Classic Bernstein 7 — The Great Fund Fee Mystery

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Classic Bernstein 7 — The Great Fund Fee Mystery

Post by SimpleGift »

PREFACE: This is the seventh in a series of posts highlighting the classic investing insights of William Bernstein from the 1990s and early 2000s. Many new Bogleheads have never been exposed to his early writings — and while the data sets used may seem antiquated, his portfolio concepts and novel analyses are still helpful to investors today, new and old alike.

Previous topics in the series: 1-Asset Allocation & Time Horizon, 2-Choosing Portfolio Bond Duration, 3-Diversifying Portfolio Equities, 4-Stocks Always Beat Bonds?, 5-What’s a Thing Worth?, 6-The Value Premium & Inflation
------------------------------------------------------------------------------------------

If all Bogleheads could agree on just one point, it’s likely to be the dictum that “costs matter” when it comes to mutual fund investing. In fact, in this post, we may be surprised to discover that fees matter much more than we ever thought. To begin, most of us have the following mental model of investor returns and expense ratios:
  • Image
The Analysis. To test this formula and hypothesis, Mr. Bernstein regressed the returns of all domestic mutual funds against their expense ratios for the 5-year period from 4/94-3/99 (see chart at left below). It’s clear that this fund cloud has a tendency to slope down and to the right — the higher the fund expense, the lower the return. But what’s surprising is the regression slope of the line is -2.22. In other words, every added dollar of expense deducts more than 2 dollars of return. Statistically, this is not a fluke, as the 95% confidence levels were between -2.54 and -1.91.
  • Image
    Note: The circled funds at upper left are S&P 500 index funds.
    Source: William Bernstein, Efficient Frontier (7/99)
To investigate further, Mr. Bernstein performed the same analysis for each of the 9 Morningstar style boxes over the same period (table at right above). Note that the regression slope was greater than 1.00 in 8 of 9 cases, and greater than 2.00 in 4 of 9 cases. How can this be? What can account for the 2-1 effect of fund expenses on returns?

The Mystery. In a follow-up analysis, Mr. Bernstein examined different variables that might explain the fund return deficit above-and-beyond just the expense ratio:
  • a) He looked first at fund turnover — but found that while turnover hurt value stock funds somewhat, it seemed to help growth funds (managers were perhaps purchasing momentum-related excess return with their turnover expense).

    b) He also looked at moral turpitude, thinking fund companies that saw nothing wrong with high fund loads might also tolerate other activities harmful to shareholders (e.g., fund front-running, lax trading execution, etc.). But, inexplicably, he found load funds had less of an excess expense penalty than no-load funds.
In short, to paraphrase John Bogle, costs matter much more than you’d expect. It’s just not clear why. Thoughts?
Last edited by SimpleGift on Sat May 28, 2016 8:08 pm, edited 1 time in total.
User avatar
Topic Author
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by SimpleGift »

Just to add more details about fund turnover (which seems a logical place to look to solve this mystery). John Bogle, in his own analysis of large blend funds — where he found the return/expense regression slope was -1.80 — also proposed fund turnover as a likely explanation for the nearly 2-1 effect of expenses on returns:
John Bogle, in [i]Common Sense on Mutual Funds[/i] wrote:Our intuition might tell us that each point of cost should cost exactly one point of return, but something much more onerous is taking place. Although the causative factors are not exactly clear, one explanation seems to hold some extra merit: High-cost funds tend to have high turnover, and portfolio transactions carry a substantial cost of their own.
However, Mr. Bernstein tested the fund turnover theory in depth and found it lacking:
William Bernstein, in [i]Efficient Frontier[/i] (7/99) wrote:If one adds in turnover as a second variable in the regression, the return/turnover slope is negative, but with only 0.47% of return lost for each 100% of turnover. This is not nearly as impressive as one would expect. Further, superimposing turnover adds only a minimal amount of statistical power to the analysis, with an adjusted R-squared of 0.131, versus 0.129 for expense alone. Finally, doing two-factor expense/turnover regressions for the 9 Morningstar boxes shows no particular effect of turnover with small caps, where one would expect to see it most clearly.
Thus the mystery lives on…
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by lack_ey »

Maybe

(1) fund managers feel like they need to take greater risks to make up for the higher expense ratio and chase worse investments (for example, see high/low beta anomaly, though specifically high beta would probably be a good thing in a bull market like that studied) with a higher potential payoff but lower expected return

(2) high expenses are correlated with recent past outperformance (or maybe a fund hitting a cold streak can't charge high fees or everyone would run away), and there's reversion to the mean in future returns

(3) the types of fund companies running expensive funds are perhaps smaller and less established? (it doesn't make much sense but most of the relatively cheaper funds would be from the likes of Vanguard, maybe Fidelity, T. Rowe Price, even American Funds because this stuff doesn't take into account loads, etc., which did relatively well)
User avatar
JoMoney
Posts: 9765
Joined: Tue Jul 23, 2013 5:31 am

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by JoMoney »

It's interesting that he didn't find any added explanation from turnover.
This study: Cross Trading by Investment Advisers: Implications for Mutual Fund Performance
seems to have found significant penalty when (rather than generic turnover) it's measured on the level of "cross trading" activity of funds, that is where the turnover transactions may be principled or arranged outside of the normal market channels.
Not only was high "cross trading" activity associated with a significant penalty, it also found higher amounts of cross trading where there are "soft dollar" arrangements between brokers and funds.
... total speculation, but I could imagine the situation being that brokers that have a relationship with a fund company have less of a reason to try and make additional money off of funds they offer from which they get sales load commissions. If the broker is offering the funds at no-load they may be trying to make up that revenue through the "soft dollar" transactions when the fund has turnover/trading activity that uses that broker...
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham
User avatar
neurosphere
Posts: 3778
Joined: Sun Jan 17, 2010 1:55 pm

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by neurosphere »

Simplegift wrote: To investigate further, Mr. Bernstein performed the same analysis for each of the 9 Morningstar style boxes over the same period (table at right above). Note that the regression slope was greater than 1.00 in 8 of 9 cases, and greater than 2.00 in 4 of 9 cases.
Fyi, vanguard has graphs of the relationship between cost and 10-year return for each of the style boxes, to compliment Bernsteins table. The charts below are taken from a document called "The case for index-fund investing" (March 2015). I use this image in my financial presentations. I explain that we can't choose, in advance, a fund which will be "above the line". But we CAN choose, in advance, a fund from either the LEFT side of each plot (low-fee) or the RIGHT side (high-fee).

Here is the one for stock funds:

Image

And here is the one for bond funds:

Image

Here is the caption:
Notes: Each plotted point represents a fund within the specific size, style, and asset group. The funds are plotted to represent the relationship of their expense ratio (x-axis) versus the ten-year annualized excess return relative to their style benchmark (y-axis). The straight line represents the linear regression, or the best-fit trend line, showing the general relationship of expenses to returns within each asset group. The scales are standardized to show the slopes’ relationships to each other, with expenses ranging from 0% to 3% and returns ranging from –15% to 15% for U.S. equities and –5% to 5% for U.S. fixed income. Some funds’ expense ratios and returns go beyond the scales and are not shown. See the Appendix, on page 18, for benchmarks used for each Morningstar style box. Data reflect periods ended December 31, 2014. Sources: Vanguard calculations, using data from Morningstar, Inc., MSCI, CRSP, and Standard & Poor’s.
Note that the previous year's PDF of the entire 20 page document with data through 2013 is available here: https://pressroom.vanguard.com/nonindex ... 9.2014.pdf (I'm not sure if the most recent version is available publicaly yet, I found it at the advisor's site at Vanguard).

Neurosphere
Last edited by neurosphere on Sat May 14, 2016 10:32 am, edited 1 time in total.
If you have to ask "Is a Target Date fund right for me?", the answer is "Yes".
User avatar
Topic Author
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by SimpleGift »

lack_ey wrote:(1) fund managers feel like they need to take greater risks to make up for the higher expense ratio and chase worse investments…
If I had to hazard a guess, the answer to the mystery might be somewhere in his realm: fund managers taking greater risks and chasing worse investment ideas to compensate for their higher expense ratios — though I don’t have a clue how one might quantitatively test this theory.

After all, we know bond managers routinely engage in this practice all the time. The greater the bond fund expense ratio, the riskier the credits in which the fund manager is forced to invest, just to be competitive with the yields of other bond funds in their category. See Neurosphere’s excellent charts above.

If bond fund managers do it, why not equity fund managers as well?
dbr
Posts: 33842
Joined: Sun Mar 04, 2007 9:50 am

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by dbr »

Simplegift wrote:
lack_ey wrote:(1) fund managers feel like they need to take greater risks to make up for the higher expense ratio and chase worse investments…
If I had to hazard a guess, the answer might be somewhere in his realm: fund managers taking greater risks and chasing worse investments to compensate for their higher expense ratios — though I don’t have a clue how one might quantitatively test this idea.

After all, we know bond managers routinely engage in this practice all the time. The greater the bond fund expense ratio, the riskier the credits in which the fund manager is forced to invest, just to be competitive with the yields of other bond funds in their category. See Neurosphere’s excellent charts above.

If bond fund managers do it, why not equity fund managers as well?
I suppose if we can't explain a result by doing simple arithmetic, then the currently conventional explanation is that something behavioral is involved. So the above sounds plausible.
User avatar
neurosphere
Posts: 3778
Joined: Sun Jan 17, 2010 1:55 pm

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by neurosphere »

I was thinking about how in the short run, the distribution of a fund's return is almost random with respect to fees. E.g. in a one year period, many active funds might be very likely to beat their benchmark. But with each passing year, history has shown it gets harder and harder to overcome the drag on fees. So as time goes by, the percentage of funds which beat the benchmark (i.e. those dots which are "above the line" in the previous charts), will decrease each year.

What I would love to see (and if I had the skills/time I would do this myself), is a sequence of graphs for a given style box which shows the scatter plots from my previous post for 1-year returns, 2-year returns, 3-year returns...30+ year returns. I think it would be fun and illustrate two things clearly: 1) each year the "dots" on the graph (an individual mutual fund) will trend downward, particularly on the right of the graph and 2) each year the total number of dots will decrease (reflecting closed or merged funds).

Vanguard has a bar chart which shows this trend for all active fund over time (but not broken down by fee). It's figure 12 at this link:
https://pressroom.vanguard.com/nonindex ... 9.2014.pdf

As one goes from the bottom (1-year returns) up to the top (15-year returns) one can clearly see that their is "more blue" above the line (i.e. a greater percentage of funds underperform their benchmark) as time goes by. In addition, the "light blue" increases (funds which are merged or closed).

But it would be very nice to show this information in two dimensions (performance vs fee) for each mutual fund, and then have a movie or step-wise images for the third-dimension (time).

Something to put on my to-do list I guess. Actually, now that I think about it, I'm not sure it would be THAT hard to do with conventionally available software. The problem would be getting and arranging the dataset in the first place (historical fee and return information for each mutual fund for each period, including merged/closed funds). Ok, so now that I re-think about it, it sounds like a giant task for a "wouldn't it be nice if..." :D

Neurosphere
If you have to ask "Is a Target Date fund right for me?", the answer is "Yes".
User avatar
Topic Author
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by SimpleGift »

neurosphere wrote:I was thinking about how in the short run, the distribution of a fund's return is almost random with respect to fees. E.g. in a one year period, many active funds might be very likely to beat their benchmark. But with each passing year, history has shown it gets harder and harder to overcome the drag on fees.
In his analysis of fund expenses and returns covered in the OP, Mr. Bernstein speaks to your point:
William Bernstein, in Efficient Frontier (7/99) wrote:The other interesting piece of data to fall out of the analysis is the increasing importance of expense over time. When 1-year returns are examined, the R-squared for fund expenses is only 0.005. In other words, only 0.5% of 1-year returns can be explained by costs. However, at 5 years this rises to 12.9%, and at 15 years fully 36% of fund return is explained by expenses.

What is happening here is that over time the variation of manager performance, which is random, "washes out," leaving expense as the single most important factor determining return.
There were no charts accompanying his comments — and I don’t recall seeing any visuals elsewhere directly showing this effect over time — but your idea of creating them sounds intriguing.
dbr
Posts: 33842
Joined: Sun Mar 04, 2007 9:50 am

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by dbr »

The set of data for results over time would be a really good idea.
User avatar
siamond
Posts: 5560
Joined: Mon May 28, 2012 5:50 am

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by siamond »

Simplegift wrote:
lack_ey wrote:(1) fund managers feel like they need to take greater risks to make up for the higher expense ratio and chase worse investments…
If I had to hazard a guess, the answer to the mystery might be somewhere in his realm: fund managers taking greater risks and chasing worse investment ideas to compensate for their higher expense ratios — though I don’t have a clue how one might quantitatively test this theory.
Yup. Exactly what I was thinking when reading the first post. Except that I might not worded it in such a polite manner, and maybe spoken of greed and incompetence often going together...

Thank you for sharing this mystery, this is something I had missed in the Bernstein's mine of wonders. Intriguing indeed, and also very reassuring that we Bogleheads seem to be doing the right thing!
User avatar
LadyGeek
Site Admin
Posts: 66320
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Classic Bernstein 7 — The Great Fund Fee Mystery

Post by LadyGeek »

A complete list of Simplegift's series is now in the wiki: Classic Bernstein
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
Post Reply