DFA small value... acitve management?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
packer16
Posts: 1044
Joined: Sat Jan 04, 2014 2:28 pm

Re: DFA small value... acitve management?

Post by packer16 » Mon Oct 22, 2018 9:51 am

Dave,

You do not think human judgement is involved in factor or DFA funds? How do they decide if the active trading strategy is worth it to do it or not for a given stock? How do they decide the weighting of each stock in the portfolio? In each case human judgement. The only difference in quant investing is the computer is helping the human to make judgments based upon the logic the human puts into the computer algorithm. The ultimate choice is made by human not a computer.

As to agnostic to stock selection and timing, I disagree. A small cap value fund is all about stock selection as most of the stocks in broad market indices are not included. Rather than the manager selecting stocks, you are doing the selection based upon an empirically derived inductive relationship you think will occur in the future (SCV will provide a better risk/reward relationship than TSM). Also, there is timing as firms move in & out of the SCV indices all the time. So a SCV is an active fund that practices both stock selection & timing versus a total stock market index where these activities only occur if there is an index change. Also, SCV does not utilize the market signal of pricing to weight stocks (as the holder thinks the PMs weights are better than what the market can provide) versus TSM which incorporates these market signals.

Packer
Buy cheap and something good might happen

Random Walker
Posts: 3171
Joined: Fri Feb 23, 2007 8:21 pm

Re: DFA small value... acitve management?

Post by Random Walker » Mon Oct 22, 2018 9:54 am

packer16 wrote:
Mon Oct 22, 2018 9:24 am
Dave,

Is not the smart trading you mention an active trading strategy? If they are doing this how can this be passive? I agree it is formulaic but passive IMO means no or little trading (only when the index constituents change, not when they think they can gain a trading advantage). It appears to be part of the strategy of the fund so how can it be passive? These guys think they know how to beat the market (market-cap indices) & are running their fund that way just like other active managers.

Packer
We just see it differently. It’s formulaic and mechanical access to an additional factor that has been shown to explain returns, momentum. Passive because it’s implemented in that fashion without paying any attention to what the stock is or market environment is. It’s not necessarily trying to beat the market, it’s accessing the factors shown to drive returns. Trading can be very passive as long as it’s formulaic. It’s about the implementation.

Dave

Random Walker
Posts: 3171
Joined: Fri Feb 23, 2007 8:21 pm

Re: DFA small value... acitve management?

Post by Random Walker » Mon Oct 22, 2018 10:12 am

Packer,
Don’t believe that SCV provides superior risk/reward relationship to TSM, just a different one. One of the big F-F findings is that size and value are independent unique factors from each other and from market factor. TSM investor is accessing one source of risk, market beta. SCV investor is accessing 3 sources: market, size, value. Not superior risk/reward, just risk/ reward more diversified across the sources of risk.

As far as I know, the DFA funds are market cap weighted. That is the starting point within a universe first defined my market cap and price. Excluding some of the lottery ticket stocks is icing on the market cap weighted cake. Patient trading, buy/hold ranges, tax management, I view as pretty passive. Guess I can see how someone can view it as active compared to rigid index adherence, but its a completely different universe from human intuition. If the human is rigidly implementing well defined rules consistently, that’s pretty passive in my book.

Dave

User avatar
nedsaid
Posts: 10486
Joined: Fri Nov 23, 2012 12:33 pm

Re: DFA small value... acitve management?

Post by nedsaid » Mon Oct 22, 2018 10:27 am

nisiprius wrote:
Sat Oct 20, 2018 6:00 am
(Shrug) DFA funds are mildly active, in a gentle sort of way. That is why none of them call themselves "index funds." I don't see why there should be any mystery about this except that people want to obfuscate the issue. Their "patient trading" technique is also mildly active, they do some market timing of individual stock purchases although they don't call it that.
Yes, I have pointed this out many times. There is a continuum between active and passive and the factor funds are mostly passive but have elements of active management. Even some forms of indexing fall into this continuum, lets take for example a Small Value Index or a Momentum Index. To me, no matter what you call it, when portfolio turnover gets to be beyond a certain point, it is no longer passive. Also, the more complex the rules for selecting stocks for a "passive" fund, the smaller the universe and the greater the portfolio turnover. Pretty much, if the robots do it, it is passive because it is all rules based. If people do it, it is active almost by definition. I don't buy it.
A fool and his money are good for business.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Mon Oct 22, 2018 2:56 pm

moghopper wrote:
Sat Oct 20, 2018 8:15 pm
jalbert wrote:
Sat Oct 20, 2018 12:29 pm
Yes, but the index funds that track those indices are passively managed. The difference between that and what DFA does is that index fund managers do not control the index changes.
I find this to be arbitrary. If DFA had an outside consultant make an "index rule" and then stuck to that - then it would be an index? Or it would only then be passive?

Sorry, but I think most of this is mental gymnastics to make people feel better.
It would be fully passive if a 3rd party controlled the index they tracked. This is not an arbitrary distinction.

One important difference is seen when looking at fund tracking error. When a passively managed fund tracks an index managed by a 3rd party, tracking error may be measured objectively by comparing fund return to index return. Tracking error is purely a measure of how well the fund managers replicated the performance of the index they track.

When the index being tracked is proprietary non-public intellectual property of the fund manager, things get more squishy. You can compare performance to some 3rd party index, but this blurs whether you are seeing passive tracking error or variance in performance due to holding a portfolio different from the 3rd party index portfolio. This blurring of the two sources of tracking error is a major reason why such a fund is more active than one which tracks an index controlled by a 3rd party.

I think the real issue is that passive vs active is not an all-or-nothing thing but a range of strategies. DFA funds are more passive than a fund actively managed using fundamental analysis, but less passive than a fund that tracks an externally controlled index.
Risk is not a guarantor of return.

Elysium
Posts: 1291
Joined: Mon Apr 02, 2007 6:22 pm

Re: DFA small value... acitve management?

Post by Elysium » Mon Oct 22, 2018 5:40 pm

jalbert wrote:
Mon Oct 22, 2018 2:56 pm
moghopper wrote:
Sat Oct 20, 2018 8:15 pm
jalbert wrote:
Sat Oct 20, 2018 12:29 pm
Yes, but the index funds that track those indices are passively managed. The difference between that and what DFA does is that index fund managers do not control the index changes.
I find this to be arbitrary. If DFA had an outside consultant make an "index rule" and then stuck to that - then it would be an index? Or it would only then be passive?

Sorry, but I think most of this is mental gymnastics to make people feel better.
It would be fully passive if a 3rd party controlled the index they tracked. This is not an arbitrary distinction.

One important difference is seen when looking at fund tracking error. When a passively managed fund tracks an index managed by a 3rd party, tracking error may be measured objectively by comparing fund return to index return. Tracking error is purely a measure of how well the fund managers replicated the performance of the index they track.

When the index being tracked is proprietary non-public intellectual property of the fund manager, things get more squishy. You can compare performance to some 3rd party index, but this blurs whether you are seeing passive tracking error or variance in performance due to holding a portfolio different from the 3rd party index portfolio. This blurring of the two sources of tracking error is a major reason why such a fund is more active than one which tracks an index controlled by a 3rd party.

I think the real issue is that passive vs active is not an all-or-nothing thing but a range of strategies. DFA funds are more passive than a fund actively managed using fundamental analysis, but less passive than a fund that tracks an externally controlled index.
DFA funds are not tracking any 3rd party indexes, but they do have 3rd party indexes as benchmark. This is mainly for legal reasons I guess. Otherwise they do their own thing about capturing whatever factor that they intend to capture. I don't know if they have any internal measures to check whether they are tracking it accurately, I guess they do, but this isn't shared with public. Instead the public can see the comparison with 3rd party index which they don't track closely, so it is sort of absurd anyway. I have never paid attention to the performance of such indexes to compare the DFA funds I have owned, and instead thought of their returns in absolute terms. If they were competitive, then it didn't bother me, and if they weren't then I would start worrying, but it would take a lot more than couple of hundred basis points and more than five years for me to start getting concerned. Because you have to let these funds have flexibility that they need to execute their strategy perfectly in order to capture the maximum benefit they can from it, and that requires patience. In that regard, they are somewhat similar to holding an active managed fund.

Random Walker
Posts: 3171
Joined: Fri Feb 23, 2007 8:21 pm

Re: DFA small value... acitve management?

Post by Random Walker » Mon Oct 22, 2018 5:55 pm

Elysium wrote:
Mon Oct 22, 2018 5:40 pm
jalbert wrote:
Mon Oct 22, 2018 2:56 pm
moghopper wrote:
Sat Oct 20, 2018 8:15 pm
jalbert wrote:
Sat Oct 20, 2018 12:29 pm
Yes, but the index funds that track those indices are passively managed. The difference between that and what DFA does is that index fund managers do not control the index changes.
I find this to be arbitrary. If DFA had an outside consultant make an "index rule" and then stuck to that - then it would be an index? Or it would only then be passive?

Sorry, but I think most of this is mental gymnastics to make people feel better.
It would be fully passive if a 3rd party controlled the index they tracked. This is not an arbitrary distinction.

One important difference is seen when looking at fund tracking error. When a passively managed fund tracks an index managed by a 3rd party, tracking error may be measured objectively by comparing fund return to index return. Tracking error is purely a measure of how well the fund managers replicated the performance of the index they track.

When the index being tracked is proprietary non-public intellectual property of the fund manager, things get more squishy. You can compare performance to some 3rd party index, but this blurs whether you are seeing passive tracking error or variance in performance due to holding a portfolio different from the 3rd party index portfolio. This blurring of the two sources of tracking error is a major reason why such a fund is more active than one which tracks an index controlled by a 3rd party.

I think the real issue is that passive vs active is not an all-or-nothing thing but a range of strategies. DFA funds are more passive than a fund actively managed using fundamental analysis, but less passive than a fund that tracks an externally controlled index.
DFA funds are not tracking any 3rd party indexes, but they do have 3rd party indexes as benchmark. This is mainly for legal reasons I guess. Otherwise they do their own thing about capturing whatever factor that they intend to capture. I don't know if they have any internal measures to check whether they are tracking it accurately, I guess they do, but this isn't shared with public. Instead the public can see the comparison with 3rd party index which they don't track closely, so it is sort of absurd anyway. I have never paid attention to the performance of such indexes to compare the DFA funds I have owned, and instead thought of their returns in absolute terms. If they were competitive, then it didn't bother me, and if they weren't then I would start worrying, but it would take a lot more than couple of hundred basis points and more than five years for me to start getting concerned. Because you have to let these funds have flexibility that they need to execute their strategy perfectly in order to capture the maximum benefit they can from it, and that requires patience. In that regard, they are somewhat similar to holding an active managed fund.
Elysium,
I agree with you. I once made a post stating that I was told by an advisor that “DFA funds sort of serve as their own indexes”. I commented that I think that is totally true, and was lambasted for being naive. But it is true. They target certain factor exposures, and as long as they adhere to those exposures, they are doing their job. When the factors outperform they will look good and when the factors underperform they will not. I assume DFA performs attribution analyses that they review internally and with their advisors.

Dave

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Mon Oct 22, 2018 6:25 pm

DFA funds are not tracking any 3rd party indexes, but they do have 3rd party indexes as benchmark. This is mainly for legal reasons I guess.
All funds should be benchmarked against some objective measure of return. It is interesting that they recommend benchmarking with the Russell 2000 which has had well documented hits to performance from front-running. This gives any product benchmarked against it a leg up if the measurement period includes times when this was an issue.

I don’t know why this is even a contentious issue. DFA has excellent products if you want the types of products they offer. Semantically, you can define passive management however it suits your fancy. DFA funds do not achieve the same level of transparency as an index fund that tracks an externally controlled index.

DFA uses internal de facto indices by screening for stocks to include in their investable domain based on factor exposure, and holding them at market cap weight.
Risk is not a guarantor of return.

User avatar
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: DFA small value... acitve management?

Post by Robert T » Mon Oct 22, 2018 6:55 pm

.
On “Dimensional Indexes” - the DFA Matrix Book 2018 includes many of them
https://www.fplcapital.com/wp-content/u ... k-2018.pdf

For example it has something called the “Dimensional US Small Cap Value Index”. One would think that this is what is broadly tracked by the DFA US Small Value Fund (DFSVX) – at least since 2014 when the DFA funds had started including profitability screens (as in the indexes in the matrix book).

Here’s a % return comparison.

Dimensional US Small Value Index / DFSVX

2014…..3.4 /.…3.5
2015…-8.4 /…-7.8
2016.. 37.2 /...28.3
2017…. 7.6 /….7.2

Annualized return over last 4 years (2014-2017):

Dimensional US Small Value Index = 8.7%
DFA US Small Value Fund (DFSVX) = 7.0%
Difference = -1.7% (driven mainly by the large difference in returns in 2016)

Not sure why there is such a large difference between the DFA fund and the Dimensional Index.

Robert
.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Tue Oct 23, 2018 6:54 pm

Robert T wrote:
Mon Oct 22, 2018 6:55 pm
On “Dimensional Indexes” - the DFA Matrix Book 2018 includes many of them
https://www.fplcapital.com/wp-content/u ... k-2018.pdf

For example it has something called the “Dimensional US Small Cap Value Index”. One would think that this is what is broadly tracked by the DFA US Small Value Fund (DFSVX) – at least since 2014 when the DFA funds had started including profitability screens (as in the indexes in the matrix book).

Here’s a % return comparison.

Dimensional US Small Value Index / DFSVX

2014…..3.4 /.…3.5
2015…-8.4 /…-7.8
2016.. 37.2 /...28.3
2017…. 7.6 /….7.2

Annualized return over last 4 years (2014-2017):

Dimensional US Small Value Index = 8.7%
DFA US Small Value Fund (DFSVX) = 7.0%
Difference = -1.7% (driven mainly by the large difference in returns in 2016)

Not sure why there is such a large difference between the DFA fund and the Dimensional Index.

Robert
.



well the index says this:
wrote: "DIMENSIONAL US SMALL CAP VALUE INDEX
January 1975–present
Compiled by Dimensional from CRSP and Compustat data. Targets
securities of US companies traded on the NYSE, NYSE MKT (formerly AMEX),
and Nasdaq Global Market whose relative price is in the bottom 35% of the
Dimensional US Small Cap Index after the exclusion of utilities, companies
lacking financial data, and companies with negative relative price. The
index emphasizes securities with higher profitability, lower relative price,
and lower market capitalization. Profitability is measured as operating
income before depreciation and amortization minus interest expense
scaled by book. Exclusions: non-US companies, REITs, UITs, and investment
companies. The index has been retroactively calculated by Dimensional
and did not exist prior to March 2007. The calculation methodology was
amended in January 2014 to include profitability as a factor in selecting
securities for inclusion in the index.
whereas the prospectsus for the fund says this:
wrote: "The Advisor may adjust the representation in the U.S. Small Cap Value Portfolio of an eligible company, or exclude a company, after considering such factors as free float, momentum, trading strategies, liquidity, value, profitability, and other factors that the Advisor determines to be appropriate, given market conditions"
it looks like the main factor the fund uses that its benchmark index does not is momentum. so i suspect 2016 is a story of momentum bets gone awry

yep. here is an excerpt from the fama and french series for the momentum factor:
wrote: 2008 13.22
2009 -82.68
2010 6.01
2011 7.26
2012 1.57
2013 7.9
2014 1.6
2015 20.66
2016 -20.34
2017 5.04
so 2016 was a bad momentum year at -20.3%. Personally i am a real skeptic on momentum. I think even before trading costs the CAGR returns to the momentum factor are negative or poor as far back as the eye can see. It's a typical hedge fund strategy that works in a small way for a while until it blows up in a big way.

cheers,
grok
Keep calm and Boglehead on. KCBO.

User avatar
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: DFA small value... acitve management?

Post by Robert T » Tue Oct 23, 2018 8:56 pm

.
Thanks grok87.

Yes - looks like the 'momentum' screens cost DFSVX in 2016 relative to the Dimensional US Small Value Index.

Interestingly DFA added the momentum screens in 2003, as I recall, and the relatively large negative momentum in 2003, 2009, and 2016 put a drag on DFSVX returns in those years relative to no momentum screens.

Returns: DFSVX / Dimensional US Small Value Index

2003 ...59,4 / ...66.0
2009...33.6 / ....51.2
2016 ...28.3 / ...37.2

Annualized returns 2003-2017:
  • DFA US Small Value (DFSVX) = 11.9%
    Dimensional US Small Value Index = 13.2%
    Difference = -1.3% (most of which is explained by differences in 2003, 2009, 2016).
Would just note that over this 15 period (2003-2017) the 'momentum premium' for small caps was negative (in contrast to all other 15 year periods since 1927). No knowing what the next 15 years will bring.

Robert
.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Wed Oct 24, 2018 6:47 am

Robert T wrote:
Tue Oct 23, 2018 8:56 pm
.
Thanks grok87.

Yes - looks like the 'momentum' screens cost DFSVX in 2016 relative to the Dimensional US Small Value Index.

Interestingly DFA added the momentum screens in 2003, as I recall, and the relatively large negative momentum in 2003, 2009, and 2016 put a drag on DFSVX returns in those years relative to no momentum screens.

Returns: DFSVX / Dimensional US Small Value Index

2003 ...59,4 / ...66.0
2009...33.6 / ....51.2
2016 ...28.3 / ...37.2

Annualized returns 2003-2017:
  • DFA US Small Value (DFSVX) = 11.9%
    Dimensional US Small Value Index = 13.2%
    Difference = -1.3% (most of which is explained by differences in 2003, 2009, 2016).
Would just note that over this 15 period (2003-2017) the 'momentum premium' for small caps was negative (in contrast to all other 15 year periods since 1927). No knowing what the next 15 years will bring.

Robert
.
Thanks Robert. It’s interesting that as soon as they added it as a factor it stopped working! I wonder if that is a coincidence...
Cheers
Grok
Keep calm and Boglehead on. KCBO.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Wed Oct 24, 2018 4:15 pm

Quite possibly it is not a coincidence. But there are only 5 independent 15-year periods from 1927 to 2003 so that was also a very small sample to establish the effect in the first place. It’s absence or presence could just be an artifact of chance.

But if the time period for assessing a momentum effect is known, it seems like it would be susceptible to front-running. The periods being exploited to measure momentum effects would tend to get shorter over time as market participants tried to be the first to buy a stock deemed to have momentum until any such effect were gone, if it indeed existed in the first place.

I also am skeptical that if a stock exhibits momentum for some period of time, the time period would be constant. This would mean that you have no way of knowing if the effect is exhausted or exploitable. This behavior is also statistically indistinguishable from the whole effect being an artifact of chance.
Last edited by jalbert on Thu Oct 25, 2018 2:40 am, edited 1 time in total.
Risk is not a guarantor of return.

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Wed Oct 24, 2018 4:22 pm

jalbert wrote:
Mon Oct 22, 2018 6:25 pm
I don’t know why this is even a contentious issue. DFA has excellent products if you want the types of products they offer. Semantically, you can define passive management however it suits your fancy. DFA funds do not achieve the same level of transparency as an index fund that tracks an externally controlled index.

DFA uses internal de facto indices by screening for stocks to include in their investable domain based on factor exposure, and holding them at market cap weight.
I just sat through a pretension by DFA today. They are not passive.

They are top down, discipline, quantitative and academically driven. They are consonantly fiddling around with their factors.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Wed Oct 24, 2018 4:42 pm

alex_686 wrote:
Wed Oct 24, 2018 4:22 pm
jalbert wrote:
Mon Oct 22, 2018 6:25 pm
I don’t know why this is even a contentious issue. DFA has excellent products if you want the types of products they offer. Semantically, you can define passive management however it suits your fancy. DFA funds do not achieve the same level of transparency as an index fund that tracks an externally controlled index.

DFA uses internal de facto indices by screening for stocks to include in their investable domain based on factor exposure, and holding them at market cap weight.
I just sat through a pretension by DFA today. They are not passive.

They are top down, discipline, quantitative and academically driven. They are consonantly fiddling around with their factors.
Thanks
Is “pretension” a typo or deliberate?
:)
Any other tidbits? Like which factor are they most jazzed about now?
Keep calm and Boglehead on. KCBO.

retiredjg
Posts: 34158
Joined: Thu Jan 10, 2008 12:56 pm

Re: DFA small value... acitve management?

Post by retiredjg » Wed Oct 24, 2018 4:57 pm

privatefarmer wrote:
Sat Oct 20, 2018 12:12 am
....does this not sound like active management?
Did you think it was an index fund?

There is nothing in the name or in (a quick look at) the prospectus that would indicate it is an index fund. And the turnover is WAY too high to be an index fund. I'd have to say that of course, it is actively managed.

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Wed Oct 24, 2018 7:13 pm

grok87 wrote:
Wed Oct 24, 2018 4:42 pm
Is “pretension” a typo or deliberate?
typo.

The 5 factor model is the way to go. Self serving in my opinion.

Ignore the dividend factor. It is the value & profitability factors that matter.

Momentum is real but can't be exploited. (I had a issue with their arguments)

Low Beta is not a factor.

Dead Man Walking
Posts: 717
Joined: Wed Nov 07, 2007 6:51 pm

Re: DFA small value... acitve management?

Post by Dead Man Walking » Wed Oct 24, 2018 8:14 pm

Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.

DMW

finite_difference
Posts: 1043
Joined: Thu Jul 09, 2015 7:00 pm

Re: DFA small value... acitve management?

Post by finite_difference » Wed Oct 24, 2018 9:11 pm

alex_686 wrote:
Wed Oct 24, 2018 7:13 pm
grok87 wrote:
Wed Oct 24, 2018 4:42 pm
Is “pretension” a typo or deliberate?
typo.

The 5 factor model is the way to go. Self serving in my opinion.

Ignore the dividend factor. It is the value & profitability factors that matter.

Momentum is real but can't be exploited. (I had a issue with their arguments)

Low Beta is not a factor.
Thanks for these tidbits.

There’s a small page on the wiki but it’s helpful: https://www.bogleheads.org/wiki/Fama_an ... ctor_model

The 5 factors for explaining equity returns are:

1. Equity risk (self-explanatory??)
2. Value risk (company book-to-market equity?)
3. Size risk (company size?)
4. Investment (?)
5. Profitability (?)

However, 2 of the references are broken:

3. The Five Dimensions of Risk, by Rick Ferri
5. Research, DFA

It would be nice to have a sentence summarizing each factor and to clean-up the links. Perhaps Fama and French’s new paper is not a working paper any more?

typical.investor
Posts: 343
Joined: Mon Jun 11, 2018 3:17 am

Re: DFA small value... acitve management?

Post by typical.investor » Wed Oct 24, 2018 10:24 pm

finite_difference wrote:
Wed Oct 24, 2018 9:11 pm
alex_686 wrote:
Wed Oct 24, 2018 7:13 pm
grok87 wrote:
Wed Oct 24, 2018 4:42 pm
Is “pretension” a typo or deliberate?
typo.

The 5 factor model is the way to go. Self serving in my opinion.

Ignore the dividend factor. It is the value & profitability factors that matter.

Momentum is real but can't be exploited. (I had a issue with their arguments)

Low Beta is not a factor.
Thanks for these tidbits.

There’s a small page on the wiki but it’s helpful: https://www.bogleheads.org/wiki/Fama_an ... ctor_model

The 5 factors for explaining equity returns are:

1. Equity risk (self-explanatory??)
2. Value risk (company book-to-market equity?)
3. Size risk (company size?)
4. Investment (?)
5. Profitability (?)

However, 2 of the references are broken:

3. The Five Dimensions of Risk, by Rick Ferri
5. Research, DFA

It would be nice to have a sentence summarizing each factor and to clean-up the links. Perhaps Fama and French’s new paper is not a working paper any more?
I think it’s been found that 2) value risk isn’t explained by book-to-market.

The problem is contributed capital. It boosts book but doesn’t predict earning the value premium. Retained earnings (which is also included in book) is a better predictor if whether a stock will return the value premium if there is one.

That’s why RAFI funds serve as value funds. They aren’t strictly book and so people say they aren’t valuey enough, but their measures do capture retained earnings and so work. So for example they include Apple where a traditional value fund or even DFA wouldn’t.
Last edited by typical.investor on Thu Oct 25, 2018 3:04 am, edited 1 time in total.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Thu Oct 25, 2018 2:42 am

Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.

DMW
It is correct that Fidelity managing and tracking their own indices for the so-called Fidelity zero funds reduces transparency.
Risk is not a guarantor of return.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Thu Oct 25, 2018 7:07 am

alex_686 wrote:
Wed Oct 24, 2018 7:13 pm
grok87 wrote:
Wed Oct 24, 2018 4:42 pm
Is “pretension” a typo or deliberate?
typo.

The 5 factor model is the way to go. Self serving in my opinion.

Ignore the dividend factor. It is the value & profitability factors that matter.

Momentum is real but can't be exploited. (I had a issue with their arguments)

Low Beta is not a factor.
Thanks, that’s very helpful.

The comments on momentum are interesting. That is consistent with my understanding of their approach. That they don’t actively use momentum as a factor- like they would never run a momentum fund. But I think based on the prospectus for the small value fund they may use it as a screen for when they buy or sell.

So for example if they have an overweight position in stock abc and want to reduce to index weight, they may hold off if the momentum signal looks bad, etc.

That being sAid there momentum screen strategy seems to have cost them.
Keep calm and Boglehead on. KCBO.

garlandwhizzer
Posts: 2016
Joined: Fri Aug 06, 2010 3:42 pm

Re: DFA small value... acitve management?

Post by garlandwhizzer » Thu Oct 25, 2018 1:29 pm

Robert T wrote:

Would just note that over this 15 period (2003-2017) the 'momentum premium' for small caps was negative (in contrast to all other 15 year periods since 1927). No knowing what the next 15 years will bring.
Interesting, 15 years of negative returns for the first time in 90 years. I thought that MOM was supposed to be the most consistent of factors. We know that value has long periods of underperformance. Value has also had a rough spell for about a decade. Apparently so does SC MOM now. Is this just a random event, soon to be reversed, or is it evidence that factor premiums are being arbitraged away by quants? Thoughts?

Garland Whizzer

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Thu Oct 25, 2018 1:53 pm

That they don’t actively use momentum as a factor- like they would never run a momentum fund. But I think based on the prospectus for the small value fund they may use it as a screen for when they buy or sell.
That is my understanding as well. Many deep value stocks are ones where the market has drastically repriced the risk of future revenue being delivered causing a big drop in stock price. A momentum screen would delay buying those until the negative momentum had dissipated.

Conversely, when a value stock delivers a robust return, it will appreciate in price potentially enough to no longer be a value stock. A momentum screen might hold the stock until the positive momentum dissipates instead of selling when it was no longer a value stock.

This would differ from a momentum fund by not buying growth stocks with positive momentum as one example. But it still reduces value exposure and increases exposure to positive momentum.

It is instructive to run a 4-factor regression on the ETF RZV which has deep value exposure without a momentum screen along with DFSVX which has a momentum screen:

https://www.portfoliovisualizer.com/fac ... sion=false

RZV is weighted by value exposure not cap-weighted so had more size factor exposure. I think DFA would contend that they did not give up too much value exposure from the negative momentum screen.
Risk is not a guarantor of return.

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Thu Oct 25, 2018 6:19 pm

Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.
So, what is your definition of a index? What DFA does not seem like indexing.

The standard characteristics are:
Arm's length
Low Subjectivity
Low Change in Methodology
Changes are announced beforehand

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Thu Oct 25, 2018 6:24 pm

jalbert wrote:
Thu Oct 25, 2018 1:53 pm
That is my understanding as well. Many deep value stocks are ones where the market has drastically repriced the risk of future revenue being delivered causing a big drop in stock price. A momentum screen would delay buying those until the negative momentum had dissipated.

Conversely, when a value stock delivers a robust return, it will appreciate in price potentially enough to no longer be a value stock. A momentum screen might hold the stock until the positive momentum dissipates instead of selling when it was no longer a value stock.
So, at the DFA presentation I was at - which was given by the research side, not the fund management side - they did not advocate small cap value. If you did that your portfolio would be under 50% of the small cap market space. Rather they were advocating small cap companies while excluding growth companies with low profitability. For those of you who are counting, that is 3 factors. Same returns as small cap value, but you are only excluding 25% of the small cap market space. So lower diversification risk.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Thu Oct 25, 2018 7:26 pm

Were they expressing recency bias or outright giving up on the idea that the size and value factors are persistent and pervasive across all markets?
Risk is not a guarantor of return.

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Thu Oct 25, 2018 7:40 pm

jalbert wrote:
Thu Oct 25, 2018 7:26 pm
Were they expressing recency bias or outright giving up on the idea that the size and value factors are persistent and pervasive across all markets?
First, when pressed by questioners if the factor premiums were stable or if they were losing power, the presenter really danced around the issue. Like a mantra he just repeated that the factors were persistent. Sigh, it made me a little unhappy. But this is more of a comment on the presenter.

Now, back to your question. Nothing of the kind. Let me see if I can make this clear.

Portfolio 1 = {Small, Value}, about 50% of small value space, about an extra 1.44 in extra return.

Portfolio 2 = {Small, not growth and not high profits}, about 75% of small value space, about an extra 1.4 in extra return.

Return premiums are from memory, so take them with a heavy grain of salt. Presenter just did not think giving up 25% of market space was worth the extra statically insignificant extra return.

They did exam the factors over time (before and after the 70s and markets (US, DM, EM). Found that the factors did give extra premium but that the premiums differed.

typical.investor
Posts: 343
Joined: Mon Jun 11, 2018 3:17 am

Re: DFA small value... acitve management?

Post by typical.investor » Thu Oct 25, 2018 10:51 pm

alex_686 wrote:
Thu Oct 25, 2018 6:24 pm
jalbert wrote:
Thu Oct 25, 2018 1:53 pm
That is my understanding as well. Many deep value stocks are ones where the market has drastically repriced the risk of future revenue being delivered causing a big drop in stock price. A momentum screen would delay buying those until the negative momentum had dissipated.

Conversely, when a value stock delivers a robust return, it will appreciate in price potentially enough to no longer be a value stock. A momentum screen might hold the stock until the positive momentum dissipates instead of selling when it was no longer a value stock.
So, at the DFA presentation I was at - which was given by the research side, not the fund management side - they did not advocate small cap value. If you did that your portfolio would be under 50% of the small cap market space. Rather they were advocating small cap companies while excluding growth companies with low profitability. For those of you who are counting, that is 3 factors. Same returns as small cap value, but you are only excluding 25% of the small cap market space. So lower diversification risk.
Is that what DFA funds do now?

It doesn’t really seem like it.

For instance, DFA large value doesn’t hold Apple or Microsoft, but blend funds do along with Amazon. Apple and Microsoft are the types of companies I’d want but not Amazon due to profitability.

In any case, I’m sympathetic to DFA’s research views. Schwab’s fundamental indexes have always behaved that way. I still consider them value and not really blend funds even they they include holdings outside the value half.

Dead Man Walking
Posts: 717
Joined: Wed Nov 07, 2007 6:51 pm

Re: DFA small value... acitve management?

Post by Dead Man Walking » Fri Oct 26, 2018 3:34 am

alex_686 wrote:
Thu Oct 25, 2018 6:19 pm
Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.
So, what is your definition of a index? What DFA does not seem like indexing.

The standard characteristics are:
Arm's length
Low Subjectivity
Low Change in Methodology
Changes are announced beforehand
Alex,

I invest in retail mutual funds that claim to track indices determined by indexers such as CRSP, FTSE, S&P, Russell, etc. I allow fund providers to determine which index they choose to track for their funds. As a consumer, my job is to determine which funds best track the index they have chosen to track. Vanguard is notorious for changing the indices they choose track. Their historical performance data often include an asterisk which explains that the performance history is based on a spliced index. What the hell is a spliced index?

I'm a dyi investor who doesn't have access to an advisor who has access to DFA funds; consequently, any discussion of DFA funds is purely academic. Since they screen a segment of the market for various factors, I consider their funds to be quasi active. Their patient trading has a passive nature, but isn't as disciplined as traditional index funds. I agree that DFA funds are not index funds. They are active funds with index fund characteristics.

The Fidelity zero funds are too new to evaluate. I agree with jalbert's comment about the transparency of the indices that Fidelity is developing for the funds.

Sorry that I can't define what an index is. I simply accept the indices provided by the major indexers. I choose an index fund that best suits my investment plan.

DMW

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Fri Oct 26, 2018 7:07 am

Dead Man Walking wrote:
Fri Oct 26, 2018 3:34 am
alex_686 wrote:
Thu Oct 25, 2018 6:19 pm
Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.
So, what is your definition of a index? What DFA does not seem like indexing.

The standard characteristics are:
Arm's length
Low Subjectivity
Low Change in Methodology
Changes are announced beforehand
Alex,

I invest in retail mutual funds that claim to track indices determined by indexers such as CRSP, FTSE, S&P, Russell, etc. I allow fund providers to determine which index they choose to track for their funds. As a consumer, my job is to determine which funds best track the index they have chosen to track. Vanguard is notorious for changing the indices they choose track. Their historical performance data often include an asterisk which explains that the performance history is based on a spliced index. What the hell is a spliced index?

I'm a dyi investor who doesn't have access to an advisor who has access to DFA funds; consequently, any discussion of DFA funds is purely academic. Since they screen a segment of the market for various factors, I consider their funds to be quasi active. Their patient trading has a passive nature, but isn't as disciplined as traditional index funds. I agree that DFA funds are not index funds. They are active funds with index fund characteristics.

The Fidelity zero funds are too new to evaluate. I agree with jalbert's comment about the transparency of the indices that Fidelity is developing for the funds.

Sorry that I can't define what an index is. I simply accept the indices provided by the major indexers. I choose an index fund that best suits my investment plan.

DMW
Keep calm and Boglehead on. KCBO.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Fri Oct 26, 2018 7:10 am

Dead Man Walking wrote:
Fri Oct 26, 2018 3:34 am
alex_686 wrote:
Thu Oct 25, 2018 6:19 pm
Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.
So, what is your definition of a index? What DFA does not seem like indexing.

The standard characteristics are:
Arm's length
Low Subjectivity
Low Change in Methodology
Changes are announced beforehand
Alex,

I invest in retail mutual funds that claim to track indices determined by indexers such as CRSP, FTSE, S&P, Russell, etc. I allow fund providers to determine which index they choose to track for their funds. As a consumer, my job is to determine which funds best track the index they have chosen to track. Vanguard is notorious for changing the indices they choose track. Their historical performance data often include an asterisk which explains that the performance history is based on a spliced index. What the hell is a spliced index?

I'm a dyi investor who doesn't have access to an advisor who has access to DFA funds; consequently, any discussion of DFA funds is purely academic. Since they screen a segment of the market for various factors, I consider their funds to be quasi active. Their patient trading has a passive nature, but isn't as disciplined as traditional index funds. I agree that DFA funds are not index funds. They are active funds with index fund characteristics.

The Fidelity zero funds are too new to evaluate. I agree with jalbert's comment about the transparency of the indices that Fidelity is developing for the funds.

Sorry that I can't define what an index is. I simply accept the indices provided by the major indexers. I choose an index fund that best suits my investment plan.

DMW
Splicing- think of splicing two ends of rope together.

“Active funds with index fund characteristics” what evidence do we have, besides dfa marketing material, that this is true?
Keep calm and Boglehead on. KCBO.

Dead Man Walking
Posts: 717
Joined: Wed Nov 07, 2007 6:51 pm

Re: DFA small value... acitve management?

Post by Dead Man Walking » Fri Oct 26, 2018 3:01 pm

grok87 wrote:
Fri Oct 26, 2018 7:10 am
Dead Man Walking wrote:
Fri Oct 26, 2018 3:34 am
alex_686 wrote:
Thu Oct 25, 2018 6:19 pm
Dead Man Walking wrote:
Wed Oct 24, 2018 8:14 pm
Although the original question is about DFA Small Cap Value, many responses have discussed various indices. As I recall, the new Fidelity zero funds will invest in indices developed by Fidelity. DFA seems to develop their own indices for various market segments. Fidelity seems to be following the same path with their new "total" market funds. Of course, the returns for all of these funds will be compared to the appropriate indices from all of the established indexers; however, they may not track any of them.
So, what is your definition of a index? What DFA does not seem like indexing.

The standard characteristics are:
Arm's length
Low Subjectivity
Low Change in Methodology
Changes are announced beforehand
Alex,

I invest in retail mutual funds that claim to track indices determined by indexers such as CRSP, FTSE, S&P, Russell, etc. I allow fund providers to determine which index they choose to track for their funds. As a consumer, my job is to determine which funds best track the index they have chosen to track. Vanguard is notorious for changing the indices they choose track. Their historical performance data often include an asterisk which explains that the performance history is based on a spliced index. What the hell is a spliced index?

I'm a dyi investor who doesn't have access to an advisor who has access to DFA funds; consequently, any discussion of DFA funds is purely academic. Since they screen a segment of the market for various factors, I consider their funds to be quasi active. Their patient trading has a passive nature, but isn't as disciplined as traditional index funds. I agree that DFA funds are not index funds. They are active funds with index fund characteristics.

The Fidelity zero funds are too new to evaluate. I agree with jalbert's comment about the transparency of the indices that Fidelity is developing for the funds.

Sorry that I can't define what an index is. I simply accept the indices provided by the major indexers. I choose an index fund that best suits my investment plan.

DMW
Splicing- think of splicing two ends of rope together.

“Active funds with index fund characteristics” what evidence do we have, besides dfa marketing material, that this is true?
When I think of splicing two ends of rope together, I think of a square knot. :happy

Other than DFA marketing materials, the only evidence that I have are posts here by Bogleheads who advocate investing in DFA funds. Larry Swedroe and Random Walker frequently explain the strategies used by Dimensional Fund Advisors. I'm not sure how closely their strategies adhere to the characteristics described by alex_686. Their strategies don't seem to be transparent from my point of view.

Investors who seek to invest in indices should be able to find an ETF that fits their investment plan. I have never understood the hullabaloo over DFA funds. When the advisor fee is deducted from the total return, comparable performance can surely be achieved by an appropriate ETF.

DMW

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Fri Oct 26, 2018 3:46 pm

It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Risk is not a guarantor of return.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Fri Oct 26, 2018 4:25 pm

jalbert wrote:
Fri Oct 26, 2018 3:46 pm
It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Thanks, that’s interesting. So rzv is the way to go then?
Keep calm and Boglehead on. KCBO.

alex_686
Posts: 3920
Joined: Mon Feb 09, 2015 2:39 pm

Re: DFA small value... acitve management?

Post by alex_686 » Fri Oct 26, 2018 4:42 pm

jalbert wrote:
Fri Oct 26, 2018 3:46 pm
It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Interesting - at the DFA presentation I was at, the research guy was dismissive of liquidity. On the other hand he felt that the joint factor of {Value, Profitability} had a higher premium than either {Value} or {Profitability} by themselves.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Fri Oct 26, 2018 6:11 pm

alex_686 wrote:
Fri Oct 26, 2018 4:42 pm
jalbert wrote:
Fri Oct 26, 2018 3:46 pm
It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Interesting - at the DFA presentation I was at, the research guy was dismissive of liquidity. On the other hand he felt that the joint factor of {Value, Profitability} had a higher premium than either {Value} or {Profitability} by themselves.
that's interesting
Keep calm and Boglehead on. KCBO.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Fri Oct 26, 2018 11:25 pm

grok87 wrote:
Fri Oct 26, 2018 4:25 pm
jalbert wrote:
Fri Oct 26, 2018 3:46 pm
It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Thanks, that’s interesting. So rzv is the way to go then?
RZV tracks the sp600 pure value index, which I believe lacks the investability screens of the regular sp600 indices. I was thinking more along the lines of IJS, VIOV, or SLYV. These load on size, value, and quality to varying extents (multi-factor funds before multi-factor funds became fashionable, but these do not try to exploit momentum).

I’m not the right person to ask for recommendations on heavily factor-loaded portfolios though. I do relatively mild tilts and am not sold on the free lunch stories based on diversification across high factor exposures. DFA clearly has changed their portfolio construction methodologies a fair bit over the years.

SLYV has the lowest ER of the three SP600 value ETFs but has distributed cap gains in the past, so I would only hold it in a tax-qualified account. I’ve used DFFVX (available in a 401K) and VSIAX, but think my preference moving forward would be VIOV in taxable space and SLYV in tax-qualified space.
Risk is not a guarantor of return.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Sat Oct 27, 2018 5:10 am

jalbert wrote:
Fri Oct 26, 2018 11:25 pm
grok87 wrote:
Fri Oct 26, 2018 4:25 pm
jalbert wrote:
Fri Oct 26, 2018 3:46 pm
It is interesting is that funds that track the S&P 600 Small-Cap Value Index are more aligned with the new focus on quality than DFA funds have been in the past. The liquidity and profitability screens of the S&P600 index family are a form of quality screen (but does not include the investment factor).
Thanks, that’s interesting. So rzv is the way to go then?
RZV tracks the sp600 pure value index, which I believe lacks the investability screens of the regular sp600 indices. I was thinking more along the lines of IJS, VIOV, or SLYV. These load on size, value, and quality to varying extents (multi-factor funds before multi-factor funds became fashionable, but these do not try to exploit momentum).

I’m not the right person to ask for recommendations on heavily factor-loaded portfolios though. I do relatively mild tilts and am not sold on the free lunch stories based on diversification across high factor exposures. DFA clearly has changed their portfolio construction methodologies a fair bit over the years.

SLYV has the lowest ER of the three SP600 value ETFs but has distributed cap gains in the past, so I would only hold it in a tax-qualified account. I’ve used DFFVX (available in a 401K) and VSIAX, but think my preference moving forward would be VIOV in taxable space and SLYV in tax-qualified space.
well i could be wrong here. but i thought rzv started with the s&p 600 (small caps) and then threw out everything that wasn't "pure value".
it holds like 160 stocks or about 27% of the 600. ijs by contrast holds about 460 stocks or 77% of the 600.
Keep calm and Boglehead on. KCBO.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Sat Oct 27, 2018 3:24 pm

Yeah, I think you are correct. RZV loads negatively on the investment factor but that is not part of the SP600 screen. It actually has a higher loading on profitability than IJS:

https://www.portfoliovisualizer.com/fac ... sion=false
Risk is not a guarantor of return.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Sat Oct 27, 2018 3:41 pm

jalbert wrote:
Sat Oct 27, 2018 3:24 pm
Yeah, I think you are correct. RZV loads negatively on the investment factor but that is not part of the SP600 screen. It actually has a higher loading on profitability than IJS:

https://www.portfoliovisualizer.com/fac ... sion=false
thanks. i may give it a look...
Keep calm and Boglehead on. KCBO.

typical.investor
Posts: 343
Joined: Mon Jun 11, 2018 3:17 am

Re: DFA small value... acitve management?

Post by typical.investor » Sun Oct 28, 2018 12:01 am

jalbert wrote:
Sat Oct 27, 2018 3:24 pm
Yeah, I think you are correct. RZV loads negatively on the investment factor but that is not part of the SP600 screen. It actually has a higher loading on profitability than IJS:

https://www.portfoliovisualizer.com/fac ... sion=false
OK, RZV has a higher loading on profitability. It also has a much more negative alpha. So what explains that?

Does the fund actually load more on value (which had a negative return over the period) than showing? Is it momentum? Or is it just an artifact of the way factors get measured that RZV is showing a strong negative alpha to offset the high loading on profitability which had a strong return over the period. Or is it's market loading not actually as high as being reported? Or is it a combination of those things?

Alpha is actually anything unexplained isn't it? I really don't have a good sense of what is explaining RZV returns by looking at the loadings.

I don't believe we can just omit the alpha from the picture and conclude anything from the loadings. The loadings show RZV should have out performed IJS don't they. That's what the Factor Performance Attribution in Basis Points shows me. The Monthly Factor Premiums in basis points for RZV is higher. But that didn't happen in actual returns.

Or am I misunderstanding this?

I just think factor loadings are too fuzzy to really be that useful.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Sun Oct 28, 2018 2:03 am

The loadings show RZV should have out performed IJS don't they?
No, it shows RZV has more exposure to certain risk factors. There is never a guarantee that a risk factor will deliver its risk premium, or that factors outside the regression are not explanatory over the period. If not included in the regression, they will show up in alpha.

RZV loads negatively on momentum. That could be the source of the negative alpha or it could be something else.

Factor regressions are also imprecise models. In some cases they explain return well. In other cases, they fall over. That’s one reason I’m not willing to use heavy factor tilts.
Risk is not a guarantor of return.

typical.investor
Posts: 343
Joined: Mon Jun 11, 2018 3:17 am

Re: DFA small value... acitve management?

Post by typical.investor » Sun Oct 28, 2018 4:27 am

jalbert wrote:
Sun Oct 28, 2018 2:03 am
The loadings show RZV should have out performed IJS don't they?
No, it shows RZV has more exposure to certain risk factors. There is never a guarantee that a risk factor will deliver its risk premium, or that factors outside the regression are not explanatory over the period. If not included in the regression, they will show up in alpha.
Well what I was looking at is the Factor Performance Attribution in Basis Points tab which shows what each premium returned over the period. It also shows how much in basis points the fund returned per it's exposure to that factor.

Are you saying that knowing a funds risk factor exposure (loading) and returns of a factor over that period (positive or negative), that we shouldn't expect a fund that loads on more on value in a period when value returned the highest premium to have the best returns? I don't think that's what you intend to say.

I'm asking what it is that the Performance Attribution tab purports to show. https://www.portfoliovisualizer.com/fac ... sion=false

Isn't it how much the fund earned in basis points from it's exposure to each risk factor?

jalbert wrote:
Sun Oct 28, 2018 2:03 am
RZV loads negatively on momentum. That could be the source of the negative alpha or it could be something else.
Sure, and it shows negative momentum when you include that factor. But that's FF 4 factor and now we have lost sight of profitability which you seemed to be comparing two different funds on.

My point was it's so murky that I'm not sure we can really compare funds so accurately, and certainly shouldn't ignore alpha when doing so.

Anyway, I don't agree that it's things not included in the regression which show up in alpha. Why should adding a factor leave less explained? Look at RZV:

Three factors -0.03% annual alpha
Four factors 0.73% annual alpha
Five factors -1.65% annual alpha

For the four factor model, MOM is pretty negative. I suspect that MOM isn't being measured accurately and it takes some positive alpha to offset it. I mean if size, value and market explain RZV pretty well, why would adding a MOM suggest there is an additional factor or factors returning a premium (as shown in the positive alpha). Why wouldn't that factor also be there when only measuring size, value and market?

With five factors the negative MOM is gone (as that's not in the model) and the market loading and size loading are stronger. So we need negative alpha to offset that because market and size returned a premium in the time period.

It's not that the four or five factor models have more unexplained factors that the three factor models do. It's that loadings are related to what factors are included in the model, and value is the catchall. Sure it might be an unexplained factor, or it may be there to adjust for when an included factor's loadings aren't entirely accurately measured.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Sun Oct 28, 2018 4:42 am

typical.investor wrote:
Sun Oct 28, 2018 4:27 am
jalbert wrote:
Sun Oct 28, 2018 2:03 am
The loadings show RZV should have out performed IJS don't they?
No, it shows RZV has more exposure to certain risk factors. There is never a guarantee that a risk factor will deliver its risk premium, or that factors outside the regression are not explanatory over the period. If not included in the regression, they will show up in alpha.
Well what I was looking at is the Factor Performance Attribution in Basis Points tab which shows what each premium returned over the period. It also shows how much in basis points the fund returned per it's exposure to that factor.

Are you saying that knowing a funds risk factor exposure (loading) and returns of a factor over that period (positive or negative), that we shouldn't expect a fund that loads on more on value in a period when value returned the highest premium to have the best returns? I don't think that's what you intend to say.

I'm asking what it is that the Performance Attribution tab purports to show. https://www.portfoliovisualizer.com/fac ... sion=false

Isn't it how much the fund earned in basis points from it's exposure to each risk factor?

jalbert wrote:
Sun Oct 28, 2018 2:03 am
RZV loads negatively on momentum. That could be the source of the negative alpha or it could be something else.
Sure, and it shows negative momentum when you include that factor. But that's FF 4 factor and now we have lost sight of profitability which you seemed to be comparing two different funds on.

My point was it's so murky that I'm not sure we can really compare funds so accurately, and certainly shouldn't ignore alpha when doing so.

Anyway, I don't agree that it's things not included in the regression which show up in alpha. Why should adding a factor leave less explained? Look at RZV:

Three factors -0.03% annual alpha
Four factors 0.73% annual alpha
Five factors -1.65% annual alpha

For the four factor model, MOM is pretty negative. I suspect that MOM isn't being measured accurately and it takes some positive alpha to offset it. I mean if size, value and market explain RZV pretty well, why would adding a MOM suggest there is an additional factor or factors returning a premium (as shown in the positive alpha). Why wouldn't that factor also be there when only measuring size, value and market?

With five factors the negative MOM is gone (as that's not in the model) and the market loading and size loading are stronger. So we need negative alpha to offset that because market and size returned a premium in the time period.

It's not that the four or five factor models have more unexplained factors that the three factor models do. It's that loadings are related to what factors are included in the model, and value is the catchall. Sure it might be an unexplained factor, or it may be there to adjust for when an included factor's loadings aren't entirely accurately measured.
I suspect that RZV does have a negative loading on momentum. If you look at the bad momentum years, 2009 and 2016 then RZV outperformed in those years:
2009 RZV = 67.1%, morningstar category = 31.3%
2016 RZV = 33.6%, mornngstar category = 26.0%
Keep calm and Boglehead on. KCBO.

garlandwhizzer
Posts: 2016
Joined: Fri Aug 06, 2010 3:42 pm

Re: DFA small value... acitve management?

Post by garlandwhizzer » Sun Oct 28, 2018 12:06 pm

RZV has been a disaster ever since it opened 13 yrs ago. Morningstar rates it 2 stars, suggesting it takes on the highest level of risk and produces only average returns. Actually average returns overstates RZV's real returns. Vanguard's Small Cap Blend ETF, has massively outperformed RZV since RZV's inception 13 years ago (123% versus 79%). RZV took on considerably more risk and produced considerably lower returns than a simple small cap blend index in spite of the fact that RZV maximizes exposure to the value (deep value) and small (micro-cap) factors. Factor apologists always have an explanation to offer for failure of factor funds (negative MOM, negative alpha, you're not waiting long enough for the outperformance to show up, etc.,). These explanations are always seen in hindsight. However when RZV started its factor analysis going forward projected a brilliant future of outperformance due to its micro-cap deep value focus. The exact opposite has happened consistently for 13 years. I believe a certain level of skepticism is appropriate when someone tells you they know what the future performance of real factor funds will be based on 3, 4, or 5 factor models. Models are academically derived attempts to describe reality, not reality itself, which is a far more elusive target. The important concept to keep in mind in investing for outperformance is "No Guarantees."

Garland Whizzer

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Sun Oct 28, 2018 2:56 pm

Sure, and it shows negative momentum when you include that factor. But that's FF 4 factor and now we have lost sight of profitability which you seemed to be comparing two different funds on.

My point was it's so murky that I'm not sure we can really compare funds so accurately, and certainly shouldn't ignore alpha when doing so.

Anyway, I don't agree that it's things not included in the regression which show up in alpha. Why should adding a factor leave less explained? Look at RZV:

Three factors -0.03% annual alpha
Four factors 0.73% annual alpha
Five factors -1.65% annual alpha
I only use the regressions as a tool to see what kind of exposures a fund has. I do not think factor regressions are very reliable. They certainly are not very predictive of future performance. I’m not trying to defend the idea of holding RZV, just noting that if exposure to size, value, and profitability are what you care about, RZV seems to have a high exposure to them. Its lifespan has been during a period of underperformance of value. Is there some reason to expect it to have lots of morning stars?
Risk is not a guarantor of return.

grok87
Posts: 8443
Joined: Tue Feb 27, 2007 9:00 pm

Re: DFA small value... acitve management?

Post by grok87 » Sun Oct 28, 2018 4:16 pm

jalbert wrote:
Sun Oct 28, 2018 2:56 pm
Sure, and it shows negative momentum when you include that factor. But that's FF 4 factor and now we have lost sight of profitability which you seemed to be comparing two different funds on.

My point was it's so murky that I'm not sure we can really compare funds so accurately, and certainly shouldn't ignore alpha when doing so.

Anyway, I don't agree that it's things not included in the regression which show up in alpha. Why should adding a factor leave less explained? Look at RZV:

Three factors -0.03% annual alpha
Four factors 0.73% annual alpha
Five factors -1.65% annual alpha
I only use the regressions as a tool to see what kind of exposures a fund has. I do not think factor regressions are very reliable. They certainly are not very predictive of future performance. I’m not trying to defend the idea of holding RZV, just noting that if exposure to size, value, and profitability are what you care about, RZV seems to have a high exposure to them. Its lifespan has been during a period of underperformance of value. Is there some reason to expect it to have lots of morning stars?
agree
Keep calm and Boglehead on. KCBO.

jalbert
Posts: 3812
Joined: Fri Apr 10, 2015 12:29 am

Re: DFA small value... acitve management?

Post by jalbert » Mon Oct 29, 2018 12:52 am

Anyway, I don't agree that it's things not included in the regression which show up in alpha. Why should adding a factor leave less explained? Look at RZV:

Three factors -0.03% annual alpha
Four factors 0.73% annual alpha
Five factors -1.65% annual alpha
Nothing inconsistent there. Adding momentum contributed to negative factor return, moving some return into alpha for the four factor model.

The five factor model drops momentum, but captures more positive return in quality factors, a net gain in factor return both from adding the positive contributing profitability and dropping the negative contributing momentum, which leaves a negative alpha to compensate.

It is counterintuitive because the limits in explanatory power of factor models are more significant than often articulated.
Risk is not a guarantor of return.

User avatar
vineviz
Posts: 2005
Joined: Tue May 15, 2018 1:55 pm

Re: DFA small value... acitve management?

Post by vineviz » Mon Oct 29, 2018 9:07 am

typical.investor wrote:
Sun Oct 28, 2018 4:27 am
Anyway, I don't agree that it's things not included in the regression which show up in alpha.
By definition in a linear regression, alpha represents the amount of non-random variance which cannot be explained by the independent variables. So it is certainly true that omitting an important independent variable which is a consistent source of variance (in this case, a risk factor) can result in all or some of that variance showing up in alpha.

In this case I think the "missing" factor is actually the fixed income "term" factor: adding that risk factor improves the adjusted r-squared notably and reduces alpha to an amount that is statistically insignificant.

https://www.portfoliovisualizer.com/fac ... ssion=true

This could have something to do with the perceived safety of RZV's holdings relative to IJS and/or the discounting of cash flows: RZV has a higher dividend yield. RZV also has a considerably higher weight in consumer cyclical companies with lower weights in financial services and utilities.

RZV dramatically outperformed IJS in 2009, for instance, and the nature of this fund may be that it exerts its premium (assuming their is one) in infrequent but intense bursts.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Post Reply