Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Support MFO
Donate through PayPal
Mutual Funds' 5-Star Curse: MFO's David Snowball Comments
Hello fellow MFO members ... David Snowball is quoted in this article (see below) and is worthy of reading not only for his comments but the whole article is good reading.
'(David Snowball, publisher of the Mutual Fund Observer website, highlights the Marsico and Fairholme funds as examples of funds that were top performers under charismatic managers— Tom Marsico and Bruce Berkowitz, respectively—and have faltered. "[Marsico] was a family very much focused on Tom Marsico and his bold take on investing," Mr. Snowball says. The same is true of Fairholme, he says, "which has been struggling in a number of ways lately.")'
The referenced article is indeed excellent. However, it really contains no surprising information. It just reinforces a wide body of earlier research findings that grow more compelling with each year’s additional data.
Mutual fund performance persistence is largely a myth, with a few outstanding exceptions. The game plan should be to find and stick with those noteworthy exceptions. The article identifies a few of them.
In physics, the law of gravity dominates. Although not quite as universal, within the investment community, a regression-to-the-mean law seems almost as pervasive.
The evidence against fund persistent outperformance is overwhelming. In its early years, the Morningstar 5-star rating system proved so vulnerable to performance decay, and subsequently to star erosion, that the firm was forced to modify and expand its ranking methods to salvage its fading reputation.
The numerous periodic table of returns published annually by a host of industry giants demonstrate the fragile nature of last year’s winners. The rotations are swift. Buying last year’s top performers has proven to be a Loser’s game.
Each year, Fortune magazine issues a list of the most respected and the most despised companies. The list does have some stability year over year. However, the market returns from owning the most accomplished companies over the long haul do not match those yielded by the least honored outfits. From an investment rewards perspective, this is yet another illustration of a regression-to-the-mean pull.
All MFO regulars should be familiar with the often referenced Standard and Poor’s SPIVA and Persistency reports. These documents have registered fund persistency failures for many years. Here is the Link to those data rich reports:
If motivated, you can click on the specific SPIVA and Persistency items to explore more deeply. The SPIVA mid-year report was just released. I have not yet accessed it.
Michael Mauboussin has written extensively about the tradeoffs between luck and skill in determining outcomes. He concludes that as the skill level of professional money managers has escaladed over time, the luck component becomes the more dominant factor. There are fewer and fewer Excess Return opportunities. Hence, expect that persistent superior performance should degrade even more in the future and should be more difficult to identify.
That’s an imperfect signal that favors an Index-like approach. Institutions have recognized that signal, and are moving more aggressively in that direction. Perhaps we should too?
In brief exchanges such as yours, it is sometimes a challenge to interpret your intentions. Were you simply pulling my leg or were you asking serious questions? I’ll assume the serious purpose.
A reversion-to-the-mean is surely not a law in the same league as a physics law. I used the term just to draw a parallel between a hard science and a soft science. More properly, a regression-to-the-mean is probably better characterized as a statistical concept.
It is a way statisticians describe an outlier event that deviates from some average, expected outcome. In their earlier years, Kahneman and Tversky observed that when pilots performed either above or below their normal performance standard, subsequent flights reverted to their average skill levels. Instructors like to think it was their training guidance; the Behavioral scientists judged otherwise.
If a ball player scores 4 hits in a single game, it is highly likely that he will not repeat that exceptional output in the next game; he will display a regression-to-the-mean tendency. Likewise, if he hits at a 400 batting average for a month, it is doubtful that his hot hand will persist. Again, he will revert towards his historically normal batting average.
The time horizon for the regression-like pull is unspecified. When investing, it depends upon the investors specific timeframe. If he trades frequently, the comparison period must be representative of that short timeframe. If the investor is planning for the long-term, the representative “mean” value should adequately reflect a long time horizon.
My preference is for longer running data collection averages for comparative purposes. I recognize that this bias introduces data staleness risk. So some experimentation is needed to properly identify a meaningful data period to calculate a “Mean” value.
Sorry for wasting your time with this reply if you were just playing “a pulling of the leg” game. I’m sure you knew everything I just reviewed.
But did you know that Earth’s gravity is NOT precisely constant? It depends on location, ground water level, and shifts under tectonic plate activity. Social laws are temporal and uncertain. Even physics laws are approximations and sometimes infrequently revised. Change happens.
>> whole article is good reading. >> The referenced article is indeed excellent.
Omg, how the heck so? Seems extremely feeble to me, even by the feeble standards of WSJ and MW service pieces. The rules are lame, not really or demonstrably or consistently true, the quotes from Kinnel and Snowball highly selective (you can always find such an example), the absence of broad data lazy and troubling, and the whole thing a yawn and worse, useless, unhelpful, unactionable. 1 - Sure, some hot managers ain't no more. Maybe most. Oh, but wait, Danoff ahead. Oh, wait, he's longterm. 2 - Sure, GLRBX has been good forever, and you shoulda bought and held, period, full stop, you dope. 3 - Everyone has been writing forever about size, which matters absolutely, sure, yes, except when it doesn't, and most managers say it doesn't, but they are wrong, or are they? Wait, Danoff ahead. Good thing Tillinghast and D&C were overlooked. 4 - KIS, okay, so indexing is best. Or is it? And cherish sc and value, oky-doky and all righty, then. Twas ever thus. 5 - duh. 6 - Balance and blend and mix are important. Wow. And 7, the last rule, the future, well, read those last three questions out loud and see if you can do it without laughing.
The whole thing is a weak, but actually worse than weak, September yawn, to my eye. I worry when smart people here reflexively write that something is good or excellent. Is there anything whatsoever in this article that would not have occurred to any regular visitor to this forum??
It is the classic debate of skill vs luck. My view is that skill exists, especially in certain talented and hardworking individuals. However, we (meaning both the experts and their followers) tend to overestimate our skills. Long term trend is still an average trend based on overall fundamentals (e.g., overall economic growth). There is some statistical distribution of performance around this average trend and it is very hard to show that persistence in performance is due to skill or a consequence of perceived patterns in a underlying randomly distributed outcomes (kind of like a sequence of five heads in a coin toss). We are all attempting to tilt odds in our favor by studying the art, understanding our behavioral limitations, keeping things simple enough to manage, be humble to change when our views turn out to be wrong and have the courage to take manageable risks when we are convinced about our understanding.
@Kaspa- Yes, I agree, and this is exactly why some of us maintain perhaps 20 or so funds rather than just 2 or 3. Of course there's overlap- that's the whole idea!
We are all attempting to tilt odds in our favor by studying the art, understanding our behavioral limitations, keeping things simple enough to manage, be humble to change when our views turn out to be wrong and have the courage to take manageable risks when we are convinced about our understanding.
When you questioned the appropriateness of calling the investment world’s regression-to-the-mean observed phenomena a law, it did prompt me to do some deeper thinking on the matter. Well, that puts us into some dangerous waters.
Certainly investment wisdom and operational rules assembled over decades of practical experience deserve some elevated status when they address investment process and not investment outcomes. Probably assigning these assembled rules the title of “laws” might be a “bridge-too-far”. Why did I do it?
In thinking about my readings over the last month, I recalled a White Paper produced by the respectable GMO investment house. It was written by James Montier, Montier titled his work “The Seven Immutable Laws of Investing”. Eureka! That was likely my inspiration. Here is a list of his laws, immutable no less:
1. Always insist on a margin of safety 2. This time is never different 3. Be patient and wait for the fat pitch 4. Be contrarian 5. Risk is the permanent loss of capital, never a number 6. Be leery of leverage 7. Never invest in something you don’t understand
There’s a ton of Warren Buffett and Benjamin Graham embedded in these “laws”. I believe they serve as excellent generic investment guidance. Maybe they do rise to the level of laws.
If you are interested, here is a Link to the complete White Paper:
These rules do emphasize investment process over investment outcomes. A nice way to illustrate the tension between process and outcomes is to visualize a 2 X 2 matrix with Process on the vertical axis and Outcomes on the horizontal axis. Both dimensions can have either a good or a bad result.
The 2 X 2 boxes can be filled with the following conclusions: Good process/good outcome is an earned reward, good process/bad outcome is unlucky, bad process/good outcome is lucky and is likely a misleading signal towards overconfidence, and bad process/bad outcome is true justice.
The key to continued long-term investment success is a solid Process that is NOT abandoned when the markets turn South. It is amazing how many investors are lured by the Sirens' song of a lucky outcome. The likelihood of any future success with a poor investment Process is dubious at best.
Enough of this serious stuff for now. I too liked Kaspa's submittal; good stuff, very well summarized.
Comments
'(David Snowball, publisher of the Mutual Fund Observer website, highlights the Marsico and Fairholme funds as examples of funds that were top performers under charismatic managers— Tom Marsico and Bruce Berkowitz, respectively—and have faltered. "[Marsico] was a family very much focused on Tom Marsico and his bold take on investing," Mr. Snowball says. The same is true of Fairholme, he says, "which has been struggling in a number of ways lately.")'
The referenced article is indeed excellent. However, it really contains no surprising information. It just reinforces a wide body of earlier research findings that grow more compelling with each year’s additional data.
Mutual fund performance persistence is largely a myth, with a few outstanding exceptions. The game plan should be to find and stick with those noteworthy exceptions. The article identifies a few of them.
In physics, the law of gravity dominates. Although not quite as universal, within the investment community, a regression-to-the-mean law seems almost as pervasive.
The evidence against fund persistent outperformance is overwhelming. In its early years, the Morningstar 5-star rating system proved so vulnerable to performance decay, and subsequently to star erosion, that the firm was forced to modify and expand its ranking methods to salvage its fading reputation.
The numerous periodic table of returns published annually by a host of industry giants demonstrate the fragile nature of last year’s winners. The rotations are swift. Buying last year’s top performers has proven to be a Loser’s game.
Each year, Fortune magazine issues a list of the most respected and the most despised companies. The list does have some stability year over year. However, the market returns from owning the most accomplished companies over the long haul do not match those yielded by the least honored outfits. From an investment rewards perspective, this is yet another illustration of a regression-to-the-mean pull.
All MFO regulars should be familiar with the often referenced Standard and Poor’s SPIVA and Persistency reports. These documents have registered fund persistency failures for many years. Here is the Link to those data rich reports:
http://us.spindices.com/resource-center/thought-leadership/spiva/
If motivated, you can click on the specific SPIVA and Persistency items to explore more deeply. The SPIVA mid-year report was just released. I have not yet accessed it.
Michael Mauboussin has written extensively about the tradeoffs between luck and skill in determining outcomes. He concludes that as the skill level of professional money managers has escaladed over time, the luck component becomes the more dominant factor. There are fewer and fewer Excess Return opportunities. Hence, expect that persistent superior performance should degrade even more in the future and should be more difficult to identify.
That’s an imperfect signal that favors an Index-like approach. Institutions have recognized that signal, and are moving more aggressively in that direction. Perhaps we should too?
Best Wishes.
Is regression to the mean a law?
Been wondering about that lately.
And, which mean?
Category mean? Decade mean? Post WWII mean? US equity mean? Etc, etc.
Hope all is well.
c
Thank you for reading my post.
In brief exchanges such as yours, it is sometimes a challenge to interpret your intentions. Were you simply pulling my leg or were you asking serious questions? I’ll assume the serious purpose.
A reversion-to-the-mean is surely not a law in the same league as a physics law. I used the term just to draw a parallel between a hard science and a soft science. More properly, a regression-to-the-mean is probably better characterized as a statistical concept.
It is a way statisticians describe an outlier event that deviates from some average, expected outcome. In their earlier years, Kahneman and Tversky observed that when pilots performed either above or below their normal performance standard, subsequent flights reverted to their average skill levels. Instructors like to think it was their training guidance; the Behavioral scientists judged otherwise.
If a ball player scores 4 hits in a single game, it is highly likely that he will not repeat that exceptional output in the next game; he will display a regression-to-the-mean tendency. Likewise, if he hits at a 400 batting average for a month, it is doubtful that his hot hand will persist. Again, he will revert towards his historically normal batting average.
The time horizon for the regression-like pull is unspecified. When investing, it depends upon the investors specific timeframe. If he trades frequently, the comparison period must be representative of that short timeframe. If the investor is planning for the long-term, the representative “mean” value should adequately reflect a long time horizon.
My preference is for longer running data collection averages for comparative purposes. I recognize that this bias introduces data staleness risk. So some experimentation is needed to properly identify a meaningful data period to calculate a “Mean” value.
Sorry for wasting your time with this reply if you were just playing “a pulling of the leg” game. I’m sure you knew everything I just reviewed.
But did you know that Earth’s gravity is NOT precisely constant? It depends on location, ground water level, and shifts under tectonic plate activity. Social laws are temporal and uncertain. Even physics laws are approximations and sometimes infrequently revised. Change happens.
Best Wishes.
>> The referenced article is indeed excellent.
Omg, how the heck so? Seems extremely feeble to me, even by the feeble standards of WSJ and MW service pieces. The rules are lame, not really or demonstrably or consistently true, the quotes from Kinnel and Snowball highly selective (you can always find such an example), the absence of broad data lazy and troubling, and the whole thing a yawn and worse, useless, unhelpful, unactionable. 1 - Sure, some hot managers ain't no more. Maybe most. Oh, but wait, Danoff ahead. Oh, wait, he's longterm. 2 - Sure, GLRBX has been good forever, and you shoulda bought and held, period, full stop, you dope. 3 - Everyone has been writing forever about size, which matters absolutely, sure, yes, except when it doesn't, and most managers say it doesn't, but they are wrong, or are they? Wait, Danoff ahead. Good thing Tillinghast and D&C were overlooked. 4 - KIS, okay, so indexing is best. Or is it? And cherish sc and value, oky-doky and all righty, then. Twas ever thus. 5 - duh. 6 - Balance and blend and mix are important. Wow. And 7, the last rule, the future, well, read those last three questions out loud and see if you can do it without laughing.
The whole thing is a weak, but actually worse than weak, September yawn, to my eye. I worry when smart people here reflexively write that something is good or excellent. Is there anything whatsoever in this article that would not have occurred to any regular visitor to this forum??
Underneath the emotion, I like to think that markets are driven by fundamentals.
And fundamentals are driven by people, ideas/innovation, resources/capital, legal system, etc.
So different times and different places will warrant different medians. The normal changes, like you say.
Folks that have a good sense of such changes likely make good active investors, even if they are only active with allocation.
talented and hardworking individuals. However, we (meaning both the experts and
their followers) tend to overestimate our skills. Long term trend is still an average
trend based on overall fundamentals (e.g., overall economic growth).
There is some statistical distribution of performance around this average trend
and it is very hard to show that persistence in performance
is due to skill or a consequence of perceived patterns in a underlying randomly
distributed outcomes (kind of like a sequence of five heads in a coin toss).
We are all attempting to tilt odds in our favor by studying the art, understanding
our behavioral limitations, keeping things simple enough to manage, be humble to
change when our views turn out to be wrong and have the courage to take manageable risks when we are convinced about our understanding.
When you questioned the appropriateness of calling the investment world’s regression-to-the-mean observed phenomena a law, it did prompt me to do some deeper thinking on the matter. Well, that puts us into some dangerous waters.
Certainly investment wisdom and operational rules assembled over decades of practical experience deserve some elevated status when they address investment process and not investment outcomes. Probably assigning these assembled rules the title of “laws” might be a “bridge-too-far”. Why did I do it?
In thinking about my readings over the last month, I recalled a White Paper produced by the respectable GMO investment house. It was written by James Montier, Montier titled his work “The Seven Immutable Laws of Investing”. Eureka! That was likely my inspiration. Here is a list of his laws, immutable no less:
1. Always insist on a margin of safety
2. This time is never different
3. Be patient and wait for the fat pitch
4. Be contrarian
5. Risk is the permanent loss of capital, never a number
6. Be leery of leverage
7. Never invest in something you don’t understand
There’s a ton of Warren Buffett and Benjamin Graham embedded in these “laws”. I believe they serve as excellent generic investment guidance. Maybe they do rise to the level of laws.
If you are interested, here is a Link to the complete White Paper:
http://conferences.pionline.com/uploads/conference_admin/JM_0311.pdf
These rules do emphasize investment process over investment outcomes. A nice way to illustrate the tension between process and outcomes is to visualize a 2 X 2 matrix with Process on the vertical axis and Outcomes on the horizontal axis. Both dimensions can have either a good or a bad result.
The 2 X 2 boxes can be filled with the following conclusions: Good process/good outcome is an earned reward, good process/bad outcome is unlucky, bad process/good outcome is lucky and is likely a misleading signal towards overconfidence, and bad process/bad outcome is true justice.
The key to continued long-term investment success is a solid Process that is NOT abandoned when the markets turn South. It is amazing how many investors are lured by the Sirens' song of a lucky outcome. The likelihood of any future success with a poor investment Process is dubious at best.
Enough of this serious stuff for now. I too liked Kaspa's submittal; good stuff, very well summarized.
Best Wishes to All.