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.
A very important article. If the conclusions are the article are incorrect, how are they incorrect? If they are correct, why do we invest in active funds?
@dryflower, the studies along these lines do not model the reality correctly and so come up with incorrect conclusions when they generalize it.
Imagine, if you were to conduct a similar study of multiple choice test scores of all high school students over a number of tests with no selection criterion, you may conclude that the test scores across all students were determined more by luck than skill because random answering shows similar distribution. Or if there was skill that you could not reliably select students that had skill or that they showed no persistence or if it did it was because those students studied more broadly than others (took more risk).
Studies that lie behind this indexing cult usually follow the same cyclic argument as follows, including the study linked in the other thread.
1. In the aggregate over all funds, the active funds do not perform over index funds.
But wait, you say, I do not select funds by throwing darts and there are a lot of bad funds. So what if I select the good funds?
2. Even if you were somehow able to select good funds they do not show persistence over time. So, it is futile.
But wait, you say, you require the fund to beat the index every year or in two consecutive periods and consider it a fail even if the fund underperforms in one period by a mere 0.5% while it gained by 2% in the previous period. So, it isn't necessary to be persistent in YOUR definition for the fund to do well over time.
3. But active funds in general do not beat indices over time. See 1 above.
See the cyclic argument?
I am waiting for somebody to do a simple study. Find the cumulative performance over a reasonable period of time of active funds selected with multiple criterion available at the beginning of the period and see if any of the criterion select for over performing funds with statistical significance. This is what models reality better. If none of the selection criterion comes out with a reliable way to select a fund, then there is a good case to make against choosing active funds
Because reality is complex, choosing a mathematical or statistical model to draw conclusions from is not trivial and requires some simplifying assumptions. That can lead to incorrect conclusions in the general case.
As a very simple example of incorrect modeling leading to incorrect conclusions, consider @mjg's charming anecdote in another thread of using probabilities to make shooting decisions in basketball (not to pick on @mjg here).
On the surface, it looks perfectly reasonable. The math is correct. So, the conclusion of going for 3 pointers should be valid, right? It might be, if the opposing coach is a total idiot as might happen in a junior team of 12 year olds. However, it does not apply in general because the reality is different for that simple modeling to apply.
When you have two opposing teams making strategic decisions in a zero sum game, it needs to be modeled with game theory rather than simple probability to reflect what happens in reality better. Game theory applied to this suggests that the fixed point equilibrium reduces the efficacy of the 3 pointers to a level where it provides no advantage over a 2 point attempt.
The plain English translation is that, if the 3 point tries start to win with higher probability in successive games, the opposing team will start to guard against 3 pointers more, which leaves them vulnerable to 2 point attacks and so a higher probability doing that, etc., until an equilibrium is reached.
So real world coach decisions are much more complex relying on luck and skill to outplay the other teams based on the players, how they are playing at that moment, probability of refs in that game to be lenient towards fouls in that game, the strengths or weaknesses of the opposing coach, etc. Good coaches do this by intuition and aren't helped by a probability calculator to consult. It is not just luck either.
Good fund managers and good coaches have a lot of things in common.
@cman, thank you!!! I was feeling somewhat depressed after reading Sam Lee's article (and its funny but I feel that way a lot after I read his articles) and you've cheered me up. I felt even better when I did a little research and found out Lee's newsletter, with an admittedly short history, has performed miserably compared to the Wilshire 5000 on a risk adjusted basis and even worse when unadjusted. Maybe his writing just helps to make him feel better about his own lack of both skill and luck.
So is it the unorthodox consensus around here that in fact it is not all that difficult to choose funds that will, on average, beat a relevant index over time going forward?
That is what I seem to believe, what seems to be true, and what has been true for me in experience. The common wisdom characterizes that as a delusion.
So is it the unorthodox consensus around here that in fact it is not all that difficult to choose funds that will, on average, beat a relevant index over time going forward?
I can't speak for others. But the impression I've always gotten is that Prof. Snowball and the board in general try to keep the human element in mind when talking about funds and investing. That is, people have different experiences, temperaments, and life situations that lead them to different investing goals. Beating an index may be one of those goals, but it might not be. By boiling down everything to that equation you risk square pegs in round holes and investors shooting themselves in the foot.
At the risk of putting words in @cman's mouth (not that I could), I don't believe his objection is to indexing, per se, but to a blind adherence to one style of investing without consideration of the investor. He also objects to the idea that investor decisions are like throwing darts blindly at a board. People know to avoid the stinkers.
I don't think he's making a conclusion about beating an index or not, just suggesting that reflexively trumpeting the Boglehead line isn't thoughtful, can reflect bad methodology, and ultimately can hurt investors.
So is it the unorthodox consensus around here that in fact it is not all that difficult to choose funds that will, on average, beat a relevant index over time going forward?
That is what I seem to believe, what seems to be true, and what has been true for me in experience. The common wisdom characterizes that as a delusion.
It is difficult to state what is a consensus and what is orthodox and unorthodox. They are subjective.
However, there is significant personal empirical evidence that it seems possible to pick a good actively managed fund that performs very well over time. Most people here would probably find a lot of these in their portfolios. But this appears to be in stark contradiction to the claims made by index fund proponents in the studies they cite.
It would be a useful intellectual activity to resolve this apparent contradiction. This is what science tries to do when it finds empirical data that doesn't fit a theory. You look to see possible explanations including the assumptions in the studies and the model used.
There are obvious flaws in the model used in these studies to reflect the world in which the contradictory empirical data exists. So they may not be contradictory at all. So, it would be useful to design an experiment that would actually test the conditions in which the empirical data is observed.
That is to test for performance using the selection criteria that people use rather than over the entire fund world. Prove or disprove the hypothesis that one or more selection criteria helps people find active fund managers that beat indices over time using the metric that matters, not some artificial criterion like persistence every year or every period.
Very good job, cman. It's simple arithmetic that the market returns what the market returns minus the costs of investing in the market. So if mutual funds are representative of some market then they must underperform that market by the amount of their costs. These include trading costs. It doesn't take any studies to demonstrate that, but all those 'publish or perish' academics have to do something, I guess.
It's clear that funds which shadow the market, 'Closet Indexers' they're called, must necessarily underperform in aggregate, and those that outperform are doomed to do so by a tiny margin. Only by being very different from the market (measured these days by 'Active Share') do you have a chance to outperform. However, the penalty for losing money as a 'Closet Indexer' is...well, there isn't any penalty for it. You just shrug your shoulders, say that everything was down, it was a tough year and the key is for the investor to show patience, not give up, the market always comes back, etc.
The penalty for losing money when you're being different from the market is being fired from your position as an idiot for not knowing what everyone knew. You can see why there's so much closet indexing, i.e., sure underperforming, going on, and so little real attempt to outperform the market.
Another problem with repeated success is that it increases your assets under management (AUM). As you get bigger, more and more small investments become out of your reach. For example, if a huge fund like DODGX finds a small cap company that is a great investment, so what? DODGX is so big that it can't buy enough of the small company to make any significant difference in its performance. Thus, the more successful a fund is, the more difficult it becomes to outperform the market because more and more opportunities simply become unavailable to it.
Not to speak for David Snowball, but these are two great reasons to concentrate on small boutique firms. If you own the company, you're not going to fire yourself because things go wrong one year (e.g., Intrepid Funds or FPA), and if you have the discipline you can close the fund before it grows out of its successful strategy (Grandeur Peaks being a great example).
I'll also add that these studies must be talking about risk-adjusted returns since you can always beat an index that goes up by simply layering on more and more leverage to it. Unfortunately, 'risk' is a difficult if not impossible idea to specify with precision. For my money, and to paraphrase David Hume, it takes a helluva lot of education to believe that the 'Value Effect' is one of increased risk, but each to his own.
>> if a huge fund like DODGX finds a small cap company that is a great investment, so what? DODGX is so big that it can't buy enough of the small company to make any significant difference in its performance. Thus, the more successful a fund is, the more difficult it becomes to outperform the market
One reads this an awful lot, which doesn't make it untrue, but Tillinghast sure does give the lie to it much more often than not, as does, often, Danoff.
>> if a huge fund like DODGX finds a small cap company that is a great investment, so what? DODGX is so big that it can't buy enough of the small company to make any significant difference in its performance. Thus, the more successful a fund is, the more difficult it becomes to outperform the market
One reads this an awful lot, which doesn't make it untrue, but Tillinghast sure does give the lie to it much more often than not, as does, often, Danoff.
Those two have been wonderful fund managers, but looking at Morningstar I note that they both have had sharply falling relative performances. At 15 years they both crush their benchmarks, at 10 years they win handily but not by nearly so wide a margin, and at 5 years Tillinghast still wins clearly but not by so much as before while Danoff trails the S&P 500 by a touch. And Tillinghast used to run that as a small cap fund; not anymore.
Right, and FLPSX at 3y too, very slight superiority, depending on which midcap ETF you look at. My point was only that he finds ways to add (perhaps decreasingly; would probably not call it 'sharply') value with a huge $48b fund, 85% the size of DODGX.
Comments
Directly under an interview with David Herro as part of "Beat the Market Week."
Imagine, if you were to conduct a similar study of multiple choice test scores of all high school students over a number of tests with no selection criterion, you may conclude that the test scores across all students were determined more by luck than skill because random answering shows similar distribution. Or if there was skill that you could not reliably select students that had skill or that they showed no persistence or if it did it was because those students studied more broadly than others (took more risk).
Studies that lie behind this indexing cult usually follow the same cyclic argument as follows, including the study linked in the other thread.
1. In the aggregate over all funds, the active funds do not perform over index funds.
But wait, you say, I do not select funds by throwing darts and there are a lot of bad funds. So what if I select the good funds?
2. Even if you were somehow able to select good funds they do not show persistence over time. So, it is futile.
But wait, you say, you require the fund to beat the index every year or in two consecutive periods and consider it a fail even if the fund underperforms in one period by a mere 0.5% while it gained by 2% in the previous period. So, it isn't necessary to be persistent in YOUR definition for the fund to do well over time.
3. But active funds in general do not beat indices over time. See 1 above.
See the cyclic argument?
I am waiting for somebody to do a simple study. Find the cumulative performance over a reasonable period of time of active funds selected with multiple criterion available at the beginning of the period and see if any of the criterion select for over performing funds with statistical significance. This is what models reality better. If none of the selection criterion comes out with a reliable way to select a fund, then there is a good case to make against choosing active funds
Because reality is complex, choosing a mathematical or statistical model to draw conclusions from is not trivial and requires some simplifying assumptions. That can lead to incorrect conclusions in the general case.
As a very simple example of incorrect modeling leading to incorrect conclusions, consider @mjg's charming anecdote in another thread of using probabilities to make shooting decisions in basketball (not to pick on @mjg here).
On the surface, it looks perfectly reasonable. The math is correct. So, the conclusion of going for 3 pointers should be valid, right? It might be, if the opposing coach is a total idiot as might happen in a junior team of 12 year olds. However, it does not apply in general because the reality is different for that simple modeling to apply.
When you have two opposing teams making strategic decisions in a zero sum game, it needs to be modeled with game theory rather than simple probability to reflect what happens in reality better. Game theory applied to this suggests that the fixed point equilibrium reduces the efficacy of the 3 pointers to a level where it provides no advantage over a 2 point attempt.
The plain English translation is that, if the 3 point tries start to win with higher probability in successive games, the opposing team will start to guard against 3 pointers more, which leaves them vulnerable to 2 point attacks and so a higher probability doing that, etc., until an equilibrium is reached.
So real world coach decisions are much more complex relying on luck and skill to outplay the other teams based on the players, how they are playing at that moment, probability of refs in that game to be lenient towards fouls in that game, the strengths or weaknesses of the opposing coach, etc. Good coaches do this by intuition and aren't helped by a probability calculator to consult. It is not just luck either.
Good fund managers and good coaches have a lot of things in common.
So is it the unorthodox consensus around here that in fact it is not all that difficult to choose funds that will, on average, beat a relevant index over time going forward?
That is what I seem to believe, what seems to be true, and what has been true for me in experience. The common wisdom characterizes that as a delusion.
At the risk of putting words in @cman's mouth (not that I could), I don't believe his objection is to indexing, per se, but to a blind adherence to one style of investing without consideration of the investor. He also objects to the idea that investor decisions are like throwing darts blindly at a board. People know to avoid the stinkers.
I don't think he's making a conclusion about beating an index or not, just suggesting that reflexively trumpeting the Boglehead line isn't thoughtful, can reflect bad methodology, and ultimately can hurt investors.
However, there is significant personal empirical evidence that it seems possible to pick a good actively managed fund that performs very well over time. Most people here would probably find a lot of these in their portfolios. But this appears to be in stark contradiction to the claims made by index fund proponents in the studies they cite.
It would be a useful intellectual activity to resolve this apparent contradiction. This is what science tries to do when it finds empirical data that doesn't fit a theory. You look to see possible explanations including the assumptions in the studies and the model used.
There are obvious flaws in the model used in these studies to reflect the world in which the contradictory empirical data exists. So they may not be contradictory at all. So, it would be useful to design an experiment that would actually test the conditions in which the empirical data is observed.
That is to test for performance using the selection criteria that people use rather than over the entire fund world. Prove or disprove the hypothesis that one or more selection criteria helps people find active fund managers that beat indices over time using the metric that matters, not some artificial criterion like persistence every year or every period.
It's clear that funds which shadow the market, 'Closet Indexers' they're called, must necessarily underperform in aggregate, and those that outperform are doomed to do so by a tiny margin. Only by being very different from the market (measured these days by 'Active Share') do you have a chance to outperform. However, the penalty for losing money as a 'Closet Indexer' is...well, there isn't any penalty for it. You just shrug your shoulders, say that everything was down, it was a tough year and the key is for the investor to show patience, not give up, the market always comes back, etc.
The penalty for losing money when you're being different from the market is being fired from your position as an idiot for not knowing what everyone knew. You can see why there's so much closet indexing, i.e., sure underperforming, going on, and so little real attempt to outperform the market.
Another problem with repeated success is that it increases your assets under management (AUM). As you get bigger, more and more small investments become out of your reach. For example, if a huge fund like DODGX finds a small cap company that is a great investment, so what? DODGX is so big that it can't buy enough of the small company to make any significant difference in its performance. Thus, the more successful a fund is, the more difficult it becomes to outperform the market because more and more opportunities simply become unavailable to it.
Not to speak for David Snowball, but these are two great reasons to concentrate on small boutique firms. If you own the company, you're not going to fire yourself because things go wrong one year (e.g., Intrepid Funds or FPA), and if you have the discipline you can close the fund before it grows out of its successful strategy (Grandeur Peaks being a great example).
I'll also add that these studies must be talking about risk-adjusted returns since you can always beat an index that goes up by simply layering on more and more leverage to it. Unfortunately, 'risk' is a difficult if not impossible idea to specify with precision. For my money, and to paraphrase David Hume, it takes a helluva lot of education to believe that the 'Value Effect' is one of increased risk, but each to his own.
One reads this an awful lot, which doesn't make it untrue, but Tillinghast sure does give the lie to it much more often than not, as does, often, Danoff.