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.
FYI: Larry's series evaluating the performance of the market’s most prominent actively managed mutual fund families continues today with an in-depth analysis of the offerings from Boston-based Grantham Mayo van Otterloo (GMO). Regards, Ted (Click On PDF Upper Left For Easier Reading): http://www.advisorperspectives.com/articles/2016/02/25/does-gmo-add-value-for-investors
This is all well and good...I suppose. But he needs to taken into account the volatility (standard deviation) in performance, not just the mean. It's not hard to outperform on a total value basis if your mean comes in lower than the benchmark if your volatility is lower than the benchmark's. I find it disappointing that a financial advisor does not appear to explicitly account for volatility in his/her analysis.
Also, not sure what he's doing with his factor analysis stuff. Maybe someone can explain that to me. I've used factor analysis plenty in my day job, but am not sure how he's using it here.
I write all of this, BTW, as someone who does not have a horse in the GMO race (although I enjoy their quarterly letters).
I understand Larry's take on all of this, but the real issue is getting clients to stay in these funds, DFA or not, and consistently invest in them in all market conditions. Although, concentrated and contrarian strategies can often lose too much to the point of struggling to ever recover.
I see your "time in market" argument as reinforcing my point. A lower volatility fund may reduce investor anxiety and increase time in market.
So, either way one slices it, I really think volatility / standard deviation is an important missing piece here. An active manager who underperforms based on a simple comparison of averages, can outperform in real dollars if he/she helps investors avoid a major market collapse. I think in part that is what GMO and other active managers try to do (again, I do not invest with GMO, but actually do hold two DFA funds via an employer plan).
Bottom line: Larry's point may stand even after taking volatility into account. Regardless, simply rolling out another "let's look at the averages" comparison is as underwhelming as he claims GMO is, if for no other reason than it reinforces poor checkoutcounter-magazine-style comparisons. I mean, really: we're going to draw factor analysis into this, but avoid looking at volatility and upside/downside market capture?
As President James A. Garfield noted: “Man cannot live by bread alone; he must have peanut butter”.
Similarly, an investor cannot survive on a simple one-dimensional model to capture stock market movements like CAPM is, but a fuller explanation demands a more eloquent model, perhaps like the Fama-French 3-factor formulation. The modeling became even more complex when a Momentum component was added, and still more complex when Fama-French contributed yet more elements.
The initial CAPM (Capital Asset Pricing Model) was developed by Bill Sharpe who received a Nobel award for his inspiration. Beta was his modeling parameter of choice. Later he ruefully said that he is fortunate that the Nobel community does not rescind their awards.
Fama and French expanded the pricing modeling to three parameters: Beta, Size differences, and Price-to-Book ratios. Momentum was added by Carhart to transform the three factor into a four factor model. Models are just simplifications and are never perfect. Neither was the four factor model, so Fama-French added more elements.
The goal was to develop a complete model that fully projected stock returns with a minimum of residual unexplained component. That residual (Alpha) can be interpreted as a fund manager’s skill or luck contribution when evaluating mutual fund performance. If the residual Alpha remains positive over an extended timeframe, then skill is the likely interpretation; if it does not persist, luck is the most probable interpretation.
The accumulated empirical mutual fund evidence is that luck is the predominant factor, but some limited cases of skill are identifiable. The Swedroe work finds skill within the American and Vanguard fund groups. The Vanguard finding is consistent with a claim reported in a recent MFO posting. Here is an Internal Link to that article:
By way of full disclosure, I own a substantial number of Vanguard products, both passively and actively managed.
I agree with your basic observation that the research would be better served with some acknowledgement of standard deviations. That’s always a necessary element when reporting statistical work.
When generating an Alpha for a fund, the standard deviation of that Alpha for an extended time period is also needed to formulate a test of its significance. The Swedroe paper did not include that data. The Alpha standard deviations are needed to calculate a “t” test value which is a measure of how meaningful the statistical results really are.
A “t” test value of plus or minus 2 or greater is the usual criterion for a statistically meaningful finding. These were likewise missing in action from the Swedroe document. Still, I congratulate him for his project. I concluded that his work is useful. It tells us where to search if Alpha is a singular selection criterion. It is not so for me.
I hope this is helpful. Please understand that I’m not a statistician, just an investor with an interest in statistical data.
@MJG: your points on n-factor models are well taken. It would be interesting to compare fit statistics and information criteria for each of the models for each of the funds to better determine if funds' all have the same underlying factor structure. Which model is correct? Is the correct model the "correct" all the time? Leave that as it is for another time. There's also a methodological issue in terms of getting the random variable distributions of the observed indicators correct. Without that, the factor analysis results are apt to be biased, and statistical inferences on parameters from the factor solution are apt to be incorrect. Same, of course, goes for a t-test, etc. Would the biases be consistent across funds/comparisons? Hard to say. But, leave that as it is. There is also the question of how appropriate linear or fully parametric models (such as factor analysis) are for something as opaque as financial markets and asset performances. Perhaps a parametric model should explicitly account for trending (is "momentum" sufficient) in a different manner. Again, leave that as it is.
Because, what I really want to know is -- in terms of actual, realized, dollar value returns -- what is GMO's performance when we factor in their contribution to risk reduction. I would think that is what GMO with their "2-sigma" logic, etc., is trying to do: not just pick attractive and/or undervalued assets per se, but avoid market bubbles. This is how I often hear Grantham speaking of GMO's mission. Here I think a different accounting for volatility is important before dismissing what GMO does or doesn't do. And it doesn't have to be all that complicated, but simply comparing means to means doesn't account for this aspect of what GMO attempts to do (perhaps adding "momentum" to a factor analysis does capture this in some tangential way; I don't think so, but its possible). I would think an analyst such as Larry savvy enough to wax poetic about factor analysis would know this. For example, it is quite possible for Fund A to have a lower mean annualized return than Fund B, but outperform Fund B, if it has lower volatility. As Buffett says: "Rule #1 is to not lose money; Rule #2 is to not forget Rule #1". Who cares if the market index has a higher average return but takes costly bites out of your principle along the way? That's really the point of my critique. If GMO lagged the market average by 0.5% but reduced investors' exposure to the 2001 or 2008 downturns, they would have added significant value that would likely not be captured in simple mean comparisons or (as I understand it) an n-factor model.
'Let the jury consider their verdict,' the King said, for about the twentieth time that day.
'No, no!' said the Queen. 'Sentence first - verdict afterwards.'
'Stuff and nonsense!' said Alice loudly. 'The idea of having the sentence first!'
'Hold your tongue!' said the Queen, turning purple.
Mr. Swedroe is a marketer. His title, Director of Research, strikes me as modestly misleading since we often associate that title with someone leading the stock-by-stock, market-by-market analytics effort. BAM's official description of his role is this: "Larry regularly reviews the findings published in dozens of peer-reviewed financial journals, evaluates the outcomes and uses the result to inform the firm’s formal investment strategy recommendations." What does that mean? The most recent research I read concluded that frequent portfolio disclosure depresses performance by facilitating front-running of the fund. Okay, and therefore ... ? Several recent studies found that most active funds don't outperform their benchmark's raw returns. Uh-huh. "Finding successful funds ex-ante is extremely difficult." Someone got published for that nugget. At base, Larry's conclusion was reached a long time ago, BAM is deeply invested in it, and I'm not sure what he's actually contributing. The nature of peer-reviewed articles is that they aren't timely, so it's unlikely that the answer to the question "what on earth are the Chinese doing?" will be found there.
On the risk-adjustment stuff, I looked at GMO's five largest funds.
Benchmark-Free Allocation (GMBFX) is largest and it's beaten its peers since inception with dramatically higher returns (about 300 bps/year) and lower risk (a downside deviation of 4.4 versus 7.2 for its peers). It's Sharpe ratio is 0.99 versus 0.43 for the peers.
International Equity (GMOIX) is second largest and it's beaten its peers since inception with modestly higher returns (about 120 bps) and lower risk (DSDEV 11.2 vs 12.5), resulting in a higher Sharpe (0.24 versus 0.16).
Quality (GQEFX) is third, a Great Owl (i.e., a consistent winner), with higher returns (40 bps), lower DSDEv (7.8 versus 10.3) and higher Sharpe (0.44 versus 0.32).
Emerging Markets (GMOEX) is fourth. It has better performance (180 bps), very slightly higher volatility (16.9 versus 16.8) and a higher Sharpe (0.14 versus 0.06).
US Equity Allocation (GMUEX) is fifth. Better performance (90 bps), lower vol (9.8 versus 10.6) and higher Sharpe (0.51 versus 0.42).
If you play with the comparison groups or the time frames, you can substantially change the results. That said, most of these funds lead most of their peers, mostly with lower volatility, over most full market cycles.
Hi David, I just glanced at M*, and for the last ten years, GBMFX has returned 4.82% annually, beating its category by a highly respectable 1.03%. http://www.morningstar.com/funds/XNAS/GBMFX/quote.html But that's the institutional share class, with a $10 million minimum. I presume that the higher fees for non-institutional investors will eat up some of that alpha, and then it's very inefficient taxwise (1.78% tax cost) so unless you're in a tax-sheltered account, there goes the rest of that alpha and then some. Vanguard's balanced index fund, VBINX, turned in a 5.99% annual return over that ten year period, with a 0.80% tax cost. In a taxable account, the annual return would be nearly double GBMFX, and even in a tax-sheltered account it would be superior. And of course there's no manager-transition risk, concern about luck, etc. I don't know about the since-inception results, maybe in the 90s GBMFX had glory years, but MJG's thoughts -- that even funds run by brilliant, ethical, humble, and careful managers like Mr. Grantham have a hard time outperforming -- are becoming ever more convincing.
@expatsp: you're still ignoring volatility. You do yourself a potentially great disservice if you simply consider mean annualized returns. This is especially important for firm such as GMO, whose strategy and value-add is (ostensibly) that they seek to avoid bubbles rather than simply shoot the lights out year to year.
@shostakovich. A fair point, sir. It indeed seems that actively-managed funds can successfully offer lower volatility in return for lower returns, and for retirees, that is probably a good option.
But since over a ten year period -- one that included the biggest housing bubble and market crash since the Great Depression -- a simple balanced index fund outperformed GMO's flagship fund on an aftertax basis by an average of 226 basis points a year, and on a pretax basis by 117 basis points a year (and that's assuming you were lucky enough to have $10 million to buy the cheaper, institutional share class), I think that for most people with at least 10 years till retirement, the index fund is the better bet.
But that investor will have to close his eyes and even add more if possible during the inevitable downturns. Not every investor can do that.
@expatsp: agree. The 2008 downturn was so severe, that many of the conservative, deep value stalwarts took a major hit. Since GMO recognized that a 3-sigma was on the horizon, they should have been *especially* prone to go ultra-defensive. But in a severe crash everyone wants out of everything, which leaves little place to hide.
I think we are on the edge of another bubble at some point in the next 1-3 years. I wonder if, for younger investors who have lived through the GR, and now another bubble, if they won't avoid the market altogether. Problem is, many are also very wary of owning property as well (and/or have student debt). Where will they invest for the future? Will they simply spend (buy dividend-stalwart consumer stocks !!!)?
We are "blessed" to be living in interesting times.
My parents caught some of the dot-com boom; but otherwise enjoyed what in retrospect seems to have been a very good market throughout their investing careers. I don't think many retail investors investing today will enjoy such good fortunes.
I was ready to whip out my checkbook to purchase a few shares of a GMO fund when I realized I was a tad short of the membership entry fee this month. Too, too bad. I suppose a few other MFOers also find themselves in that situation.
I was not unfamiliar with the stiff entry requirements that GMO directly imposes on potential customers.
Given the likely paucity of a large general clientele base, I’m somewhat puzzled by Jeremy Grantham’s frequent talks at public events like the MoneyShow. I benefited from his global insights on several occasions at such presentations. He is an impressive, controlled speaker who inspires trust and confidence. Wallets open when Grantham speaks.
After one such presentation, I did seek an alternate source of his funds working through Wells Fargo. That was a few years ago. The initial purchasing requirements were very modest, but I rejected the opportunity because of a heavy initial purchase fee. I still blanch from such ruthless and completely unnecessary charges. I consistently stay away because of that fee structure.
Comments
Also, not sure what he's doing with his factor analysis stuff. Maybe someone can explain that to me. I've used factor analysis plenty in my day job, but am not sure how he's using it here.
I write all of this, BTW, as someone who does not have a horse in the GMO race (although I enjoy their quarterly letters).
So, either way one slices it, I really think volatility / standard deviation is an important missing piece here. An active manager who underperforms based on a simple comparison of averages, can outperform in real dollars if he/she helps investors avoid a major market collapse. I think in part that is what GMO and other active managers try to do (again, I do not invest with GMO, but actually do hold two DFA funds via an employer plan).
Bottom line: Larry's point may stand even after taking volatility into account. Regardless, simply rolling out another "let's look at the averages" comparison is as underwhelming as he claims GMO is, if for no other reason than it reinforces poor checkoutcounter-magazine-style comparisons. I mean, really: we're going to draw factor analysis into this, but avoid looking at volatility and upside/downside market capture?
Color me unimpressed.
As President James A. Garfield noted: “Man cannot live by bread alone; he must have peanut butter”.
Similarly, an investor cannot survive on a simple one-dimensional model to capture stock market movements like CAPM is, but a fuller explanation demands a more eloquent model, perhaps like the Fama-French 3-factor formulation. The modeling became even more complex when a Momentum component was added, and still more complex when Fama-French contributed yet more elements.
The initial CAPM (Capital Asset Pricing Model) was developed by Bill Sharpe who received a Nobel award for his inspiration. Beta was his modeling parameter of choice. Later he ruefully said that he is fortunate that the Nobel community does not rescind their awards.
Fama and French expanded the pricing modeling to three parameters: Beta, Size differences, and Price-to-Book ratios. Momentum was added by Carhart to transform the three factor into a four factor model. Models are just simplifications and are never perfect. Neither was the four factor model, so Fama-French added more elements.
The goal was to develop a complete model that fully projected stock returns with a minimum of residual unexplained component. That residual (Alpha) can be interpreted as a fund manager’s skill or luck contribution when evaluating mutual fund performance. If the residual Alpha remains positive over an extended timeframe, then skill is the likely interpretation; if it does not persist, luck is the most probable interpretation.
The accumulated empirical mutual fund evidence is that luck is the predominant factor, but some limited cases of skill are identifiable. The Swedroe work finds skill within the American and Vanguard fund groups. The Vanguard finding is consistent with a claim reported in a recent MFO posting. Here is an Internal Link to that article:
http://www.mutualfundobserver.com/discuss/discussion/26202/finding-active-managers-to-beat-the-market
By way of full disclosure, I own a substantial number of Vanguard products, both passively and actively managed.
I agree with your basic observation that the research would be better served with some acknowledgement of standard deviations. That’s always a necessary element when reporting statistical work.
When generating an Alpha for a fund, the standard deviation of that Alpha for an extended time period is also needed to formulate a test of its significance. The Swedroe paper did not include that data. The Alpha standard deviations are needed to calculate a “t” test value which is a measure of how meaningful the statistical results really are.
A “t” test value of plus or minus 2 or greater is the usual criterion for a statistically meaningful finding. These were likewise missing in action from the Swedroe document. Still, I congratulate him for his project. I concluded that his work is useful. It tells us where to search if Alpha is a singular selection criterion. It is not so for me.
I hope this is helpful. Please understand that I’m not a statistician, just an investor with an interest in statistical data.
Best Wishes.
Because, what I really want to know is -- in terms of actual, realized, dollar value returns -- what is GMO's performance when we factor in their contribution to risk reduction. I would think that is what GMO with their "2-sigma" logic, etc., is trying to do: not just pick attractive and/or undervalued assets per se, but avoid market bubbles. This is how I often hear Grantham speaking of GMO's mission. Here I think a different accounting for volatility is important before dismissing what GMO does or doesn't do. And it doesn't have to be all that complicated, but simply comparing means to means doesn't account for this aspect of what GMO attempts to do (perhaps adding "momentum" to a factor analysis does capture this in some tangential way; I don't think so, but its possible). I would think an analyst such as Larry savvy enough to wax poetic about factor analysis would know this. For example, it is quite possible for Fund A to have a lower mean annualized return than Fund B, but outperform Fund B, if it has lower volatility. As Buffett says: "Rule #1 is to not lose money; Rule #2 is to not forget Rule #1". Who cares if the market index has a higher average return but takes costly bites out of your principle along the way? That's really the point of my critique. If GMO lagged the market average by 0.5% but reduced investors' exposure to the 2001 or 2008 downturns, they would have added significant value that would likely not be captured in simple mean comparisons or (as I understand it) an n-factor model.
'No, no!' said the Queen. 'Sentence first - verdict afterwards.'
'Stuff and nonsense!' said Alice loudly. 'The idea of having the sentence first!'
'Hold your tongue!' said the Queen, turning purple.
Mr. Swedroe is a marketer. His title, Director of Research, strikes me as modestly misleading since we often associate that title with someone leading the stock-by-stock, market-by-market analytics effort. BAM's official description of his role is this: "Larry regularly reviews the findings published in dozens of peer-reviewed financial journals, evaluates the outcomes and uses the result to inform the firm’s formal investment strategy recommendations." What does that mean? The most recent research I read concluded that frequent portfolio disclosure depresses performance by facilitating front-running of the fund. Okay, and therefore ... ? Several recent studies found that most active funds don't outperform their benchmark's raw returns. Uh-huh. "Finding successful funds ex-ante is extremely difficult." Someone got published for that nugget. At base, Larry's conclusion was reached a long time ago, BAM is deeply invested in it, and I'm not sure what he's actually contributing. The nature of peer-reviewed articles is that they aren't timely, so it's unlikely that the answer to the question "what on earth are the Chinese doing?" will be found there.
On the risk-adjustment stuff, I looked at GMO's five largest funds.
Benchmark-Free Allocation (GMBFX) is largest and it's beaten its peers since inception with dramatically higher returns (about 300 bps/year) and lower risk (a downside deviation of 4.4 versus 7.2 for its peers). It's Sharpe ratio is 0.99 versus 0.43 for the peers.
International Equity (GMOIX) is second largest and it's beaten its peers since inception with modestly higher returns (about 120 bps) and lower risk (DSDEV 11.2 vs 12.5), resulting in a higher Sharpe (0.24 versus 0.16).
Quality (GQEFX) is third, a Great Owl (i.e., a consistent winner), with higher returns (40 bps), lower DSDEv (7.8 versus 10.3) and higher Sharpe (0.44 versus 0.32).
Emerging Markets (GMOEX) is fourth. It has better performance (180 bps), very slightly higher volatility (16.9 versus 16.8) and a higher Sharpe (0.14 versus 0.06).
US Equity Allocation (GMUEX) is fifth. Better performance (90 bps), lower vol (9.8 versus 10.6) and higher Sharpe (0.51 versus 0.42).
If you play with the comparison groups or the time frames, you can substantially change the results. That said, most of these funds lead most of their peers, mostly with lower volatility, over most full market cycles.
David
I just glanced at M*, and for the last ten years, GBMFX has returned 4.82% annually, beating its category by a highly respectable 1.03%. http://www.morningstar.com/funds/XNAS/GBMFX/quote.html
But that's the institutional share class, with a $10 million minimum. I presume that the higher fees for non-institutional investors will eat up some of that alpha, and then it's very inefficient taxwise (1.78% tax cost) so unless you're in a tax-sheltered account, there goes the rest of that alpha and then some.
Vanguard's balanced index fund, VBINX, turned in a 5.99% annual return over that ten year period, with a 0.80% tax cost. In a taxable account, the annual return would be nearly double GBMFX, and even in a tax-sheltered account it would be superior. And of course there's no manager-transition risk, concern about luck, etc.
I don't know about the since-inception results, maybe in the 90s GBMFX had glory years, but MJG's thoughts -- that even funds run by brilliant, ethical, humble, and careful managers like Mr. Grantham have a hard time outperforming -- are becoming ever more convincing.
But since over a ten year period -- one that included the biggest housing bubble and market crash since the Great Depression -- a simple balanced index fund outperformed GMO's flagship fund on an aftertax basis by an average of 226 basis points a year, and on a pretax basis by 117 basis points a year (and that's assuming you were lucky enough to have $10 million to buy the cheaper, institutional share class), I think that for most people with at least 10 years till retirement, the index fund is the better bet.
But that investor will have to close his eyes and even add more if possible during the inevitable downturns. Not every investor can do that.
I think we are on the edge of another bubble at some point in the next 1-3 years. I wonder if, for younger investors who have lived through the GR, and now another bubble, if they won't avoid the market altogether. Problem is, many are also very wary of owning property as well (and/or have student debt). Where will they invest for the future? Will they simply spend (buy dividend-stalwart consumer stocks !!!)?
We are "blessed" to be living in interesting times.
My parents caught some of the dot-com boom; but otherwise enjoyed what in retrospect seems to have been a very good market throughout their investing careers. I don't think many retail investors investing today will enjoy such good fortunes.
I was ready to whip out my checkbook to purchase a few shares of a GMO fund when I realized I was a tad short of the membership entry fee this month. Too, too bad. I suppose a few other MFOers also find themselves in that situation.
I was not unfamiliar with the stiff entry requirements that GMO directly imposes on potential customers.
Given the likely paucity of a large general clientele base, I’m somewhat puzzled by Jeremy Grantham’s frequent talks at public events like the MoneyShow. I benefited from his global insights on several occasions at such presentations. He is an impressive, controlled speaker who inspires trust and confidence. Wallets open when Grantham speaks.
After one such presentation, I did seek an alternate source of his funds working through Wells Fargo. That was a few years ago. The initial purchasing requirements were very modest, but I rejected the opportunity because of a heavy initial purchase fee. I still blanch from such ruthless and completely unnecessary charges. I consistently stay away because of that fee structure.
Best Wishes.