To pick a fund manager look beyond three year trailing returns
Introduction
Picking an investment fund is hard. Researchers found that investment managers selected by U.S. pension plans, foundations and endowments underperform their opportunity set (the universe of funds which could have been selected) by an aggregate 1 per cent over three years.[i] This note summarises evidence on the characteristics of high performing investment funds, providing some guidance on selection criteria. There are four key points.
- Fund selection matters. An investor picking an Australian equities fund, with the ability to select a top quartile fund, could conceivably earn an above-benchmark return of 2 per cent a year over 10 years after fees.
- Persistence and style matter. Consider whether high returns in a prior period, whether it be three, five or ten years, is simply the result of an investment style performing well (which won’t persist if another investment style outperforms next period) or whether the investment fund outperformed compared to funds with a similar investment style.
- Active share matters. A reasonably high active share is a pre-requisite for an investment fund to achieve above-benchmark performance because it represents the non-index bets made by the portfolio manager. High active share at the stock selection level does not necessarily mean higher portfolio risk, given that portfolio exposure to common shocks to industry or value versus growth stocks can be diversified away.
- Fees matter as part of a holistic assessment. Fees are a function of an investment process, not a signal of future returns. While on average low return funds charge slightly higher fees than high return funds, screening on fees without consideration of returns persistence, style and active share will result in the investor excluding funds that generate high returns after fees.
What level of outperformance can investors reasonably expect?
In the 10 years ending August 2023, an investor holding an index fund that replicated the S&P/ASX 200 or the S&P/ASX 300 would have earned an annual return of 7.3 per cent after fees. Amongst non-index managers with a 10-year track record available for analysis, the after-fee returns are about the same. This is from a sample of 615 investment funds classified as Australian equities by Lipper with $97 billion under management. The actual 10-year return on $84 billion over 10 years was 7.8 per cent a year, representing outperformance of 0.4 per cent, but funds which did not survive for 10 years, and which are likely to have below-average returns, do not form part of the sample. So, consistent with evidence from around the world, non-index managers in aggregate earn just enough returns to offset expenses of about 0.7 per cent.
However, what matters in fund selection is the potential for above- or below-average returns. The top quartile of investment funds, on a dollar-weighted basis, earned annual above-benchmark returns of 1.8 per cent a year, while the bottom quartile underperformed by 0.9 per cent. Over shorter intervals there is greater dispersion of performance. For instance, over three years, top quartile managers earned above-benchmark returns of 5.1 per cent a year, while bottom quartile managers underperformed by 3.1 per cent. Again, this is due to the style-specific exposure of portfolios. Top quartile managers do not continue to beat their benchmarks by 5 per cent a year because market conditions change, and their investment style will either track or lag the market in different conditions. This means the goal is to select an investment fund that is likely to generate above-benchmark returns on average, not just in any one particular market.
Fund selection matters. The implication is that an investor picking an Australian equities fund, with the ability to select a top quartile fund, could conceivably earn an above-benchmark return of 2 per cent a year over 10 years after fees.
Is there persistence in performance?
Yes, to a modest degree, but we need to be specific about what we mean by persistence. Ranking investment funds according to 3-year trailing returns shows weak evidence of being able to predict the subsequent rankings over the next three years.[ii] A manager could easily be in the top half in one three-year period and the bottom half in the next three-year period.
However, there is more persistence in investment fund alpha, specifically if an investment fund outperforms a basket of stocks in the same style there is a reasonable chance the investment fund will exhibit the same alpha with respect to style next period.[iii] This occurs because investment managers generally have a style bias. For instance, a manager may have a persistent overweight position in value or growth stocks. And there are random fluctuations in whether value or growth stocks will perform well in the next period.
This is linked to the evidence of where high-performing managers deliver above-benchmark returns. When managers do generate above-benchmark returns it is most often due to security selection (picking the best stocks or the best bonds) rather than being able to time entry and exit into entire industries or asset classes. A manager who has persistent ability to pick stocks from within a particular style has a reasonable chance of earning positive alpha in consecutive periods.
Persistence and style matter. The implication is that investors should consider whether high returns in a prior period, whether it be three, five or ten years, is simply the result of an investment style performing well (which won’t persist if another investment style outperforms next period) or whether the investment fund outperformed compared to funds with a similar investment style.
How active is your fund?
Active share is the proportion of fund holdings that do not overlap with index weights. An index fund has zero active share. An Australian equities fund that holds just ASX: BHP has 90 per cent active share relative to the S&P/ASX 300, given that BHP is 10 per cent of the index; a fund that holds BHP and Commonwealth Bank (ASX: CBA) in proportion to their market capitalisation has 82 per cent active share. A typical non-index fund has about 30 per cent to 70 per cent active share.[iv] In an influential research paper published in 2009, researchers documented the relationship between active share and fund returns.[v] They reported that a 10 per cent increase in active share is associated with an increase in benchmark-adjusted returns of 0.7 per cent. The return premium increases to 0.8 per cent for funds of below-median size. In addition, once active share in stocks is accounted for, a concentrated industry position was not associated with a return premium.
The reason active share is important is that it represents the portion of an investment fund that is a bet against the market, with half of the active share being a long position and half being a short position. Recall that a top quartile investment manager, on average, earns returns about 2 per cent above benchmark over 10 years after fees. This is only possible with a sufficient active share. For example, if the active share is only one-third of the portfolio, to beat the benchmark by 2 per cent before fees means that the active positions need to generate a return of 6 per cent (in other words the BUYS have to, on average, beat the SELLS by 6 per cent, which is a challenging task). In contrast, if the active share is two-thirds of the portfolio, to beat the benchmark by 2 per cent means that the active positions only need to generate a return of 3 per cent. Once fees are accounted for, taking an active position becomes even more important.
All else equal, increasing portfolio concentration entails more risk. But in portfolio construction, all is not equal. Holding Australia’s five largest banks would result in an active share of 78 per cent. Holding BHP, CBA and ASX: CSL would result in an active share of 76 per cent. But the latter portfolio will exhibit considerably less volatility. Investment funds taking concentrated positions in individual stocks generally offset this concentration with positions that diminish risk (for example, by diversifying across industries, or across value and growth stocks).
Active share matters. A reasonably high active share is a pre-requisite for an investment fund to achieve above-benchmark performance because it represents the non-index bets made by the portfolio manager. High active share at the stock selection level does not necessarily mean higher portfolio risk, given that portfolio exposure to common shocks to industry or value versus growth stocks can be diversified away.
Fees are not a signal of quality
There has been a recent focus on superannuation fund performance benchmarking, given the $3.5 trillion invested in this asset class, including $1.0 trillion in MySuper assets.[vi] Investment management fees have been in a secular downturn for many years, driven by economies of scale, technological advances and competition. Researchers have consistently failed to find that fees are a signal of high prospective returns. Fees generally reflect inputs rather than results: An index fund benefits from economies of scale and very low turnover, while a highly concentrated fund requires more expensive due diligence on any one particular investment. In some cases, that diligence pays off, while in other cases the expense simply represents a loss of value to investors. Over periods of one, three, five and ten years, bottom quartile funds on a dollar-weighted basis charged slightly higher fees than top quartile funds, around 0.2 per cent to 0.5 per cent. But the fee differential between top and bottom quartile funds is far lower than the performance differential, which on an annual basis is 2.7 per cent over ten years and 3.3 per cent over five years.[vii]
Fees matter as part of a holistic assessment. Fees are a function of an investment process, not a signal of future returns. While on average low return funds charge slightly higher fees than high return funds, screening on fees without consideration of returns persistence, style and active share will result in the investor excluding funds that generate high returns after fees.
Conclusion
Selecting an investment manager is hard because we have limited data to separate skill from luck. Ranking on three years of past performance is analogous to tipping the winner of the NRL or the AFL after three rounds of competition. But consideration of four issues can help.
First, understand the realistic potential for above-benchmark performance. Top quartile managers, on average, have the potential to outperform their benchmarks by around 2 per cent a year after fees over a 10-year period.
Second, there is low persistence of investment fund rankings on total returns or returns versus benchmark because investment funds have a style focus and different styles perform well in different market conditions. But there is more persistence in fund returns relative to an investment fund’s style exposure. Think value versus growth, or income versus capital gains, as style dimensions. Consider whether an investment fund performed well compared to other funds with similar value or growth characteristics. That helps to determine whether past performance reflects the ability of the investment manager to pick high return stocks, or whether the track record merely reflects a style that was in favour for a while.
Third, outperformance can only result from a minimum active share: The non-index bets taken by an investment fund. Typical active share is around 30 per cent to 70 per cent, and research suggests a 10 per cent increase in active share is associated with a 0.7 per cent increase in annual returns (0.8 per cent for small funds). Too low an active share means that the active bets have little chance of generating returns that offset fees.
Fourth, fees are important as part of a holistic approach to fund selection. High fees do not signal high quality. Rather, fees are a function of the investment approach. On average, low returns funds charge slightly higher fees than high return funds (about 0.2 per cent). But the fee differential is small compared to the return difference (about 3 per cent). Obviously, high fees raise the bar to outperformance and investment managers compete, in part, on fees. But a myopic focus on fees has the potential to exclude high return funds from consideration.
References
Endnotes
[i] Goyal, Wahal and Yavuz (2023)
[ii] Cornell, Hsu and Nanigian (2017)
[iii] If you rank managers by alpha compared to a model based upon style characteristics like value and growth there is persistence in alpha (Vidal-Garcia, 2013), especially if alpha is measured on a pre-fee basis and in dollar terms rather than in percentages (Berk and van Binsbergen, 2015). And there is persistence in performance of funds with similar size (for example, in Europe, small winning funds beat small losing funds again next year on average; Leippold and Rueegg, 2020).
[iv] I measured the active share for 52 funds for which portfolio holdings were available, comprising $23 billion total net assets. 40 funds with $21 billion total net assets had an active share within the range of 30 per cent to 70 per cent.
[v] Cremers and Petajisto (2009)
[vi] (VIEW LINK) accessed on 3 October 2023.
[vii] In a study of funds holding equities listed in Europe and Asia (Guido, Beckers, Hazenberg and Van Der Scheer, 2022) returns from 2008 to 2020 were 0.7 per cent higher than average for the sub-set of funds that met the following criteria over the prior year: lowest quartile fees, highest quartile information ratio, highest quartile tracking error (a different proxy to active share for how much the fund deviates from benchmark).
5 topics
3 stocks mentioned