend to show lower Active Share measures than other groups: only 6% of the group had Active Shares above 90, while in small cap core U.S. equities, 68% of managers rated 90 or higher. That difference arises from small cap benchmarks having more constituents, and a flatter distribution of positions, than large cap indexes. Similarly, the nature of the opportunity set results in higher Active Share measures for managers of global large cap core and emerging markets equities as well. As a result, analyses of Active Share across managers should take into account market segments to ensure valid comparisons. Exhibit 2: Active Share Distribution for Equity Manager Universes, Two Years Ended March For example, Tim Cohen, et al., Active Share: A Misunderstood Measure in Manager Selection, Fidelity Investments, Feb (available at www. fidelity.com/bin-public/060_www_fidelity_com/documents/leadership-series_active-share.pdf, accessed Aug. 11, 2014); and Scott Wittman and Vinod Chandrashekaran, Debunking Active Share, Investment Viewpoints, American Century Investments, Sept (available at institutional. americancentury.com/content/dam/americancentury/institutional/pdfs/viewpt/ debunking-active-share.pdf; accessed Aug. 11, 2014). 3

4 Exhibit 3: Active Share Distribution for Equity Manager Universes, Two Years Ended March Off-benchmark securities: Active Share is influenced by the proportion of non-benchmark securities a manager has selected. One typical example is the strategy known as EAFE plus, which is tied to the MSCI EAFE, but often with a significant portion of emerging market stocks. Such managers can vary their emerging markets holdings widely, with direct effects on Active Share: Holding 20% of a portfolio in off-benchmark emerging market stocks drives Active Share versus the MSCI EAFE higher by 20 points Portfolio cash: Equity benchmarks hold no cash, but active managers do, and in the case of unconstrained strategies these can carry significant amounts of cash when they see few opportunities in the market. All else being equal, a higher cash position brings higher Active Share by the proportion of cash held. 4. Benchmark choice: A manager s level of Active Share is obviously influenced by the benchmark used in the calculation. Exhibit 4 considers the case of managers in the evestment global growth universe. The left panel shows that the distribution of Active Share calculated from the MSCI World index (in light blue) tends to be higher than that based on the MSCI World Growth index (in dark blue) for the same manager group. Similarly, shifting the global growth managers to the MSCI ACWI index yields very high Active Shares, although less so when the metric is calculated against the betterfitting MSCI ACWI Growth index. 4 In the calculation, 20% divided by 2 produces 10 points of Active Share, but by investing 20% in off-benchmark securities, the benchmark is underweighted by 20%, which divided by 2 yields an additional 10 points. Taken together, the allocation increases active share by 20 points. 4

5 Exhibit 4: Effects on Active Share of Changing Benchmarks 5. Active Share and excess return: The most contentious point in the industry s debate on Active Share is the headline that managers with higher Active Share measures tend to earn higher excess returns. Research subsequent to Cremers and Petajisto shows that some, but not all, high Active Share managers outperform in certain market regimes, and that the relationship is not a simple one. In the evestment universe, we have found a strong and consistent relationship between Active Share and tracking error; however, high Active Share is not necessarily associated with high excess return. For both our U.S. large cap core and U.S. small cap core manager groups, for the three years ended March 2014, tracking error and Active Share tended to rise together (the dark blue bars in Exhibit 5, versus the Active Share scale on the left). The relationship between Active Share and excess return was inconsistent (shown by the light blue bars). In U.S. large cap core, the highest excess return for that period accrued to managers with Active Share scores of 90 to 95 and to those with scores of 55 to 65, but not to those in between. For the evestment U.S. small cap core group, excess return rose as Active Share increased from 60 through 90, but dropped off at the highest levels. Moreover, some of the group s highest excess returns were earned by managers with low Active Share. 5

6 Exhibit 5: Active Share Versus Tracking Error and Excess Return for Two evestment Manager Universes, for the Three Years Ending March 2014 Monitoring Managers through Active Share While Active Share may or may not be a reliable predictor of performance, it is certainly a valuable tool for selecting and monitoring individual managers and the combination of managers within portfolios. Active Share is measured at a point in time, typically at quarter end. When Active Share is viewed as a time series, however, an investor or consultant can plot the course of how actively the manager has driven the portfolio in different market environments. Active Share can also be viewed alongside other measures of portfolio construction, such as the proportion in top-ten holdings, to determine that managers are meeting their commitments on concentration and conviction. Similarly, material changes in Active Share can be weighed against the size of a portfolio, to detect issues in liquidity of positions or changes in a manager s philosophy or process. Differing degrees of Active Share can indicate whether a given benchmark is suited to a manager s style. Active Share is a statistic for strategies in the evestment database, and investors can readily calculate a manager s score against a variety of benchmarks. In a portfolio setting, investors can measure the Active Shares of one manager versus another, to ensure that managers portfolio holdings are complementary and do not create a significant overlap. Likewise, the portfolios of several managers can be combined, and the Active Share of the aggregate portfolio calculated against a relevant index, to measure how actively they manage in combination. Individual managers can rate highly in Active Share versus a growth or value benchmark, but show lower Active Share when their strategies are combined and measured against an index that is style-neutral. 6

7 Introducing Active Share Efficiency Cremers and Petajisto have made a valuable addition to the toolbox of investment analysis. However, Active Share s meaning is limited. Although it capably measures the active portion of portfolio construction, a high score on Active Share is not desirable in isolation: It demonstrates a manager s willingness to take active bets, but does not reflect the excess returns the active positions generate. Therefore, we propose an extension to Active Share, which we have named Active Share Efficiency, as a new measure of manager skill. Active Share Efficiency evaluates excess return relative to Active Share, to measure the rewards resulting from a manager s active positions. Active Share Efficiency is therefore similar in nature to information ratio. (The Appendix provides details of calculations.) Active Share Efficiency is a useful measure for monitoring the value added of individual managers through time. It also allows comparisons of efficiency across managers with different degrees of active management, providing a level playing field. This point is critical, because managers ranking lower in Active Share can still be highly efficient, and can turn the Active Share conversation to their advantage by emphasizing their Active Share Efficiency. As an illustration, a portfolio that equally weights the stocks in the S&P 500 has an Active Share of just 44, but has earned returns well above the capitalization-weighted benchmark (0.6% and 3.5% annualized for the three years and five years ended June 30, 2014), and ranks highly in terms of Active Share Efficiency. Just as the ranges of Active Share vary for managers in different segments of the equity market, Active Share Efficiency varies as well. Exhibit 6 compares the Active Share Efficiency scores of managers of U.S. large cap core and U.S. small cap core strategies in the evestment universe for the three years ended March Measures for the bulk of large cap core managers were evenly distributed between minus 3 and 3, while Active Share Efficiency for small cap core managers was skewed to higher scores. These distributions are a result of the market environment for active managers in their respective universes: More U.S. small cap managers were able to add excess returns versus their small cap benchmarks, while U.S. large cap core managers had greater difficulty in adding value. Exhibit 6: Distributions of Managers Active Share Efficiency, U.S. large cap core Versus U.S. small cap core, Three Years Ended March

8 The clustering of Active Share Efficiency scores for managers in different market segments leads us to introduce another new measure: the Peer Share Efficiency score. This refinement groups managers according to their particular styles, market cap ranges, geographies and other portfolio characteristics, and compares a manager s efficiency to the group. As many institutional investors and their consultants currently favor concentrated and unconstrained strategies, we believe evaluating managers results in a focused and relevant context is especially important. Peer Share, a feature in evestment analytics, measures the Active Share of an individual manager versus an evestment peer group rather than against a conventional benchmark, and provides a useful additional measure for manager selection and monitoring. Holdings of managers within the peer group are aggregated, giving each manager equal weight, to form a peer composite portfolio, which is used to measure an Active Share for each manager. Adding in each manager s excess returns derives Peer Share Efficiency, to monitor Active Share Efficiency on the relevant group of managers and holdings. Peer Share Efficiency can be especially valuable, as it compares managers on the opportunity set of securities the peer group actually holds, which can differ widely from conventional cap-weighted or style benchmarks. Conclusion Active Share was first posited as a predictor of active managers excess returns, but has since been adapted as a tool for evaluating portfolio construction as well. However, a meaningful interpretation of a given manager s Active Share requires context: Levels of Active Share vary according to the equity market segment a strategy invests in and the manager s reliance on off-benchmark securities and holdings of cash, as well as the chosen benchmark. Such considerations are especially important in the evaluation of unconstrained and concentrated portfolios, which pursue more active investment styles. Our research on managers in the evestment universes shows that the relationships between Active Share and excess return are not simple or consistent. However, Active Share is nonetheless highly useful in selecting and monitoring managers individually or in combination. We propose new measures, Active Share Efficiency and Peer Share Efficiency, which build on the notion of Active Share. Active Share Efficiency is similar to information ratio, in that it evaluates excess return relative to Active Share, to measure the effectiveness of a manager s approach to active management. Peer Share Efficiency is based on Active Share Efficiency, but is calculated from a peer group s universe of holdings, rather than standard benchmarks. 8

9 Appendix Active Share Active Share is one-half of the sum of the absolute values of differences between the benchmark and portfolio for all positions, expressed as an integer between 0 and 100. Zero Active Share represents a portfolio identical to the benchmark, while an Active Share of 100 indicates that the portfolio and benchmark have no positions in common. The measure is calculated at a point in time, likely on quarter-end positions, and is often presented as a three-year moving average. Active Share = 1 2 N i=1 w fund,i w index,i Where w fund,i and w index,i are the portfolio weights of asset i in the fund and in the index, and the sum is taken over the universe of all assets. Active Share Efficiency Active Share Efficiency is the ratio of a portfolio s excess return for a given quarterly period to its active share at the end of the period. Active Share Efficiency = Re AS where Re is excess return (expressed as an integer) and AS is Active Share (measured as a decimal). Similar to other measures, it can be viewed as an average excess return over multiple time periods (average rolling periods) as to limit endpoint sensitivity. In that case, the formula would change slightly, (where n equals number of periods) to include average excess return divided by average active share: Peer Share Efficiency Active Share Efficiency = n ( Re i i=1 ) ( AS i i=1 ) Peer Share Efficiency is an extension of Active Share Efficiency to evestment s Peer Share statistic. Peer Share measures the Active Share of equity strategies against their evestment Peer Alpha universe. evestment has assembled 49 peer universes, constructed on investment style factors such as capitalization, quality, and sector emphasis. Peer groups are reconstituted quarterly. Peer Share Efficiency, then, compares a given manager s Active Share Efficiency to the average Active Share Efficiency for the specified peer group. n n n 9

10 Selected References Cohen, Tim, et al. Active Share: A Misunderstood Measure in Manager Selection. Leadership Series/Investment Insights, Fidelity Investments, Feb Available at accessed Aug. 11, Cremers, K.J. Martijn, and Petajisto, Antti. How Active Is Your Portfolio Manager? A New Measure That Predicts Performance. Working Paper, Available at papers.ssrn.com/sol3/papers.cfm?abstract_id=891719, accessed Aug. 12, Khusainova, Erianna, and Mier, Juan. Taking a Closer Look at Active Share. Investment Research, Lazard Asset Management. March 6, 2013, revised and republished June 25, Available at accessed Aug. 11, Mercer. Perspectives on Equity Investments Vol. 3 (3). Sept Deb Clarke and Matt Reckamp, eds. Available at www. merceris.com/uploads/documents/ vfrhgmwwc10589.pdf, accessed Aug. 11, Petajisto, Antti. Active Share and Mutual Fund Performance. Financial Analysts Journal Vol. 69 (4). July/Aug Available at accessed Aug. 12, Sharpe, William F. Mutual Fund Performance. The Journal of Business Vol. 39 (1, Part 2: Supplement on Security Prices). Jan. 1966, pp Available at accessed Aug. 11, Wittman, Scott, and Chandrashekaran, Vinod. Debunking Active Share, Investment Viewpoints, American Century Investments, Sept Available at institutional.americancentury.com/content/dam/americancentury/institutional/pdfs/viewpt/ debunking-active-share.pdf, accessed Aug. 11, About evestment evestment provides a flexible suite of easy-to-use, cloud-based solutions to help the institutional investing community identify and capitalize on global investment trends, better select and monitor investment managers and more successfully enable asset managers to market their funds worldwide. With the largest, most comprehensive global database of traditional and alternative strategies, delivered through leading-edge technology and backed by fantastic client service, evestment helps its clients be more strategic, efficient and informed. 10