Larry Swedroe: A Tale Of 3 Small Value Funds

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 10:36 am

https://www.etf.com/sections/index-inve ... nopaging=1

Larry distinguishes pure index funds from passive asset class funds, and shows how some passive asset class funds can avoid some weaknesses of indexing. Different passive asset class funds will have different exposures to the factors that drive returns. Cost per unit factor exposure is more important than absolute cost alone; and to target the same portfolio factor loads, one will use less of the more expensive fund with deeper exposures. In deciding on potential asset class funds for one’s portfolio, the investor needs to look beyond the fund name.

Dave

User avatar
hdas
Posts: 1355
Joined: Thu Jun 11, 2015 8:24 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by hdas » Thu Jan 31, 2019 2:45 pm

Random Walker wrote:
Thu Jan 31, 2019 10:36 am
https://www.etf.com/sections/index-inve ... nopaging=1

Larry distinguishes pure index funds from passive asset class funds, and shows how some passive asset class funds can avoid some weaknesses of indexing. Different passive asset class funds will have different exposures to the factors that drive returns. Cost per unit factor exposure is more important than absolute cost alone; and to target the same portfolio factor loads, one will use less of the more expensive fund with deeper exposures. In deciding on potential asset class funds for one’s portfolio, the investor needs to look beyond the fund name.

Dave
What is this??......Why did Mr. Swedroe pick the "easy to beat" example in his tale? Why did he forget to add IJS to the mix?

Since the start of the Omni fund it outperformed the other two for 1/2 of the cost

Since the start of IJS, Ishares matches DFA fund for 1/2 the cost. (lower std deviation by the way)

I think little details like this explain forum members reactions against the incumbent.

Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 3:41 pm

Hdas,
Are you serious? Larry was simply making a few specific points with his examples. Structured asset class funds can improve on straight indexing. The returns of funds, for good or bad, depend on their exposures to the factors, and different funds with similar names will have different exposures. His examples make the points. Those points, supported by examples, are what counts. Which specific examples he chooses not really significant.

Dave

User avatar
hdas
Posts: 1355
Joined: Thu Jun 11, 2015 8:24 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by hdas » Thu Jan 31, 2019 4:03 pm

Random Walker wrote:
Thu Jan 31, 2019 3:41 pm
Hdas,
Are you serious? Larry was simply making a few specific points with his examples. Structured asset class funds can improve on straight indexing. The returns of funds, for good or bad, depend on their exposures to the factors, and different funds with similar names will have different exposures. His examples make the points. Those points, supported by examples, are what counts. Which specific examples he chooses not really significant.

Dave
Yes I'm serious. , specifically about this:
Cherry picking, when used figuratively, refers to selective extraction of points in an argument in order to refute or affirm them while ignoring others which will not support the point(s) being made. It derives from the obvious reluctance to harvest unripe, or overripe, fruit and to select only those which will make profit (or pie).
Mr. Swedroe adds:
While VISVX has an expense ratio of 0.19% (their Admiral shares version costs just 0.07%) and DFSVX has an expense ratio of 0.52%, the higher costs of the Dimensional fund have been more than offset by their greater exposure to the size and value factors and their focus on adding value by minimizing the negatives of pure indexing. For the full period that both have existed—June 1998 through December 2018—DFSVX returned 8.6%, outperforming the lower cost VISVX, which returned 7.9%.
This quote from him ends up being an endorsement of DFA which his firm uses (this disclosure was never included in the article). He makes a comparison where said DFA fund looks good and justifies the high expense ratio based on performance. The only reason that he was able to make this point is because the selection of funds for his example. Had he included the Ishares fund, he couldn't have made his remarks about pure (cheaper) indexing.

Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

cheezit
Posts: 268
Joined: Sat Jul 14, 2018 7:28 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by cheezit » Thu Jan 31, 2019 4:09 pm

There are several issues that are separate:

1) Can an investor benefit from a fund with increased factor exposure?
2) Does a fund need to use an algorithmic strategy (rather than following an index) to achieve increased factor exposure?
3) Do the other potential benefits of not following an index allow a fund to increase its returns by avoiding some pitfalls of indices?

With regard to issue #1, I'm not going to try to answer that here because I think there are a zillion other threads where the question of the effectiveness (and cost-effectiveness) of factor investing have been covered and I don't want to bog down this thread rehashing all that, among other things.

With regard to issue #2, here are the factor regressions for all the funds that Larry mentioned in his article, along with two other funds that have been around for at least as long as the newest fund Larry mentioned (BOSVX):

Code: Select all

Ticker	Start Date	End Date	Rm-Rf	SMB	HML	RMW	CMA	Alpha	R^2	ER	Returns (TWRR)
VBR	Sep 2011	Dec 2018	1.02	0.55	0.18	0.06	0.12	-0.06%	97.6%	0.07%	11.05
IJS	Sep 2011	Dec 2018	1.01	0.88	0.25	0.22	0.11	0.04%	97.7%	0.25%	11.53
DFSVX	Sep 2011	Dec 2018	1.05	0.86	0.43	0.09	0.01	-0.12%	98.2%	0.52%	9.61
BOSVX	Sep 2011	Dec 2018	1.02	0.97	0.60	0.18	-0.15	-0.06%	97.3%	0.60%	9.67
RZV	Sep 2011	Dec 2018	1.15	1.15	0.54	0.24	0.09	-0.25%	90.8%	0.35%	8.27
(nb. I have included returns over the period because Larry mentioned annual returns for selected years in his article; obviously, returns over a roughly seven year period are not super meaningful).


To me, it looks like just using a different index can get you a pretty high exposure. For reference, VBR (the ETF version of VISVX) tracks the CRSP small value index, while IJS tracks the S&P 600 Value index and RZV tracks the S&P 600 Pure Value index.



With regard to issue #3, I'm not sure how to measure that. My first gut instinct is to look at alpha, but the R^2 coefficients for the Fama-French 5 factor model are low enough for these funds over this period that actual fit error probably overwhelms everything else in terms of the alpha coefficient that the multivariable linear regression spits out. Comparing returns is just going to tell you how well the small and value factors did over that period. I'm all ears as to measuring the benefits of non-index passive funds.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 4:49 pm

Hdas,
I understand your point that using the DFA example creates the opportunity to show how DFA earns it's higher expense ratio. Larry does disclose his firms’s use of DFA and Bridgeway, as he does in every article I read where he discusses them. In his essay, two of the points Larry makes is cost per unit factor exposure versus cost alone and some of the benefits of not adhering to a pure indexing strategy (patient trading, screening out certain stocks, buy/hold ranges, momentum and profitability screens). So his examples are appropriate for those points as well. I don’t know anything about IJS, but I’m going to go find out now.

Dave

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 4:58 pm

Cheezit,
Regarding #3, I’ve asked that question of Larry in the past. And his answer was what you did: look at the alpha.

Dave

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Thu Jan 31, 2019 5:21 pm

IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 5:44 pm

DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
Exactly.

From portfolio visualizer, longest returns available: 1/01-12/18
DFSVX 9.51%
IJS 8.55%
https://www.portfoliovisualizer.com/bac ... ion2_2=100

And here’s the factor analysis. DFSVX has somewhat more exposure to size and value. It’s alpha is +0.20% v. IJS -0.15%. Perhaps that alpha represents patient trading, buy/hold ranges, screening out IPOs, extreme small growth, penny stocks, bankruptcy?
https://www.portfoliovisualizer.com/fac ... sion=false

Dave
Last edited by Random Walker on Thu Jan 31, 2019 5:56 pm, edited 1 time in total.

fennewaldaj
Posts: 823
Joined: Sun Oct 22, 2017 11:30 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by fennewaldaj » Thu Jan 31, 2019 5:50 pm

DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 6:04 pm

fennewaldaj wrote:
Thu Jan 31, 2019 5:50 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.
That is true, but there is an implicit cost involved: slightly less efficient portfolio. People who factor tilt are basically trying to diversify away from market beta. Market beta dominates the risk of most of our portfolios. If one uses a long only fund with deeper exposure to size and value, he can obtain the same desired exposure to those factors by taking on less additional exposure to market risk. Using deeper exposures in small value fund moves the portfolio slightly more in the direction of risk parity: same tilt to size and value with less exposure to market factor.

Dave

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Thu Jan 31, 2019 6:05 pm

fennewaldaj wrote:
Thu Jan 31, 2019 5:50 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.
Yes, it is relevant and I have a lot of respect for RobertT's comments. In my own case, I prefer to get as much SCV exposure as I can in a smaller allocation. Just another approach, but getting more SCV exposure at a somewhat higher expense leaves more room in the portfolio for other diversifying factors, including a higher allocation to bonds.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 6:17 pm

DaufuskieNate wrote:
Thu Jan 31, 2019 6:05 pm
fennewaldaj wrote:
Thu Jan 31, 2019 5:50 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.
Yes, it is relevant and I have a lot of respect for RobertT's comments. In my own case, I prefer to get as much SCV exposure as I can in a smaller allocation. Just another approach, but getting more SCV exposure at a somewhat higher expense leaves more room in the portfolio for other diversifying factors, including a higher allocation to bonds.
Which means somewhat more efficient portfolio: portfolio with same expected return, higher bond allocation, less market beta exposure, narrower SD.

Dave

fennewaldaj
Posts: 823
Joined: Sun Oct 22, 2017 11:30 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by fennewaldaj » Thu Jan 31, 2019 6:22 pm

Random Walker wrote:
Thu Jan 31, 2019 6:17 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 6:05 pm
fennewaldaj wrote:
Thu Jan 31, 2019 5:50 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.
Yes, it is relevant and I have a lot of respect for RobertT's comments. In my own case, I prefer to get as much SCV exposure as I can in a smaller allocation. Just another approach, but getting more SCV exposure at a somewhat higher expense leaves more room in the portfolio for other diversifying factors, including a higher allocation to bonds.
Which means somewhat more efficient portfolio: portfolio with same expected return, higher bond allocation, less market beta exposure, narrower SD.

Dave
right if you want more small value than can be provided by a 100% of US stocks allocation to ijs the deeper value funds are your only option.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Thu Jan 31, 2019 6:38 pm

fennewaldaj wrote:
Thu Jan 31, 2019 6:22 pm
Random Walker wrote:
Thu Jan 31, 2019 6:17 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 6:05 pm
fennewaldaj wrote:
Thu Jan 31, 2019 5:50 pm
DaufuskieNate wrote:
Thu Jan 31, 2019 5:21 pm
IJS has significantly lower value exposure than BOSVX and DFSVX. Kind of makes the point that it's not only about expenses, but also how much exposure you are getting to the factors that you want. IJS outperformed over the timeframe mentioned above BECAUSE it had lower value exposure during a time when value was not doing well. That's exactly what you would expect. So fine, add IJS to the analysis but it doesn't change ANY of the points being made.
thats true. The point RobertT has made several times is still relevant though. That is unless you are going 100% small value you can achieve what you need by simply holding more of the less factor exposed fund. So like 50% vbr 50% vti instead of 25% BOSVX 75% VTI for example.
Yes, it is relevant and I have a lot of respect for RobertT's comments. In my own case, I prefer to get as much SCV exposure as I can in a smaller allocation. Just another approach, but getting more SCV exposure at a somewhat higher expense leaves more room in the portfolio for other diversifying factors, including a higher allocation to bonds.
Which means somewhat more efficient portfolio: portfolio with same expected return, higher bond allocation, less market beta exposure, narrower SD.

Dave
right if you want more small value than can be provided by a 100% of US stocks allocation to ijs the deeper value funds are your only option.
But the point is that even if you want less SV than can obtain target SV exposure with less market exposure when use fund with deeper exposure to size and value.

Dave

cheezit
Posts: 268
Joined: Sat Jul 14, 2018 7:28 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by cheezit » Thu Jan 31, 2019 6:50 pm

If you want deeper exposure why not look at RZV, which tracks the S&P 600 Pure Value index? As you see in the chart from my earlier post, it had a higher exposure to size, value and quality than DFSVX over the period in the article. If you look at the entire lifespan of RZV (back to April of 2006), the differences become even larger, with RZV having a much larger size loading (1.29 vs 0.87), value loading (0.94 vs 0.44), quality loading (0.34 vs 0.09), and a lower correlation to the total US market (0.80 vs 0.92) compared to DFSVX. And the ER is lower to boot (0.35 vs 0.52). If I wanted to tilt hard, I'd give it a long look.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Thu Jan 31, 2019 7:31 pm

cheezit wrote:
Thu Jan 31, 2019 6:50 pm
If you want deeper exposure why not look at RZV, which tracks the S&P 600 Pure Value index? As you see in the chart from my earlier post, it had a higher exposure to size, value and quality than DFSVX over the period in the article. If you look at the entire lifespan of RZV (back to April of 2006), the differences become even larger, with RZV having a much larger size loading (1.29 vs 0.87), value loading (0.94 vs 0.44), quality loading (0.34 vs 0.09), and a lower correlation to the total US market (0.80 vs 0.92) compared to DFSVX. And the ER is lower to boot (0.35 vs 0.52). If I wanted to tilt hard, I'd give it a long look.
1) Negative momentum exposure is the biggest issue...but also..
2) Smaller AUM...and
3) Fewer number of stocks held.

Dead Man Walking
Posts: 904
Joined: Wed Nov 07, 2007 6:51 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Dead Man Walking » Fri Feb 01, 2019 2:13 am

Any discussion that includes funds that use screens to select stocks should include active funds. Active managers use screens to select stocks too. Perhaps funds from Royce or other active funds should be included in the discussion. Small cap value isn’t limited to the funds discussed here.

DMW

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Fri Feb 01, 2019 8:22 am

Dead Man Walking wrote:
Fri Feb 01, 2019 2:13 am
Any discussion that includes funds that use screens to select stocks should include active funds. Active managers use screens to select stocks too. Perhaps funds from Royce or other active funds should be included in the discussion. Small cap value isn’t limited to the funds discussed here.

DMW
A good definition of passive is no market timing and no individual security selection. With this definition, many formulaic cap weighted strategies remain agnostic to the individual assets held, there is no subjectivity involved.

Dave

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Fri Feb 01, 2019 10:44 am

Random Walker wrote:
Fri Feb 01, 2019 8:22 am
Dead Man Walking wrote:
Fri Feb 01, 2019 2:13 am
Any discussion that includes funds that use screens to select stocks should include active funds. Active managers use screens to select stocks too. Perhaps funds from Royce or other active funds should be included in the discussion. Small cap value isn’t limited to the funds discussed here.

DMW
A good definition of passive is no market timing and no individual security selection. With this definition, many formulaic cap weighted strategies remain agnostic to the individual assets held, there is no subjectivity involved.

Dave
And...with no subjectivity, there is less risk of style drift which is something to watch out for in the history of any active fund.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Fri Feb 01, 2019 10:46 am

An active fund can use formulaic screens, but after a stock passes through the screens, it’s up to an individual to subjectively decide whether to buy.

Dave

User avatar
packer16
Posts: 1306
Joined: Sat Jan 04, 2014 2:28 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by packer16 » Fri Feb 01, 2019 10:59 am

I think the implicit assumption that the human will give a poorer result is only true in some cases. Even quantitative methods use human adjusting (such as removal of certain stocks due to various factors (IPO, etc.). The machines are being used together with human judgement to arrive at a buy/sell decision. The extreme of this is just invest in an index but you seem to be arguing for an index plus some adjustments for security selection. If there is no security selection than SCV is no longer passive because you are selecting the SCV stocks (which change every period) out of the stock universe & you are timing the buy/sell to whenever the stock goes in or out of your SCV group of stocks. You may not begin your process with stock selection but in the end you are doing stock selection & market timing. Maybe a better way to describe factor investing is quantitative active investing versus passive investing so we will not confuse indexing & factor investing.

Packer
Buy cheap and something good might happen

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Fri Feb 01, 2019 11:08 am

If you compare the Royce SCV fund to DFA, the Royce fund has a lot more drift in its exposure to value, including a negative exposure at times. If you are looking for a "smart" manager that you think can produce consistent alpha, then this is not a problem. If, instead, you are looking for a consistent exposure to a factor that fits into the overall design of your portfolio, then style drift is a problem.

afan
Posts: 4477
Joined: Sun Jul 25, 2010 4:01 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by afan » Fri Feb 01, 2019 12:42 pm

The results also show alpha=zero for both funds, check the p values. As far as we can infer, the difference in alpha is purely noise.

It is essentially meaningless to pick three funds, chosen ex post, to illustrate much of anything.

It would make more sense to look at the universe of funds in this space. Consider a long enough time to draw some inferences. Eliminate survivorship bias.

Then compare risk adjusted performance vs cost (we think we know what that would show) and ask whether the advantage of details of fund management out weigh the cost. One would have to control for factor exposure.

Compare true index funds to
1. Actively managed funds that, in general, target small value but put few constraints on individual stock picking.
2. Actively managed funds that build in a set of binding rules from which the managers cannot stray as they attempt to find stocks that will do better than the index alone. I don't know how many such funds they may be.
3. Actively managed funds that promise to use some constraints but what they are, how carefully they are followed, how they are weighted and when they change are unknown to the investors.

These group 3 actively managed funds seem to be the ones called "passive" but that hardly seems to be an appropriate term.

Considering all the funds in the universe, classifying the active funds as above, using risk adjusted returns, one could ask whether the active share of the active funds was making a positive contribution to performance after costs.

Until someone does this, there is no way to know.

The comparison to "index funds" seems intentionally misleading.
There is nothing that prevents an index fund from using patient trading. There is no god of indexing who will smite down those who don't immediately trade to match changes in its components.
One can use options and futures to maintain exposure to the index and minimize tracking error.
These moves are routine in index fund management.

How often an index is reconstituted depends on a decision by whoever creates it. A fund company can pick what index it wants or create one itself.
If one wants screens on other factors, those too can be built into an index.

Stocks will move into and out of an index depending on how it is defined. There is very little such movement in VTI, which of course is an index fund.

Exactly.

From portfolio visualizer, longest returns available: 1/01-12/18
DFSVX 9.51%
IJS 8.55%
https://www.portfoliovisualizer.com/bac ... ion2_2=100

And here’s the factor analysis. DFSVX has somewhat more exposure to size and value. It’s alpha is +0.20% v. IJS -0.15%. Perhaps that alpha represents patient trading, buy/hold ranges, screening out IPOs, extreme small growth, penny stocks, bankruptcy?
https://www.portfoliovisualizer.com/fac ... sion=false

Dave
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama

garlandwhizzer
Posts: 2568
Joined: Fri Aug 06, 2010 3:42 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by garlandwhizzer » Fri Feb 01, 2019 1:04 pm

cheezit wrote:

If you want deeper exposure why not look at RZV, which tracks the S&P 600 Pure Value index? As you see in the chart from my earlier post, it had a higher exposure to size, value and quality than DFSVX over the period in the article. If you look at the entire lifespan of RZV (back to April of 2006), the differences become even larger, with RZV having a much larger size loading (1.29 vs 0.87), value loading (0.94 vs 0.44), quality loading (0.34 vs 0.09), and a lower correlation to the total US market (0.80 vs 0.92) compared to DFSVX. And the ER is lower to boot (0.35 vs 0.52). If I wanted to tilt hard, I'd give it a long look.
All this is true. The other thing that is true is that if you look all the back to RZV's inception almost 13 years ago, it has massively and consistently underperformed both TSM and Vanguard Small Cap Index. Morningstar rates RZV 1 star, their lowest rating largely for its pathetic performance in spite of its compelling factor loads. Such is the gulf between factor theory and real performance in a real funds over 13 years. So many on the Forum assume without question that factor loadings accurately project future returns. Just load up with SCV, control for junk/quality, and the magic happens. Over this time period that was clearly not true in the case of RZV which had better loadings for value, size, and quality than did DFSVX and vastly greater than TSM. What killed its results was not factor exposure at all but an annualized alpha of minus 6.02%. Alpha is the cheat factor in factor analysis which is necessary to makes the factor numbers equal reality. When factor returns don't add up to what real returns were we call that difference alpha. What alpha means in effect is that factor analysis does not tell the whole story about what future returns are going to be. Often portfolios that look very compelling on factor analysis turn out to have very high levels of negative alpha just like RZV which substantially reduces their real returns. Producing future outperformance is much more complicated than just getting attractive factor loads at least in this 13 year period with a real fund, RZV. Since its inception returns of RZV are now so far behind simple indexes now such as TSM or VB that it appears very unlikely to catch up with them for a long, long time, if ever. It is perhaps more likely that RZV will close due to its dwindling asset base as investors flee which they have been doing for years now.

Garland Whizzer

fennewaldaj
Posts: 823
Joined: Sun Oct 22, 2017 11:30 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by fennewaldaj » Fri Feb 01, 2019 1:18 pm

garlandwhizzer wrote:
Fri Feb 01, 2019 1:04 pm

but an annualized alpha of minus 6.02%.
Garland Whizzer
Hmm what i got doesn't look so bad Apr 2006 -Dec 2018
FF 3 factor Market 1.18 Size 1.19 Value 1.05 Annualized alpha -0.14%
FF 4 factor Market 1.06 Size 1.19 Value 0.72 annualized alpha 0.46%
FF 5 factor Market 1.19 Size 1.29 Value 0.94 Profitability 0.34 Annualized alpha -1.40%
Last edited by fennewaldaj on Fri Feb 01, 2019 1:28 pm, edited 1 time in total.

fennewaldaj
Posts: 823
Joined: Sun Oct 22, 2017 11:30 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by fennewaldaj » Fri Feb 01, 2019 1:28 pm

Random Walker wrote:
Fri Feb 01, 2019 10:46 am
An active fund can use formulaic screens, but after a stock passes through the screens, it’s up to an individual to subjectively decide whether to buy.

Dave
I still think this active passive thing is a continuum. Like this fund I have a position in is a low turnover quant fund. They have no tenancy to style drift. There process is entirely quantitative there is no subjective deciding. There turn over is similar to Vanguard and DFA small value funds. However there screens screen the Russel 2k down to 252 stocks.

http://portfolios.morningstar.com/fund/ ... ture=en-US

Or using the S+P 600 instead of the Russel 2k is saying that the S+P committee may be adding value. And they are getting rid of 2/3 of the selection universe.

User avatar
Socrates
Posts: 426
Joined: Sun May 13, 2018 10:27 pm
Location: Margaritaville

Big fan of IJS

Post by Socrates » Fri Feb 01, 2019 1:40 pm

for small cap value
“Don't waste your time looking back. You're not going that way.” ― Ragnar Lothbrok.

User avatar
Robert T
Posts: 2610
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Robert T » Sat Feb 02, 2019 2:42 am

.
As far as I can tell from the analysis below, for those not wanting a 100% value tilt (but still wanting some value tilt), iShares S&P SmallCap 600 Value (IJS) since its inception over 18 years ago has been a better choice than DFA Small Value (DFSVX). Obviously no guarantees going forward.

Claimed positive (and significant) alpha of DFA funds is oversold.

We were often told (not by Larry) that DFA provides 1-3% in additional alpha beyond factor exposure and that DFA’s ‘superior’ implementation was a key driver of this additional and persistent alpha (patient trading, IPO exclusions, hold ranges, block trading, momentum screens etc, etc). I would just note that this claim was at odds with Fama-French’s own research (e.g. “Luck vs. Skill in Mutual Fund Performance”), and Don Keim’s research (on “An analysis of mutual fund design: the case of investing in small-cap stocks”). Keim’s research showed that DFA’s stock inclusion/exclusion rules (e.g. hold ranges, IPO exclusions, low price/market cap restrictions) and trading strategy (e.g. patient and block trading) simply changed factor exposure rather than added alpha.
If anything, the alpha on IJS has been larger than on DFSVX, but would just note that both are not statistically different from zero. So over this period, no 1-3% alpha from DFSVX, and importantly no better alpha than lower costs alternatives.

What about the view that DFSVX has higher small cap and value loads so you just need to use less of it and can combine it with a total market fund for lower overall cost.
  • August 2000 (since inception of IJS) to December 2018:

    Market / Size / Value loads
    1.02 / 0.94 / 0.56 = DFSVX
    0.95 / 0.82 / 0.49 = IJS
    1.02 / 0.83 / 0.49 = 88% DFSVX / 12% VTSMX
It is true that over this period DFSVX had higher factor loads than IJS. Combining 88% DFSVX with 12% VTSMX/VTI would have given very similar size and value loads as IJS but with a 0.21% higher expense ratio for the combination.
  • Expense ratio
    0.52 = DFSVX
    0.25 = IJS
    0.46 = 88% DFSVX / 12% VTSMX*

    * the lower expense ratio on VTI was used in the calculation
And over this period IJS would have outperformed the DFSVX/VTSMX combination (costs matter), even though the DFSVX/VTSMX combination had a higher market beta https://www.portfoliovisualizer.com/bac ... ion3_2=100 . The above example is illustrative, but for any factor-matched combination using a higher share of VTSMX/VTI, the expense ratio of the portfolio with IJS would be lower than with DFSVX. For factor-matched portfolios total cost is the only thing that matters. The above analysis excludes any advisor fee to use DFSVX.

What about the view that using DFSVX (with its higher size and value load) allows an investor to lower overall portfolio beta while still maintaining relatively high portfolio size and value loads (be more ‘diversified’ across factors).


IJS already has lower beta that DFSVX (0.95 vs. 1.02) – comparing more equivalent market beta portfolios
  • Market / Size / Value loads
    1.02 / 0.94 / 0.56 = DFSVX
    0.95 / 0.82 / 0.49 = IJS
    0.95 / 0.87 / 0.52 = 93% DFSVX / 7% T-bills
So for the same market beta, you would have got a 0.05 higher size load, and 0.03 higher value load with the DFSVX/T-bill combination for a 0.23% higher expense ratio (+ whatever advisor fee you are paying for access). For lower market beta combinations (higher t-bill/bond allocations) there is an even smaller difference in portfolio size and value loads.

The above analysis ignores taxes, which would have been much lower for IJS (M* 15 year tax cost ratio for DFSVX = 1.15, for IJS = 0.37).

Robert
.

BetaTracker
Posts: 196
Joined: Fri Sep 21, 2012 7:57 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by BetaTracker » Sat Feb 02, 2019 6:11 am

I'm not sure either the Keim research or the Fama/French paper on luck vs. skill present the most appropriate sets of data to apply to DFA's factor-based investing process, particularly since it started to fully implement profitability screens into its asset class coverage. In research and independent analysis I've read, DFA itself references an earlier work published in the Journal of Financial Economics by Fama and French (“Profitability, Investment, and Average Returns”).
Professor Robert Novy-Marx's research has also been cited as influential in this regard. His work (“The Other Side of Value: The Gross Profitability Premium”) built on Fama/French's previous work on empirical asset pricing in relationship to excluding some forms of non-recurring costs of current profitability as a better estimate for expected future profitability.
More recently, I've seen research by Professor Sunil Wahal also being cited about continuing refinements to factor analysis that combines profitability with size and value factors.
The academic argument being made by Fama, French, Navy-Marx and Wahal (and that's being implemented in the form of DFA's funds) is that combining profitability to size and value factors smooths return patterns over time and produces better annualized returns in small cap equities.
Remember, too, the value premium hasn't been very forceful during this latest bull market and, like size, tends to occur in more variable patterns the shorter the period studied.
I'd rather be content than happy -- Lao Tzu.

User avatar
nedsaid
Posts: 12980
Joined: Fri Nov 23, 2012 12:33 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by nedsaid » Sat Feb 02, 2019 10:33 am

This is what is so frustrating about factor investing, you not only have to invest in the factors but also have to invest in the "right" funds to capture those factors. It seems whatever I am invested in are the "wrong" funds. I don't know, the Vanguard Small-Cap Value Index ETF and the S&P 600 Small Value Index ETFs have been good enough for me and both have performed fairly well. What is even more frustrating is that we have been in mostly a Large Growth market for a decade now.
A fool and his money are good for business.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Sat Feb 02, 2019 10:52 am

https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave

User avatar
GreatOdinsRaven
Posts: 551
Joined: Thu Apr 23, 2015 8:47 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by GreatOdinsRaven » Sat Feb 02, 2019 11:56 am

hdas wrote:
Thu Jan 31, 2019 4:03 pm




Mr. Swedroe adds:
While VISVX has an expense ratio of 0.19% (their Admiral shares version costs just 0.07%) and DFSVX has an expense ratio of 0.52%, the higher costs of the Dimensional fund have been more than offset by their greater exposure to the size and value factors and their focus on adding value by minimizing the negatives of pure indexing. For the full period that both have existed—June 1998 through December 2018—DFSVX returned 8.6%, outperforming the lower cost VISVX, which returned 7.9%.
0.7% higher returns with higher volatility. Reminds me of Rick Ferri who says "Alpha goes to the advisor." - because it does... In fact I still think that most people are still under-performing the factor index fund/factor index ETF after fees, because I would imagine the average AUM fee charged by most DFA-approved firms to their customers exceeds 0.7%.

People, if you want an advisor for wealth management planning, tax planning, etc go for it. They provide an incredibly valuable service that will save most people a considerable amount of money and headaches. I am a HUGE fan of RIA fiduciaries. But, don't use an AUM RIA just for DFA access unless you have no interest in doing your own portfolio and investment management.

All that work of having a SCV-tilted portfolio with its higher volatility and tracking error just to break even after advisory fees compared to a factor or market beta index fund that you can set and forget. Literally, set up automated investments in a Vanguard factor index fund or market cap beta index fund and walk away for your entire investing life. Or, if you want to get a little bit fancy, buy a cheap-ish SCV ETF like IJS (the S&P 600 SCV etf) and capture a significant factor exposure (similar size (smb), but smaller value (hml) exposure compared to the DFA SCV fund DFSVX) with NO advisory or platform access costs. That choice is a no-brainer for me.

How many people do you think stuck with the plan and even achieved that 0.7% out-performance before advisory fees during the 2011-2018 period- even with an advisor? I bet it's less than you'd think, even with the behavioral coaching.

Now, I fully accept the argument that if you don't want to handle your own portfolio and if you want a factor tilt it makes sense to use an advisor. In essence you are getting your advisory services "for nearly free" in the above example. YMMV and past performance does not guarantee future results, etc.

And, yes, I know that you can get access to DFA for pennies and not have to pay direct AUM fees (for example, I have platform access and my all-in "advisory" costs average to low single digit bps. Some people have access in their retirement accounts and 529s). But, as we all know, most people do not use such an option. I use DFA for TA-international core and targeted value/ISCV, but I enjoy the tax-efficiency of ETFs like IJS and its TLH partner VBR for my domestic SCV exposure. I also sort of have exposure to Mel's unloved mid-caps (I use MCV) via Vanguard index funds as well as a large slug of TSM. It's good enough and I'm not paying out the nose for it, relative to my net worth.

:beer :moneybag

GOR
"The greatest enemies of the equity investor are expenses and emotions." -John C. Bogle, Little Book of Common Sense Investing. | | "Winter is coming." Lord Eddard Stark.

User avatar
GreatOdinsRaven
Posts: 551
Joined: Thu Apr 23, 2015 8:47 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by GreatOdinsRaven » Sat Feb 02, 2019 12:11 pm

Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
That's a great question and I've wondered it, myself. I think so much depends on luck and timing when you're comparing investments that are arguably very similar.

As a person who is in the accumulation phase of life I've been interested in the money weighted returns of investments, too, rather than just the time weighted returns. It's just so easy to manipulate the end results of returns. For example, no doubt that DFSVX has out-performed IJS since inception as you note.

But how has it done for people who continued to make contributions to the funds since that time? Well, we can find many examples where DFA does better or where IJS does better. Depends on when and how much you are able to invest (and things that like depends on varying life circumstances).

For example, if you started with a $10,000 lump sum at inception (January 2001) and you invested $1,000 quarterly, by January 2019 your results would be different.

In this random arguably cherry picked exmaple the money weighted return of IJS was better than DFSVX, before fees (9.58% vs 9.32% dollar weighted/money weighted returns). But to be very clear IJS has a lower time weighted return of 9.2% vs 10.15%. It's the first example I tried in Portfolio Visualizer.



Portfolio Initial Balance Final Balance CAGR TWRR MWRR Stdev
DFSVX $10,000 $260,203 19.75% 10.15% 9.32% 20.55%

IJS $10,000 $267,777 19.94% 9.20% 9.58% 18.94%

VFINX(S&P 500) $10,000 $238,880 19.18% 6.00% 8.55% 14.46%
"The greatest enemies of the equity investor are expenses and emotions." -John C. Bogle, Little Book of Common Sense Investing. | | "Winter is coming." Lord Eddard Stark.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by DaufuskieNate » Sat Feb 02, 2019 1:16 pm

Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
You can do a factor analysis on PortfolioVisualizer and click on the Performance Attribution tab. According to this analysis, over the specific timeframe you mention, DFSVX earned 10.11 bp per month on average more than IJS. This breaks down as follows:

Higher SmB Exposure: 3.04 bp
Higher Market Exposure: 2.88 bp
Higher Alpha: 2.87 bp
Higher HmL Exposure: 1.32 bp

cheezit
Posts: 268
Joined: Sat Jul 14, 2018 7:28 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by cheezit » Sat Feb 02, 2019 4:39 pm

Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
I assume it was a simple oversight that caused you to omit the first few months of IJS's existence from your comparison. If you select "month to month" instead of "year to year" in your link on PV you can see them. It narrows the gap to 0.3%.

User avatar
Robert T
Posts: 2610
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Robert T » Sat Feb 02, 2019 5:07 pm

Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
Since 08/2000 (inception of IJS) to January 2019: Annualized return

DFSVX = 10.11%
IJS = 9.81%
Difference = 0.30%
https://www.portfoliovisualizer.com/bac ... ion2_2=100

As per my previous post, the difference is not caused by higher alpha. The fact is that the alpha on DFSVX was lower than IJS by 0.27% annual average over this period (using the alpha numbers from my previous post: 0.37 - 0.64 = -0.27%) This is equivalent to the expense ratio difference between IJS and DFSVX (0.25 - 0.52 = -0.27%).

DFSVX had a 0.07, 0.12, and 0.07 higher market, size, and value load respectively than IJS (i.e. higher risk following Fama-French). This only mattered if you were holding 100% DFSVX (in tax-advantaged account with close to zero adviser fee for access). As in my previous post/analysis, for lessor target portfolio market, size and value loads than DFSVX, using IJS in a portfolio to achieve these factor load targets seems to have been the better choice over this time period. Obviously no guarantees going forward.

Robert
.

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Sat Feb 02, 2019 6:21 pm

cheezit wrote:
Sat Feb 02, 2019 4:39 pm
Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
I assume it was a simple oversight that caused you to omit the first few months of IJS's existence from your comparison. If you select "month to month" instead of "year to year" in your link on PV you can see them. It narrows the gap to 0.3%.
Yes oversight. Thanks for month to month tip

Dave

Dead Man Walking
Posts: 904
Joined: Wed Nov 07, 2007 6:51 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Dead Man Walking » Sat Feb 02, 2019 10:06 pm

Random Walker wrote:
Sat Feb 02, 2019 6:21 pm
cheezit wrote:
Sat Feb 02, 2019 4:39 pm
Random Walker wrote:
Sat Feb 02, 2019 10:52 am
https://www.portfoliovisualizer.com/bac ... ion2_2=100

The scoreboard 1/01-1/19 is DFSVX 10.15% v. IJS 9.20%. So, is the difference attributable to the slightly different factor exposures or to alpha caused by DFA passive improvements over straight indexing?

Dave
I assume it was a simple oversight that caused you to omit the first few months of IJS's existence from your comparison. If you select "month to month" instead of "year to year" in your link on PV you can see them. It narrows the gap to 0.3%.
Yes oversight. Thanks for month to month tip

Dave
As usual, this thread about small cap value funds was informative and entertaining. However, I can’t help but notice that the bottom line of the discussion ended with performance comparisons. Active-passive or somewhere in between, it always boils down to performance.

DMW

Topic Author
Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: A Tale Of 3 Small Value Funds

Post by Random Walker » Sat Feb 02, 2019 10:14 pm

Dead Man Walking wrote:
Sat Feb 02, 2019 10:06 pm
As usual, this thread about small cap value funds was informative and entertaining. However, I can’t help but notice that the bottom line of the discussion ended with performance comparisons. Active-passive or somewhere in between, it always boils down to performance.

DMW
I’m not a big fan of looking at past performance comparisons myself. Of course past data is significant. But we invest looking forward. That is why I focus mostly on the rationale behind building efficient portfolios: expected returns, risk based and behavioral based explanations behind expected premia, correlations, volatilities, the effect of individual assets on the portfolio as a whole.

Dave

Post Reply