Conviction in Equity Investing
Mike Sebastian and Sudhakar Attaluri, The Journal of Portfolio Management, Summer 2014
Active management helps improve market efficiency through the identification of mispriced securities, but often the benefits do not accrue to individual investors. To maximize returns after fees, the authors argue that investors should either allocate to high conviction managers—measured as top decile active risk—or passively invest in a broad global equity index.
Gross excess returns from active management are often enough to cover fees and trading costs but are frequently not high enough to benefit investors. According to the authors, a manager would need to achieve an information ratio of 0.65 to generate alpha for investors. Roughly 70% to 90% of investment products are administered by managers with enough skill to cover fees, but the percentage of skilled managers that provide statistically significant excess returns to investors has steadily decreased from around 20% in the 1990s to less than 2% today.
With low odds, investors need to be selective when picking active managers and should consider management fees, trading costs, and time spent monitoring. Regarding the latter, a Vanguard study suggests investment committees frequently spend approximately 10% of their meetings discussing manager selection and 40% reviewing manager performance—highlighting a degree to which active management costs exceed just simple fees.
The authors use the level of active risk to determine a manager’s conviction level, concluding that only managers with active risk in the top decile earned positive returns net of fees. Skilled stock pickers that were not in the top decile often held alpha-reducing positions to minimize benchmark risk. This risk-minimization behavior can be partly attributed to the rise in popularity of benchmarking and style indicators around the early 1990s that penalized managers that missed their benchmarks. Managers in the top decile patiently invest on a long-term basis and treat investments as highly illiquid.
Given the dwindling number of skilled active managers, Sebastian and Attaluri recommend investors interested in pursuing an active management strategy commit 80% or more of public equity allocations to high-conviction managers that invest only in their best ideas. If this approach is unappealing, the authors suggest investors pursue a cost-minimization strategy and index to a broad global equity benchmark. Regardless, investors should always place a high level of scrutiny on prospective active managers and rely on indexing whenever there is uncertainty about quality.
Deactivating Active Share
Andrea Frazzini, Jacques Friedman, Lukasz Pomorski, AQR Capital Management, April 2015
The use of active share as a tool to judge a fund’s potential future return relative to a benchmark has grown increasingly popular in recent years, following academic studies linking high active share funds to outperformance. However, a new article published by AQR reviews the theoretical and empirical basis of active share and finds little support justifying its use in predicting fund returns.
Cremers and Petajisto coined the term active share in 2009 after concluding that high active share funds—or funds that are significantly different from their benchmark—typically outperform both their benchmark and low active share funds. Although institutional investors have begun relying on the metric to inform the decision-making process, with some even embedding it as a requirement in their investment guidelines, AQR identified problems when replicating the original academic studies.
A primary problem, according to the authors, is that active share has a benchmark-selection bias. Low active share funds are typically benchmarked to large-cap indexes while high active share funds have a small-cap bias. Based on the original data alone, Cremers and Petajisto suggest managers with high active share outperform by over 2% per year on a statistically significant basis. But because the styles performed differently over the sample period, inferences based on active share are misleading.
After controlling for the benchmark-selection bias, the authors note that while high active share funds still have higher fund returns, the outperformance is no longer statistically significant. In fact, benchmark-adjusted returns fell from 2.14% to 1.16%, and alphas decreased nearly two-thirds, from 2.42% to 0.88%. With the positive adjusted values likely to be just statistical noise, active share does not seem to be predictive of fund returns.
Ultimately, the authors find lacking theoretical and empirical support justifying the use of active share in predicting fund outperformance. However, active share can still be used as a tool for cost evaluation alongside other metrics like tracking error to determine if a manager is taking risks commensurate with fees.
Patient Capital Outperformance: The Investment Skill of High Active Share Managers Who Trade Infrequently
Martijn Cremers and Ankur Pareek, September 2014, White Paper
The authors examine mutual fund and institutional investor data to identify investment strategies that outperform benchmarks, concluding that among high active share portfolios only those with low turnover achieve outperformance. In other words, once an investor identifies funds with patient investment strategies, those that are distinct from their benchmarks are more likely to deliver excess returns than closet indexers.
Cremers and Pareek initially measure the active share—or the proportion of a fund’s holdings that is different from its benchmark—of a sample of mutual funds and institutional investor portfolios. They then determine the level of turnover in each fund by using a measure called fund duration, which is the weighted-average length of time that a given dollar value of stock is held. Performance is analyzed by sorting the funds into five quintiles based on each metric, delivering a total of 25 distinct portfolios.
From 1995 to 2013, only mutual funds in both the top active share and fund-duration quintiles were capable of outperforming their benchmark. The average outperformance of this portfolio was 2.30% per year net of fees, which compounded over the 19-year period to result in a cumulative excess return of 54%. The authors find that funds with long fund duration consistently and significantly outperformed funds with short fund durations across all active share quintiles.
Similarly, patiently managed institutional investor portfolios with high active share outperformed. On average, portfolios in the top quintiles of both metrics returned a statistically significant 2.22% per year in excess of their benchmark and gross of fees. This outperformance is even more pronounced when reviewing just hedge funds. While the hedge fund data reviewed was limited to long-only and quarter-end positions, the subset’s annual outperformance was 3.64% gross of fees.
The authors regressed the top quintile portfolios against seven factors to assess managerial skill. They concluded that the excess returns can partly be explained by fund managers picking safe, value, and quality stocks and maintaining conviction in their decisions over relatively long periods, a style they highlight is not much different than Warren Buffet’s. However, Cremers and Pareek find that the majority of the outperformance in the top institutional portfolios cannot be explained by these factors alone, suggesting manager successes may be linked to contrarian ideas.