Finding Star Fund Managers
January 27, 2014
Following Leaders and Laggards
With the new Fund Manager of the Year awards from Morningstar, a question of future performance of star fund managers inevitably arises. Apparently, the awards carry little short-term predictive value:
While our Fund Manager of the Year awards are recognition of past contributions rather than predictions of future results, we’re confident in each one’s long-term prospects because of their deep research resources and willingness to stick with their discipline in good times and bad. We wouldn’t expect repeat performances in 2014, as our winners and their rivals will wrestle with lofty equity valuations, policy-related volatility, a still-recovering economy, and the specter of rising rates.
Indeed, statistics on award winners presented by an article in The Wall Street Journal are not too encouraging:
||Beating the Benchmark
Generally, the best long-term predictors of fund outperformance remain a low expense ratio (fees), minimal portfolio turnover (a proxy for trading costs), and divergence from benchmark index weightings. The latter measure, known as active share should be high:
Prof. Cremers says the best funds tend to have active-shares percentages that are at least 60%. Large-stock managers should ideally have an active share above 70%. Midcap managers should have active share above 85% and small-cap managers should exceed 90%, he says.
Unfortunately, the proportion of funds with such big active shares has been falling over the years, which gave rise to “closet indexing,” as a chart from a Lazard Research study demonstrates:
The active share in the aggregate portfolio of actively-managed U.S. stock funds has been also declining:
However, active is not always a guarantee of strong performance, as shown in an earlier Alpholio™ post.
As for fund fees, an article in The New York Times points out that
The average total expense ratio, which encompasses management fees and operating expenses but not brokerage commissions and other trading costs, is 1.33 percent of assets a year for domestic stock funds and 0.97 percent for domestic bond funds, according to Morningstar.
Over time, these fees add up. According to a paper by William Sharpe, which estimated an average expense ratio of stock funds at 1.12% compared to 0.06% (now 0.05%) for the Vanguard Total Stock Market Index Fund (VTSAX, Admiral Shares):
Whether one is investing a lump-sum amount or a series of periodic amounts, the arithmetic of investment expenses is compelling… Under plausible conditions, a person saving for retirement who chooses low-cost investments could have a standard of living throughout retirement more than 20% higher than that of a comparable investor in high-cost investments.
However, as a Gerstein Fisher study found, the cheapest funds may not always provide the highest returns:
We found that the best performing quintile of funds was the second most expensive quintile (i.e., the 21-40% highest-cost ones), whether we equally weighted funds or asset-weighted them. The consistent results of the study: the cheapest quintile of funds was not the best performing, but the most expensive funds were the worst performing.
Similarly to a low correlation, a factor that is often quoted to aid stock pickers is a high degree of dispersion (measure of spreading) of stock returns. At first blush, it would seem that, just as with active share, an increased dispersion is beneficial. However, all it does is to enlarge the dispersion of active fund returns, without necessarily moving the average, as another article in The Wall Street Journal indicates:
A bigger spread between the best and worst stocks hasn’t helped active funds as a group, but it does tend to make good funds better—and bad funds worse… To put it another way: Markets that aren’t moving in lock step give active managers more rope with which to climb above the pack or to hang themselves.
While a low cost and small turnover coupled with a significant active share are generally good screening criteria, funds clearly have trouble with performance persistence. As our analyses have repeatedly demonstrated, even the star fund managers stumble, so outperformance is fleeting. This is where the Alpholio™ methodology helps by showing momentum in the smoothed cumulative RealAlpha™ for each analyzed fund, from which buy/sell signals can be derived. To learn more, please visit our FAQ.
January 4, 2014
Lack of Performance Persistence Matters in the Long Run
An article in The Wall Street Journal provides evidence that following both the recent leaders and laggards results in inferior investment performance:
According to these figures, after 20 years investors following the portfolio of the previous year’s top-performing newsletter would end up with only 2.4% of their original investment. Those following the portfolio of the previous year’s worst-performing newsletter would finish with virtually nothing. In contrast, investors in a broad market index would realize a return of over 460% in the same period.
These findings are similar to those of the recurring SPIVA reports, discussed in a previous Alpholio™ post. Mutual funds in the top half of quartile of the population are more likely to revert to the mean than could be expected by chance. On the other hand, funds in the bottom quartile are more likely to continue to underperform, and eventually end up being liquidated or merged with other funds.
The article indicates that merely focusing on lower-risk investment strategies is insufficient. Instead, the investor should extend the observation period:
You would do much better to focus on performance over far longer periods than the past 12 months. That is because, when picking an adviser based on his track record, you implicitly are betting that the future will be just like the period over which that record was produced.
While there is no magical track-record length on which you should always focus, 15 years is a good rule of thumb. The past 15 years—from the beginning of 1999 through the end of last year—encompass two powerful bull markets as well as two punishing bear markets.
As an example of a suitable mutual fund, the article offers the Turner Emerging Growth Fund (TMCGX, investor shares) with the highest annualized return among diversified U.S. equity funds in that long time span. Let’s take a closer look at that fund’s performance.
First, data from Morningstar show that the fund’s 15-year annualized return of about 17.6% was significantly higher than the 10-year return of about 10.6%. This indicates that the fund may have outperformed early on since its inception in late February 1998. Indeed, in the first couple of years the fund generated a return of over 355%, no doubt riding the wave of appreciation of small-cap stocks caused by the Internet boom. While this performance could in theory be repeated, it seems unlikely.
In the next three years through February 2003, the fund generated a cumulative loss of only 21% compared to about 52% of an average small-growth peer. This was another contribution to the high annualized 15-year return. However, the fund failed to beat the iShares Russell 2000 Growth ETF (IWO), a practical implementation of its prospectus benchmark, in four out of six most recent years. This is further illustrated by the fund’s performance relative to its reference portfolio of ETFs since early 2005:
The chart shows that all of the cumulative RealAlpha™ the fund generated through mid-2008 was subsequently lost. The fund strongly rebounded only in the second half of 2013. Moreover, the reference portfolio had a lower volatility than that of the fund.
The next chart depicts ETF membership and weight changes in the reference portfolio over the same analysis period:
The fund could have effectively been emulated by a collection of four ETFs: iShares Russell 2000 Growth (IWO; average weight 63.8%), iShares Morningstar Mid-Cap Growth (JKH; 25.4%), Vanguard Energy (VDE; 7.7%), and iShares 20+ Year Treasury Bond (TLT; 3.1%). Over time, the equivalent position in IWO became even more dominant, which implies that the fund’s characteristics were getting closer to those of its benchmark.
Only time will tell if the fund is able to outperform its reference portfolio on a consistent basis. There is no guarantee that the spectacular performance of early years will be repeated or even that the fund will be in existence in the next 15 years. In addition, as the above charts demonstrated, investor’s returns depend heavily on the timing on the initial investment. Therefore, while following the laggards is certainly not a fruitful endeavor, blindly following the leaders, even those with a long-term record, is not advisable either.
December 20, 2013
Searching for Consistent Outperformance
The December edition of the semi-annual S&P Persistence Scorecard brings better news than the July report — performance persistence of top mutual funds has slightly improved. However, percentages of domestic equity funds remaining in the top half of population for three or five consecutive 12-month periods were generally lower than those expected by mere chance.
Only 7.23% of all funds remained in the top quartile for three years and 2.11% for five years to September. While this beat random expectations of 6.25% and 0.39%, respectively, figures for the same periods ending in March were 4.69% and 0.18%.
In this context, an InvestmentNews article says that
S&P’s findings reinforce the idea that short-term performance chasing in mutual funds is likely to end badly for advisers, but for long-term investors, the lack of persistence doesn’t necessarily mean a bad outcome, provided you can stomach the volatility.
The article gives an example of this year’s top-performing large-cap blend fund, the MFS® Equity Opportunities Fund (SRFAX; Class A shares), which beat 90% of its peers and the S&P 500® index by annualized 2% in the trailing five years, despite a rocky performance:
The fund’s journey to five years of outperformance wasn’t smooth though. In 2009 it ranked in the bottom 10th percentile of large-cap blend funds and in 2011 it ranked in the bottom half of its category.
The article concludes that:
So if you can handle a fund that dips every now and again, without a big change to the investment process or management, the lack of persistence shouldn’t matter over the long run.
This example, however, forgoes important details. First, although Morningstar classifies the fund in the large-blend category, the prospectus benchmark for the fund is the Russell 1000® index. Unlike the S&P 500®, that benchmark includes medium-capitalization equities. As of this writing, the annualized five-year return of the fund was 19.09% vs. 18.41% of the iShares Russell 1000 (IWB), a difference of only 0.68%.
Second, compared to its stated benchmark the fund’s holdings are tilted toward mid- and small-cap stocks (collectively, about 46%) at the expense of giant-cap ones. This, again, indicates, that a pure large-cap index is an inappropriate reference.
Third, let’s use Alpholio™ tools to further peek under the fund’s hood. Here is a chart of weights of exchange-traded funds (ETF) in the reference portfolio for the fund:
The fund had top-four equivalent positions in the iShares S&P 500 Growth ETF (IVW; average weight of 23.8%), iShares S&P Mid-Cap 400 Growth ETF (IJK; 23%), PowerShares Dynamic Market Portfolio ETF (PWC; 21.2%), and iShares Morningstar Mid-Cap Growth ETF (JKH; 6.8%). This underscores the fund’s tilt toward not only mid-cap but also growth stocks.
Relative to the reference portfolio, the fund’s performance in the same analysis period was unimpressive:
The fund underperformed in 2009, so on a cumulative basis, its RealAlpha™ did not rebound until 2013. By this November, an investor who started with the fund in early 2005 would have not yet beat the reference ETF portfolio, as shown by the lag RealAlpha™ curve. In addition, the reference portfolio exhibited a lower volatility. (To get more information about this and other funds, please register on our website.)
The lack of performance persistence in mutual funds does matter in the long run and is more pronounced the longer the run is. In addition, as the above analysis demonstrated, relative performance heavily depends not only on the timeframe of the analysis, but also on the proper choice of the benchmark that fully adjusts for a fund’s risk.
November 20, 2013
Betting on Manager’s Past Performance
Two studies from Vanguard underscore how difficult it is to identify actively-managed mutual funds that not only survive and but also consistently outperform their benchmarks.
The first study shows that the majority of funds across all asset classes failed to outperform their prospectus benchmarks over the past 15 years through 2012:
The chart demonstrates that when assessing long-term performance, it is important to take into account liquidated and merged (“dead”) funds. Otherwise, statistics suffer from a “survivorship bias” that benefits funds still in existence. In addition, in most categories a median surviving fund exhibited a negative annualized excess return vs. the benchmark.
Statistics get worse when style benchmarks, assigned to fund categories, are used:
This is because many funds choose an inappropriate prospectus benchmark that does not reflect the fund’s actual investment style. In a majority of categories the survivors’ excess returns were even lower.
Finally, what are the chances that a fund ranked in the top quintile (20%) of U.S. actively-managed funds in terms of five-year returns through 2007 persisted in the same quintile in the next five years? About 15%, which is less than the 20% expected by chance:
As a matter of fact, about a quarter of such funds wound up in the lowest quintile, and about one-sixth disappeared altogether. Of all the funds available, only about 3% persisted in the top quintile in both five-year periods.
The second study shows that of the 1,540 U.S. domestic equity funds in the 15-year period through 2012, only about 18% survived and outperformed their respective style benchmarks:
However, almost all of these successful funds had five or more years of underperformance within the 15 years of analysis:
Moreover, about two-thirds of outperforming funds experienced at least three consecutive years of underperformance. In many cases, investors would have divested such funds and therefore not realize a full 15-year benefit.
All these findings underscore the need for a close monitoring of mutual fund performance. Outperforming funds are rare and do not persist in their winning streaks. Therefore, a dynamic analysis with a true adjustment for risk is required. Alpholio™ analyzes funds with a monthly frequency and provides buy-sell signals derived from the smoothed cumulative RealAlpha™ curves. These signals, among other inputs, can help investors make informed investment decisions. For more information about the Alpholio™ methodology, please visit our FAQ.
September 10, 2013
Persistence Is Futile
A research paper from Gerstein Fisher echoes recurring findings from SPIVA® — there is very little persistence of performance in mutual funds, even highly-ranked ones:
Unlike the Performance Scorecard that uses consecutive one-year returns, the study focused on rolling three-year trailing returns in each fund category. This multi-year approach had a smoothing effect, so percentages of the top quartile persistence were higher than those found by SPIVA®. However, it is clear that after a few years, a very small percentage of funds remained in the top quartile.
As depicted above, the study also found that
…in the rolling 3-year periods between 2002 and 2012, the likelihood that a top performer would descend to the bottom quartile of returns turns out to be exactly the same (27%) [47% in the chart?], on average, as the chance that a “dog” of a fund ends up ascending to the top quartile in the following three years.
In short, statistically speaking, an investor would have been just as well off picking a professional manager with an abysmal record of returns as he would have been with a star manager.
So, what is an investor to do? Alpholio™ provides one possible recommendation: investigate a smoothed cumulative RealAlpha™ curve to determine whether the fund will continue to add value on a truly risk-adjusted basis. To learn more, please visit the FAQ and take a look at the analyses of sample mutual funds in this blog.
July 26, 2013
The latest semi-annual S&P Persistence Scorecard study again demonstrates that the quest for performance persistence of actively-managed mutual funds is as futile as the famous resistance to the Borg in Star Trek. The majority of well-performing funds will be “assimilated” from the top quartile into the lower quartiles or from the top half into the bottom half of the fund population after both three and five years. The percentage of funds remaining in the top quartile or half after either period is much smaller than that implied by random expectations (chance), which indicates that active management skills are very fleeting.
While the study certainly brings value, it would be better if it compared actively-managed mutual funds not against each other but against an independent market benchmark specific to each fund category. Since an average fund underperforms such a benchmark by slightly more than the average expense ratio, fewer than half of the funds would beat it. In other words, if the performance bar were raised from a “relative” to an “absolute” level (as Alpholio™ does in its analyses), the distribution of funds would be much more skewed to the left, i.e. below the benchmark. However, such a distribution would truly reflect all the investment vehicles available to the investor, including index funds and exchange-traded products (ETPs). Chances are that in that case performance persistence would be even worse than the one shown in the study.
The findings of the study are corroborated by Morningstar in the analysis of its fund rating system:
“Fifteen years previously, the seminal paper on the topic, Mark Carhart’s ‘On Persistence in Mutual Fund Performance,’ offered this as its first and main conclusion: ‘Avoid funds with persistently poor performance.’
Morningstar’s research on the predictive power of its Morningstar Rating for funds (aka the star rating) further supports the notion. For most time periods, the star ratings do show mild persistence across the ratings bands, with the higher-rated funds in aggregate scoring better total returns over the next time period than the lower-rated funds. The finding is at its strongest, though, at the bottom.”
Since the statistical evidence shows that there is little long-term persistence of outperformance, the only solution is to focus on a shorter time frame. Because risk profiles in a given category vary widely across both funds and in time, a custom and dynamic benchmark has to be devised for each fund. This is exactly the methodology Alpholio™ uses. While no analytical approach based on historical data can guarantee a perfect prediction of future performance, smoothed cumulative RealAlpha™ curves do exhibit a certain degree of momentum for most funds. An investor can use this information to capitalize on fund outperformance trends, while avoiding periods of underperformance on a truly risk-adjusted basis.