An interesting Reuters article
on active share and closet indexing. Three highlights: (1) low active share = pointless waste of investor dollars. (2) Petajitso now works for BlackRock. (3) He suggests that an active share below 60 is a closet index fund. Apparently the average large-core fund is somewhere in the low 70s.
Several fund companies are starting to share, selectively at least, their active share data: Artisan (with some funds in the 99 range), Touchstone, William Blair among them. I'm trying to gather some of that data for our March issue, but it's tough since Morningstar won't give outsiders access to it.
If, perchance, you notice another fund firm that's sharing the data, would you let me know, please?
Great work David.
MFO has aggressively explored mutual fund active share issues over several years.
To refresh your memory, I posted on this topic in November, 2012. The submittal drew excellent divergent viewpoints from a number of MFO members. I enjoyed the well crafted exchanges.
Here is the internal Link to the postings:
In that exchange, I recommended some analysis and interpretations offered by Columbia University professor and prolific writer Michael Mauboussin. He generated his perspectives on active shares while preparing a piece titled “Seeking Portfolio Management Skill”. Here is the Link to his original work:
Mauboussin concludes that a few managers do demonstrate fund management skills and active shares is one component that measures a skill set.
Indeed, we must be on constant alert for active mutual fund managers who hug their benchmarks as a means to secure their jobs, and not to enhance client returns. Incentive alignment is always a difficult task, and is often not properly executed.
Thank you for reopening this discussion line.
I find this so frustrating, since M* includes the data, you can see it in the pull-down menu, but not export it.
Another path we have explored is via CRSP Survivor-Bias-Free US Mutual Fund Database. We put a request into Booth and at the discounted rate for small outfits like MFO, it will only cost us about $9K per year, so we're taking up a collection.
Maybe we should ask the mutual funds directly to start providing it and we can maintain on the MFO site for free (well effectively, thanks to our generous voluntary contributors). Hmmm...
I know, I know. There are rules about selectively releasing data. Something about equal disclosure. But somebody is getting this info; therefore, it seems only right to make it more widely accessible. Being an efficient market and all.
Thank you for reading and replying to my post.
I honestly feel that I might be the least informed MFO member, who posts semi-regularly, to respond to your sensible question. Almost any frequent MFO contributor could construct a more meaningful and deeply researched active manager listing than I can generate.
Sometimes within the last year I concluded that the incremental reward/time commitment in identifying superior active mutual fund managers was not worth the effort in terms of my realized payoffs. This bottom-line conclusion is consistent with much academic research in this arena.
That’s not to say that my personal decision is in any way universal. Successful fund management does exist and careful research can isolate the winners from the losers. I believe that a few MFO participants are very proficient at this challenging task. It does demand time and constant monitoring. At this juncture in my retirement I am not prepared to make that commitment.
In the 1990s, 100 % of my mutual fund portfolio was actively managed. Today, about half is still invested with active managers, but about half is with Index products. Within one year, my plan is to convert the mix to a 20/80 passive-heavy fund/ETF allocation. I still enjoy the excitement and the challenge of the game.
Both academic and industry studies demonstrate the steep hill that active fund management must climb. As the number of actively managed funds in a portfolio increases and as the time horizon expands, the likelihood of outdistancing an equivalent passive portfolio sharply decreases towards single digit odds.
Active fund managers find it difficult to overcome their cost drag and changing market conditions. Investing policy, manager style, and fund size inertia elements make it hard to dynamically adjust to a rapidly evolving marketplace.
The historical record demonstrates that consistent outperformance by actively managed funds is mostly nonexistent among the mutual fund population. The Standard and Poors’ SPIVA and persistency scorecard reports document this failure over an impressive timeframe. Updated versions of these reports will be shortly released covering 2013 mutual fund management performance.
If I were searching for superior actively managed funds I would look for low costs and low portfolio turnover rates as guiding criteria. I would search for positive Alpha performance results over an extended timeframe, the longer the better (at least 5 years). I would seek high Information Ratios since a fund’s volatility detracts from overall wealth accumulation. I equate positive Alpha and Information Ratio as two measures of managerial skill.
I would accept the proposition that a fund manager will not be successful each and every year; bad times and bad decisions happen so patience is a virtue. Bill Miller is an excellent illustration of an active fund manager who would test patience to a maximum.
Since you have persevered and have endured my numerous investing nuances, I am now prepared to divulge several of my favorite fund managers. I claim no prescience here, but I do currently own the following array of actively managed funds; in fact, I have owned them since the mid-1990s.
From Fidelity, I hold FCNTX and FLPSX. From Vanguard, I have sizable positions with VWINX and VWELX. And from Dodge and Cox, I have DODBX and DODGX.
I make no eureka assertions relative to the wisdom of these choices. They have served me well for over two decades. Most likely I will retain these products when I finally get my mix of actively managed funds down to my 20 % goal level. Note that some of these funds are team managed so superstars are not a necessary prerequisite.
I find these choices adequate for my purposes; they surely might not be appropriate for your goals, risk profile, or time horizon.
I hope that some MFO members, who are more diligent, more qualified, and far more current at exploring the active fund management cohort, respond to your question. I’m sure they will provide deployable specific recommendations and some useful selection guidelines.
>> concluded that the incremental reward/time commitment in identifying superior active mutual fund managers was not worth the effort in terms of my realized payoffs.
you have for two decades also gone with rightly well-regarded active managers.
My own search led me to the Sequoia fund which is low-beta and is managed by an investment team with quite a command of all aspects of the securities they own (judging by their investor day transcripts). On the other hand, its asset size has doubled recently, it charges a 1% fee, and it just barely beats a midcap index over the past ten years. Yet this is the best fund I could find.
In this terrific article on active share from a couple of weeks ago Where have all the star managers gone? [wsj] there was a very useful interactive table of funds and their active shares. Unfortunately, it looks like the article no longer links to that table. SEQUX I believe was 89%.
If you google the blue phrase and add "fund" after "star", you get the article and the link to the table. My problem with the table was that only US indexes were used, so far as I could tell, so it didn't show which international funds were better than their regional indexes.
I also have a mix of index and mutual funds. Most of my domestic holdings are in index ETFs and most international small caps are in actively managed funds, with other asset classes a mix of both.
Thanks to everyone for their enthusiastic and friendly participation; all solid stuff.
From an historic perspective, MFOer Guy’s Sequoia Fund is a wise choice for most any portfolio. It had a storied record when guided by the legendary Bill Ruane from Warren Buffett’s prestigious SuperInvestors short list. When Ruane headed the Sequoia operation it held closely to the investment principles developed by Benjamin Graham.
However, since Ruane’s passing in 2005, the fund has experienced a wild ride under its current management. Certainly, 2009 was a major disaster, but the current Sequoia management team seems to have survived and at least partially recovered some of Ruane’s mojo.
My wife has access to all the WSJ articles because of our subscription commitment. MFOer Guy is on-target with his recall that the Sequoia Fund has an active share rating of about 90 %.
For what its worth, the actively managed funds that I referenced have the following active share percentages: FCNTX (52%), FLPSX (85%), VWINX (54%), VWELX (48%), DODBX (71%), and DODGX (71%). The dispersion is rather large, and is not necessarily representative of any outperformance measure.
MFOer STB65 inventively found a key that unlocks the door to the WSJ article for non-subscribers. Wow, great stuff; please take advantage since the Table lists hundreds and hundreds of mutual funds.
The Table is based on data assembled by Professor Martijn Cremers of Notre Dame. He and Blackrock’s Antti Petajisto have been exploring the active share issue for several years. A search of the Internet uncovers a large number of references to their work product. You might want to explore these basic papers.
Vanguard recently (May, 2012) published a research paper on Active Shares. It seems to be a fair assessment. Here is the Link to their paper:
The Vanguard 16 page research report concludes that Active Share percentages are NOT the solitary magic elixir to identify superior fund managers. It is one factor. The process of identifying mutual fund manager winners has many more dimensions that include additional components like low cost structure and low trading frequency. Vanguard does conclude that Active Share percentage should be included as one element in the selection criteria list.
The Vanguard study appears to be comprehensive. Please exploit this very readable report. One of its findings reinforces the conventional investing wisdom that excess reward and risk are positively correlated.
Vanguard concludes that “The higher the active-share level, the larger the dispersion of excess returns.” In the investment world, free lunches are rare events.
I hope all you guys (and especially Guy) found this exchange useful. I did.
ASTON/River Road Independent Value Fund ARIVX 97%
Greenspring GRSPX 96%
Akre Focus Fund AKREX 93%
Queens Road Small Cap Value Fund QRSVX 93%
Oakmark Equity & Income Fund OAKBX 91%
Appleseed APPLX 87%
Berwyn Income BERIX 83%
Jensen JENSX 83%
Vanguard Healthcare VGHCX 80%
FPA Crescent FPACX 78%
Janus Balanced Fund JANBX 73%
Still questing about.
From the WSJ article
If you don't have current subscription, search for
"And the Next Star Fund Manager is" on Bing/Google and click the WSJ link to get full article and full table.
Here's a link to that active share table. May be worthy of a bookmark.
Concentrated funds seem to do the best, which is why I like these funds, but AOR is not bad for 0.30% ER, similar to Vanguard Star which I use for kids 529 plan. I think all of the above make good core funds and then go heavy into good long term aggressive growth funds such as Yacktman after a big pullback . VIX over 40 is usually good buy signal.
ARTGX, for instance, is indexed by the WSJ to the NASDAQ 100, which oddly creates an active share of 79%. In Artisan's fund literature, they list the number as 90.4%.
FPACX, to pick another popular value oriented fund, is also linked to the NASDAQ 100, for a score of 78%. FPA lists it as 90.3%.
PARNX is listed by WSJ as 79% against the NASDAQ, and by Parnassus as 92.6% against the S&P 500.
DODWX is listed by WSJ against the "Russell Top 200 Value Tr", whatever that is, at 76%. Dodge and Cox don't provide the number, but the benchmark they use is the MSCI World Index.
Some standardization would seem to be necessary for this to be useful.
edit: Looking farther, a lot of the Journal's "most similar benchmark" choices are comical. In the same category "FTSE High Dividend Yield" you have COBYX, BRUFX, and LSFIX, as well as several each of natural resources, consumer products, and healthcare funds...
Which makes the decision to assign all of the Artisan Global Value and Opportunity funds to the NASDAQ sort of odd. Odder still, Fidelity Europe is tagged to a Russell Top 200 US Growth index. The Fidelity Select portfolios to a potpouri of unrelated benchmarks. Oakmark Equity & Income is tagged to a socially-responsible stock index. All balanced funds are tagged to stock indices, but the particular index varies. Funds named "Blue Chip" get tagged to one of four different indexes.
Professor Snowball is spot on-target with respect to the sensitivity of Active Share percentages to a selected benchmark. It makes a difference, sometimes a huge enough difference to compromise its predictive power for excess returns (Alpha).
The selection of a proper benchmark is an open issue without any industry standard to provide guidance. Overall, the investment industry has accepted Active Shares as an addition to an investor’s toolkit, yet the benchmark determination remains an unresolved matter that is problematic.
Researchers Cremers and Petajisto introduced the Active Shares concept to the investment community in the mid-2000s. It was welcomed as another way to identify and choose winning active mutual fund managers. That’s excellent; we need all the tools we can get since the odds of selecting a winning manager are not high. Antti Petajisto followed that original work with an examination of mutual fund performance as a function of the Active Share parameter. He concluded that higher Active Share percentages produced superior Alpha, excess returns.
That outcome is logical since an active fund manager can only generate excess returns if he deviates from the benchmark portfolio. But therein lies a risk. If the manager does not choose his favored stocks or preferred sector weightings wisely, he is exposed to underperformance outcomes. Choosing an inappropriate benchmark can easily distort any evaluation.
In an earlier posting, I referenced a Vanguard study that used a different set of benchmarks. For convenience, I repeat the Link as follows:
Using its own set of benchmarks, the Vanguard study conclusions dramatically depart from those of Petajisto. On average, Vanguard finds few excess returns from the higher level of Active Share managers. Why? Perhaps it was the study timeframe used by Vanguard ( 2001-2011). I suspect that the selected benchmarks are a major contributor to the disparate findings. Details matter.
I recognize that all you guys are under time constraints, so asking that you read a technical report is burdensome. So allow me to focus your attention on several of the pertinent Vanguard tables and graphs that illustrate the difficulties with Active Shares, at least from the Vanguard perspective.
In Figure 4, Vanguard shows the correlation coefficient between Active Shares and Excess Returns during the Performance Period test. The correlation coefficient was 0.08, almost nonexistent at the zero level. That’s almost a random relationship, something like a fair coin toss. This finding directly contradicts the Petajisto conclusion.
In Figure 6, Vanguard presents Excess Returns for various groupings of active funds as a function of time including both the evaluation and performance periods of the study. Excess Returns for the most successful grouping (the concentrated style) degrades with time while the others are nearly constant. The overarching takeaway from this plot is that the strong pull of the reversion-to-the-mean investing axiom is fully operational here. Performance persistence is a myth in most instances.
In Figure 7, Vanguard illustrates Excess Returns as a function of Active Share. The plot pictures a widening wedge that demonstrates that outcome risk is increasing with an increasing Active Share component. Note that there are as many data points in negative Excess Returns quadrant as there are in the positive Alpha section. These data do not inspire a high confidence level.
In Figure 11, Vanguard ranks the funds by quartile on both an Active Share and cost basis. As a general observation, the Vanguard study only finds that high Active Shares produce positive Alpha when they are combined with low costs. Also, these positive Alpha zones, in terms of absolute gains, are greatly outweighed by the many negative zones with substantially higher absolute negative Alphas. Buyer beware.
I have only summarized some of the Vanguard study highlights in the hope of lessening your burden. The Vanguard report offers other investor insights.
I believe both Petajisto and Vanguard conducted honest studies. The disparate results are likely timeframe and/or benchmark related. This is not a simple problem. Newspaper legion H.L. Mencken fully captured the dangers inherent in modeling nonlinear systems with his pity observation that “For every complex problem there is an answer that is clear, simple, and wrong.” I sure don’t know who is right in this instance. Perhaps both are.
I do believe that Active Share is yet another tool to assess active management performance. But it is not a magic elixir; it should not be used in isolation. It should be merged with other discriminating signals. The ultimate utility of the Active Share measure is an unsettled issue. Whenever using it, please be sure to understand the appropriateness of its benchmark, and please use it as a part of other decision making inputs.
So, I’m coupled to Professor David’s hip on the determination of Active Share benchmark selection.
I’ll close with a Will Rogers quote that always appealed to me: “It isn’t what you don’t know that gives us trouble, it’s what we know that ain’t so.”
Where in Artisan's literature are you seeing this number?
They're on the Artisan factsheets for the individual funds.
I am just not finding it
The active share is listed under "portfolio statistics."
edit: Try here. ARTGX here.