the sellers of said funds, of course, might argue/suggest that while the average expensive fund underperforms, perhaps THEIR fund is special.
The buyer/prospective buyer of said fund, having seen your charts, is making the judgement that THEY have the ability to pick the needle in the haystack (or at the very least, hard-to-find-thing in the haystack).
Give me my passive funds and low cost funds....that's a game I can win.
Thank you for posting this super-interesting analysis. Any insights on the reason for the fact that inexpensive funds perform much better? I would not have been surprised to see high fees/low fees funds performung similarly. But the a clear inverse correlation between fees and performance is astonishing. Any insights in why?
Thanks. It's the 'you get what you don't pay for' principle that folks like Bogle et al have written/talked about at length. This is a pretty simple acid test where you're comparing a fund to the average fund in its peer group and that average fund's returns are net of the fees it levies. So if the fund you're comparing to the cat avg has a lower fee, all things being equal, it should outperform the average fund (provided we're controlling for factors like size, style, etc.). That doesn't *always* happen - pricier funds might for instance have disproportionate exposure to a particular outperforming area and perhaps that vaults them ahead of even lower-cost funds in the category. But usually that washes out and funds in a category perform similarly before fee, leaving expense differences to break the deadlock. There are some other factors that are maybe making fees even more predictable than before -- for instance, funds' returns being more alike, fewer new incepts and so more-stable peer groups, etc. But the biggie is that when funds are more-or-less alike except for fees than that's going to express itself in differences in net returns. Hope this is helpful! Regards, Jeff Ptak
....and one more comment: I wonder whether it might be more accurate to simply label the y-axis "average return vs. category average"? "excess return" implies that the top three quartiles are beating the benchmark index, or somehow justifying thier existence -or at least it might be seen that way, particularly by sellers of these funds, or owners of these funds wishing to rationalize their decision to invest.
We know from SPIVA data that the vast majority of funds underperform appropriate benchmarks.
Thanks. I compared each fund's return in a month to the average return of all funds in its peer group that month ("category average"). The difference is the excess return versus the category average. I then averaged all of those excess returns for that month for the funds in each bucket. So its an average excess return vs. the category average. By that measure, the top three groupings outperformed and the bottom two did not. Keep in mind that the 'benchmark' in this context is not a costless index (which is the acid test in popular studies of active fund performance like our Active Passive Barometer report or the S&P report you mention) but rather an average of funds' *net-of-fee* returns. That makes it easier to reconcile the outperformance of the cheapest three groupings to the more sobering results you'd see in our AP barometer or SPIVA, as this is comparing funds to an easier-to-clear hurdle than a costless index, if that makes sense. Many thanks for this feedback and hope that's clarifying! Jeff
Yes, makes total sense! I'm just always wary of presentations' vulnerability to abuse by marketers: "look, our fund is the the top quintile vis-a-vis excess return" when in fact the fund has lagged its benchmark.
None of this has to do with the point you were trying to make, which you did well!
Are you comparing funds or strategies? If funds-only, I would be concerned that you’re just visualizing the difference between MF share classes, not a comparison of various strategies and managers
All shareclasses of all funds. I've heard this argument before, typically from load shops that have a lot of shareclasses. It's a red herring. There aren't enough of these situations to move the needle and moreover there isn't the spread across the fee groups that you might assume. Nonetheless, I reran the test, this time limiting it to 'oldest shareclass' of each unique fund, which ostensibly would mitigate any potential skew that multiple share class funds could be causing. Results were quite similar with 0.6% - 1.4% annual margin of outperformance of cheapest over priciest across the trailing periods. You see the same stairstep pattern too. This isn't just a shareclass thing.
Thanks, I appreciate the response. I thought it was pretty widespread since when I run searches in Direct, unless I filter for Institutional class I seem to get at least 5 instances of every strategy, but that’s probably just the huge managers.. most boutiques do seem to have just 2-3 classes
Definitely worth posting, good visualization.
the sellers of said funds, of course, might argue/suggest that while the average expensive fund underperforms, perhaps THEIR fund is special.
The buyer/prospective buyer of said fund, having seen your charts, is making the judgement that THEY have the ability to pick the needle in the haystack (or at the very least, hard-to-find-thing in the haystack).
Give me my passive funds and low cost funds....that's a game I can win.
Thank you for posting this super-interesting analysis. Any insights on the reason for the fact that inexpensive funds perform much better? I would not have been surprised to see high fees/low fees funds performung similarly. But the a clear inverse correlation between fees and performance is astonishing. Any insights in why?
Thanks. It's the 'you get what you don't pay for' principle that folks like Bogle et al have written/talked about at length. This is a pretty simple acid test where you're comparing a fund to the average fund in its peer group and that average fund's returns are net of the fees it levies. So if the fund you're comparing to the cat avg has a lower fee, all things being equal, it should outperform the average fund (provided we're controlling for factors like size, style, etc.). That doesn't *always* happen - pricier funds might for instance have disproportionate exposure to a particular outperforming area and perhaps that vaults them ahead of even lower-cost funds in the category. But usually that washes out and funds in a category perform similarly before fee, leaving expense differences to break the deadlock. There are some other factors that are maybe making fees even more predictable than before -- for instance, funds' returns being more alike, fewer new incepts and so more-stable peer groups, etc. But the biggie is that when funds are more-or-less alike except for fees than that's going to express itself in differences in net returns. Hope this is helpful! Regards, Jeff Ptak
Very clear, thank you. Enric
....and one more comment: I wonder whether it might be more accurate to simply label the y-axis "average return vs. category average"? "excess return" implies that the top three quartiles are beating the benchmark index, or somehow justifying thier existence -or at least it might be seen that way, particularly by sellers of these funds, or owners of these funds wishing to rationalize their decision to invest.
We know from SPIVA data that the vast majority of funds underperform appropriate benchmarks.
Thanks again!
Thanks. I compared each fund's return in a month to the average return of all funds in its peer group that month ("category average"). The difference is the excess return versus the category average. I then averaged all of those excess returns for that month for the funds in each bucket. So its an average excess return vs. the category average. By that measure, the top three groupings outperformed and the bottom two did not. Keep in mind that the 'benchmark' in this context is not a costless index (which is the acid test in popular studies of active fund performance like our Active Passive Barometer report or the S&P report you mention) but rather an average of funds' *net-of-fee* returns. That makes it easier to reconcile the outperformance of the cheapest three groupings to the more sobering results you'd see in our AP barometer or SPIVA, as this is comparing funds to an easier-to-clear hurdle than a costless index, if that makes sense. Many thanks for this feedback and hope that's clarifying! Jeff
Yes, makes total sense! I'm just always wary of presentations' vulnerability to abuse by marketers: "look, our fund is the the top quintile vis-a-vis excess return" when in fact the fund has lagged its benchmark.
None of this has to do with the point you were trying to make, which you did well!
Excellent visualization of the impact of fees Jeffrey.
Thanks!
Is this research only related to stock funds or does it apply to bond funds as well?
Is this research only related to stock funds or does it apply to bond funds as well?
I can't imagine a clearer image to drive home the Boglehead philosophy. Thanks for the analysis!
Are you comparing funds or strategies? If funds-only, I would be concerned that you’re just visualizing the difference between MF share classes, not a comparison of various strategies and managers
All shareclasses of all funds. I've heard this argument before, typically from load shops that have a lot of shareclasses. It's a red herring. There aren't enough of these situations to move the needle and moreover there isn't the spread across the fee groups that you might assume. Nonetheless, I reran the test, this time limiting it to 'oldest shareclass' of each unique fund, which ostensibly would mitigate any potential skew that multiple share class funds could be causing. Results were quite similar with 0.6% - 1.4% annual margin of outperformance of cheapest over priciest across the trailing periods. You see the same stairstep pattern too. This isn't just a shareclass thing.
Thanks, I appreciate the response. I thought it was pretty widespread since when I run searches in Direct, unless I filter for Institutional class I seem to get at least 5 instances of every strategy, but that’s probably just the huge managers.. most boutiques do seem to have just 2-3 classes