DFA Funds

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
stemikger
Posts: 4950
Joined: Thu Apr 08, 2010 5:02 am

DFA Funds

Post by stemikger »

I was afraid to ask this but my curiosity is killing me.

What are DFA funds and why do so many on this forum treat them with such high regard.

I just read a book on Bernie Madoff. It tells how when Madoff was thought to be hot, the only way you can get into his fund was to know someone and even then many got turned away.

DFA seems like that from just what I read when people ask about them.

Are they that good? I personally think Vanguard is awesome. What is the difference between them and Vanguard?

Can someone explain a little about this seemingly holy grail of an investment.

Thanks.
Mazz
Posts: 77
Joined: Tue Oct 19, 2010 12:51 pm

DFA Funds

Post by Mazz »

DFA is another fund that offers low cost funds. Similar to Vanguard.

Their funds are passive and embrace a value tilt based on academic research.

They seem to offer funds is sub asset classes that other companies, including Vanguard, don't offer. Some studies have shown that investors in a diversified DFA portfolio may be able to add between 1% and 3% per year over normal index funds.

DFA, which stands for Dimensional Fund Advisors, is a real legitimate investment firm. They are not an unregulated hedge fund. I think they have about $200 billion in assets under management.

The only drawback to DFA, which is why I think it is discussed so much in forums, is that DFA makes you go through a fee only advisor to access the funds.

The key, if you are interested in their funds, is to find a low cost advisor that will work with you.

I personal use both DFA funds and Vanguard funds in my portfolio.

If you want to learn more, then you may want to check out a few informative sites. Such as:

DFAus.com - the DFA site

IFA.com - a very good site with tons of information, it is an advisor's site.

FinancialPlanning.com - another advisor site with some good information and ideas.

Hope this helps.
User avatar
Topic Author
stemikger
Posts: 4950
Joined: Thu Apr 08, 2010 5:02 am

Post by stemikger »

Awesome, Thanks Mazz.

I know I could have searched on Google, but your response is the reason why I asked it on this forum. Clear and easy to understand.

I appreciate it!!

Steve.
rustymutt
Posts: 4001
Joined: Sat Mar 07, 2009 11:03 am

Post by rustymutt »

I take issue with DFA being compared to Vanguard in regards to cost. The funds I looked into at DFA have higher cost, and the adviser fees make Vanguard funds/ETFs look even better.
Even educators need education. And some can be hard headed to the point of needing time out.
User avatar
RJB
Posts: 321
Joined: Sun Sep 21, 2008 12:39 pm
Location: /earth

Post by RJB »

From what I've read about DFA Funds, they are low cost and well diversified. They have a tendency to buy and hold stocks. One advantage that they have over index funds is that DFA Managers can buy stocks and hold them. They can also buy MicroCap Stocks. They do not have to sell stocks like a Small Cap Index Fund would if a stock rises. They do not have to buy and sell stocks based on Indexes. I've also read that they do not use a broker, but directly buy and sell there own stocks. Thus it is less likely that others can find out what they are buying and selling for both reasons.

One disadvantage DFA Funds have is that they cost slightly more than index funds.
Also not everyone has access to DFA Funds.
EyeDee
Posts: 1388
Joined: Mon Feb 19, 2007 11:15 pm

Re: DFA

Post by EyeDee »

.
Here is Larry's post with URL's corrected so the post will show:
larryswedroe wrote:Here is a series of blog posts I did on DFA and how they compare to similar index funds that I think you will find helpful

http://moneywatch.bnet.com/investing/bl ... ntent;col1
http://moneywatch.bnet.com/investing/bl ... ntent;col1
http://moneywatch.bnet.com/investing/bl ... blog-river
http://moneywatch.bnet.com/investing/bl ... blog-river
Randy | SCA - Build Savings early by living below one's means, minimize Costs including taxes, and maintain a diverse Allocation.
User avatar
Rick Ferri
Posts: 9707
Joined: Mon Feb 26, 2007 10:40 am
Location: Georgetown, TX. Twitter: @Rick_Ferri
Contact:

Re: DFA Funds

Post by Rick Ferri »

Mazz wrote:
Some studies have shown that investors in a diversified DFA portfolio may be able to add between 1% and 3% per year over normal index funds.
There is absolutely no truth to this statement. My firm offers portfolios of DFA funds, portfolios of ETFs, portfolios of Vanguard funds, and mixed portfolios based on our selection. The portfolios are all structured with the same tilts to value and small cap, and they all perform about the same, i.e. the 'noise' is about 0.5% per year, one way or the other. It's all interchangable.

This myth of DFA outperforming other index products is perpetrated by some advisor (NOT DFA) who grossly misrepresent the risk factor loads to small cap and value in the non-DFA funds they are comparing. These advisors are being unethical in their methodologies by not fully disclosing this truth upfront. Their methods certainly would not pass CFA standards for performance measurement.

DFA is a good company and they have several good funds. But there is no free lunch on Wall Street.

Rick Ferri
MM07
Posts: 43
Joined: Tue May 18, 2010 2:59 pm

Re: DFA Funds

Post by MM07 »

Let me give a disclaimer and say that I have never owned DFA funds so I am not as familiar as some with them. I just see their marketing stuff.

1. If they do ad 1-3% (no free lunch and they don't) without inscreasing any type of risk that is quickly evaporated by advisor fees to a nominal amount.

2. They talk about how active management use unfair benchmarks, for example saying it is wrong to compare an active fund that is valuey to the S&P because based on their academic research you should "expect" to get a better return. However, they then compare their super valuey quant funds to a much less valuey index like the Russell.

3. I was on one of these DFA advisor sites and they were showing how you could the same return as warren buffett by simply investing everything in emergin market value. I won't even begin to comment on this, but it is akin to me telling you that you can have the same wealth as warren buffett, just rob a federal reserve bank, never mind the risk involved.

4. They rail against active management and how even the ethical ones can't beat the indexes for long (and I happen to agree with this). They then go on to show you that while active management (which you might by luck beat the market) is a loser's game, they can beat the market with what amount to quant. index funds.

5. There are several more but those stuck out at me.


Moral of the story is if you want to use an advisor and they put you in DFA funds there is nothing wrong with that.

However expecting some type of "magic return" is dreaming.
DSimmons
Posts: 18
Joined: Mon Oct 11, 2010 1:41 pm

Post by DSimmons »

Rick,

Do you have any kind of performance on your site regarding your portfolios?
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Need less beta exposure to get same tilts with DFA

Post by Random Walker »

It seems to me that it is not quite fair to say only that VG and DFA funds perform the same when you create VG portfolios with the same tilts. First of all, the DFA core and Tax Managed funds are likely more tax efficient than the VG tilted funds. Secondly, to get the same tilts to small and value with VG funds, you need to take on more beta exposure than you would have to with DFA funds.
Not saying that one is better than the other for everybody, but likely one is better than the other for each individual. And that decision depends on many factors, is very personal, and more complex than frequently portrayed.

Dave
cheapskate
Posts: 926
Joined: Thu Apr 26, 2007 1:05 pm

Re: DFA Funds

Post by cheapskate »

Rick Ferri wrote:
Mazz wrote:
Some studies have shown that investors in a diversified DFA portfolio may be able to add between 1% and 3% per year over normal index funds.
There is absolutely no truth to this statement. My firm offers portfolios of DFA funds, portfolios of ETFs, portfolios of Vanguard funds, and mixed portfolios based on our selection. The portfolios are all structured with the same tilts to value and small cap, and they all perform about the same, i.e. the 'noise' is about 0.5% per year, one way or the other. It's all interchangable.

This myth of DFA outperforming other index products is perpetrated by some advisor (NOT DFA) who grossly misrepresent the risk factor loads to small cap and value in the non-DFA funds they are comparing. These advisors are being unethical in their methodologies by not fully disclosing this truth upfront. Their methods certainly would not pass CFA standards for performance measurement.

DFA is a good company and they have several good funds. But there is no free lunch on Wall Street.

Rick Ferri
Here is what Robert T wrote :

"My simple estimate a few years ago gave DFA a slight edge over alternatives in the 0.2 to 0.3 percent per year range (considering style consistency, block trading, securities lending, reconstitution arbitrage, tax efficiency, expense ratios) – not the 2 to 3% annual DFA outperformance (above factor exposure) sometimes claimed. Don’t think I was too far off – perhaps would be less generous in my estimates today. Just my take. Time will tell. Still think DFA is a first rate fund company and use their funds for college investments (WV529) but not for retirement investments".

In this thread

http://www.bogleheads.org/forum/viewtopic.php?p=715283

I know Robert T posted much more detailed DFA vs Vanguard (and other non-DFA option) analyses, adjusting for factor loadings. I am not able to find it on quick searches.
heyyou
Posts: 4461
Joined: Tue Feb 20, 2007 3:58 pm

DFA and VG

Post by heyyou »

DFA is a fund company that does not need the huge customer service staff that VG provides. DFA fund retail investors have to pay separately to those advisors for those services.

DFA limits turnover by encouraging their advisors to not let their investors hop in and out of DFA funds. Only in the past few years, has VG worked on that problem. VG is getting better in some areas where there was plenty of room for improvements.

DFA tries harder than VG to keep value in their value funds and to stay small in their small funds. Before VG changed their indices, some of their value funds were just the bottom half of an index, so when large caps boomed, the bottom half still had high, just not sky high P/E ratios. VG Sm or SmV fund was half midcaps for a while during the Tech boom. At the time that you needed it most, it was hard to maintain tilts to Small and Value. Again, VG is changing for the better.

Evanson Asset Management (another advisor with DFA access) has some well written articles discussing much of what has been mentioned earlier in this thread (not just this posting) without specifically naming VG. Their minimum investment was down to $500K when I last looked there, with a flat $2000 annual fee. I don’t know how their equity portfolios are structured now since the introduction of the DFA Core and Vector funds. Prior to that, equity was a 10 x 10 slice and dice including one slice that was half EM Small and half EM value. The bond portion was whatever amount suited you, separate from the equity S&D.

My opinion is if you need handholding, do hire a fee based advisor. If you want DFA funds, pay an advisor for access. If you don't need either, then do it yourself at VG. Do what suits you, but that does not mean your way is the best way for everyone else.

If you want to see a stinking pile, look at most Fidelity fund costs compared to VG funds. Their advisor shares (higher priced than investor shares!) include an annual .25% fee kicked back to the advisor. Without VG's competition, there would not be any low cost Spartan funds at Fido. That's the fund company to criticize on costs and advertising hype. Not trying to derail the topic, but Fidelity is greedy where as DFA is limiting access to make its funds easier to manage by avoiding large, fast turnover. Two vastly different cultures with VG between them.
User avatar
Random Musings
Posts: 6770
Joined: Thu Feb 22, 2007 3:24 pm
Location: Pennsylvania

Post by Random Musings »

I think some of the advantages where DFA can shine would be to utilize their SCV funds (only) with nominal bonds/TIPs. I think there have been prior backtests where this type of equity loading with bonds has some advantages. Currently, I don't think many offer EM SCV passive besides DFA.

Also, a good advisor will also help one stay the course - which may be helpful for many (most) who really can't stay the course even though they say they might.

RM
User avatar
Rick Ferri
Posts: 9707
Joined: Mon Feb 26, 2007 10:40 am
Location: Georgetown, TX. Twitter: @Rick_Ferri
Contact:

Post by Rick Ferri »

DSimmons wrote:Rick,

Do you have any kind of performance on your site regarding your portfolios?
No. As a CFA charterholder, I am under much stricter standards of performance disclousure than a vast number of advisors who have no standards. It would cost my firm several tens of thousands of dollars to have actual, audited, third party evaluated performance numbers posted to our website. We have not made that huge financial leap.

We could use make-beleive performance i.e. simulated, using the same non-sense disclaimers that many of our competitors use, but we have not stooped to that gutter level, at least not yet. Sad to say this, but the continued unethical behavior of my competitors may force us to post non-sense make-beleive (simulated) performance in the future just to level the playing field.

Rick Ferri
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

It's just as easy as this: DFA offers more efficient and targeted exposure to the desired asset class than a traditional retail index fund or index ETF in non large-growth asset classes (DFA US Large, Int'l Large, and Emerging Markets are basically interchangable with Vanguard/iShares alternatives). That means higher historical and future expected returns. More sensible trading and other value added services (securities lending) moves the dial even futher, but it is not necessary to go into that level of detail to see the advantages. In most cases, they have the added benefit of doing so with much greater security diversification.

That quickly takes non-DFA investors to the assumption that this entails greater risk (advisors make this same mistake). But large/small and value/growth have very low correlations with the market, so in most cases a more diversified multi asset class portfolio does not have the same proportion of increased volatility to match the additional historical and expected future returns (that's basic Modern Portfolio Theory).

Advisors like myself are constantly on the lookout for better (non-DFA)strategies, but the bullpen is empty at this time.

Finally, don't hide behind factor loads, they only tell you the portion of the historical returns attributable to the various market dimensions. They are noisy and imprecise when used for asset allocation decisions. An investor's goal should simply be to choose the asset allocation (ie. stocks/bonds, value/growth, small/large on a global basis) that makes the most sense for them, then choose the lowest cost, most diversified index available to them in each asset class.

Taking on more "factor exposure" through additional potfolio allocations to retail large value, small, and small value indexes (beyond what your asset class allocation calls for) has the troubling side effect of reducing the overall "market" allocation (ie. S&P 500, Russell 1000, Wilshire 5000, DFA Core 1, or DFA Core 2) and drastically increasing portfolio tracking error. And it doesn't improve your relative returns, because if you are comfortable with a given asset class asset allocation with a retail index fund approach, you will be just as comfortable with that asset class allocation using DFAs funds. (for example, if I'm willing to own "mid cap value" with Russell/Vanguard, I am just as fine holding 50% DFA large value and 50% DFA small value to get that middling effect).

So here are the historical DFA index advantages over the associated Russell indexes (which are the best we have for sufficiently long periods of time):

1) DFA Large Value index has outpaced the Russell 1000 Value by 1.2% per year since 1979.
2) 50/50 DFA large value/small value has outpaced the Russell Mid Value index by 0.7% per year since 1986.
3) DFA targeted value index has outpaced the Russell 2500 value index by 2.2% per year since 1986.
4) DFA Small Value index has outpaced the Russell 2000 Value index by 2.7% per year since 1979.
5) DFA Int'l Value index has outpaced the MSCI EAFE Value Index by 3.1% since 1994.
6) DFA Int'l Small Value index has outpaced the MSCI EAFE Small Value index by 2.4% per year since 1994.

(in each case, I went back as far as the oldest retail index was available)

There's your 1% to 3%. Pretty much makes the insight into how to properly structure a multi asset class portfolio, and guidance to stay with it through thick and thin (which is the biggest problem DIY investors have and why DFA shy's away from retail access) a free lunch.
Last edited by Multifactor Advisor on Wed Dec 15, 2010 5:41 pm, edited 2 times in total.
afan
Posts: 8191
Joined: Sun Jul 25, 2010 4:01 pm

Post by afan »

Multifactor,

That was a very informative post. Are those nominal returns, or risk adjusted? If the latter, are they single, 3 or 4 factor adjusted?

I assume these figures are pretax and before the advisor fees? Do you have comparable figures for returns in a taxable account?

For an individual investor, without a very large portfolio, say, under $10M, it seems hardly worth it to have to deal with an advisor.

Thanks
User avatar
Rick Ferri
Posts: 9707
Joined: Mon Feb 26, 2007 10:40 am
Location: Georgetown, TX. Twitter: @Rick_Ferri
Contact:

Post by Rick Ferri »

It is absurd to compare the simulated DFA fund performance to indexes that are not trying to capture deep value exposure. It's just absurd.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

Rick Ferri wrote:It is absurd to compare the simulated DFA fund performance to indexes that are not trying to capture deep value exposure. It's just absurd.
Deep value, relative value, GARP, whatever Wall Street lingo you want to apply, it all comes down to the asset class return. Either you are capturing all of it, or less than all of it. There's no way around it no matter what labels you want to use.
mikep
Posts: 3856
Joined: Wed Apr 22, 2009 9:27 pm

Post by mikep »

Aren't russell indexes subject to the most arbitrage due to their predictive annual rebalancing?
Advisors LOVE to use Russell as their comparison benchmark.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

Multifactor Advisor wrote:
Rick Ferri wrote:It is absurd to compare the simulated DFA fund performance to indexes that are not trying to capture deep value exposure. It's just absurd.
Deep value, relative value, GARP, whatever Wall Street lingo you want to apply, it all comes down to the asset class return. Either you are capturing all of it, or less than all of it. There's no way around it no matter what labels you want to use.

Also, just to clarify, my #s above are index to index, so they are perfectly comparable.
User avatar
Blue
Posts: 1178
Joined: Sat Jul 12, 2008 10:18 pm

Post by Blue »

Who owns DFA?



Who owns Vanguard?



We hold our investments with the organization whose interests are most closely aligned with our own.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

mikep wrote:Aren't russell indexes subject to the most arbitrage due to their predictive annual rebalancing?
Advisors LOVE to use Russell as their comparison benchmark.
It's better than S&P or Vanguard.

10YR Returns:

Large Value

Code: Select all

Russell 1000 Value = +2.8%
Vanguard Value Index = +1.7%
S&P 500 Value = +1.2%
Small Value

Code: Select all

Russell 2000 Value = +8.5%
Vanguard Small Value Index = +8.6%
S&P 600 Value = +8.3%
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

Blue wrote:Who owns DFA?



Who owns Vanguard?



We hold our investments with the organization whose interests are most closely aligned with our own.
As an owner, maybe you can convince them to focus more on completing a lineup of global large/small and growth/value indexes instead of continuing to add to their sector fund lineup, as well as a myriad of active strategies?

PS--CRSP is developing a new suite of indexes that may help to reduce the reconstitution issues associated with retail indexes, while also offering more sensible and targeted value/growth exposure that Vanguard may eventually adopt.

Whose wealth do you think is driving the CRSP work (at the University of Chicago)? DFA CEO David Booth.

In an ironic twist of fate, David Booth (and indirectly DFA) may turn out to do more good for Vanguard than their own CEOs ever did!
User avatar
Blue
Posts: 1178
Joined: Sat Jul 12, 2008 10:18 pm

Post by Blue »

Multifactor Advisor wrote:
Blue wrote:Who owns DFA?



Who owns Vanguard?



We hold our investments with the organization whose interests are most closely aligned with our own.
As an owner, maybe you can convince them to focus more on completing a lineup of global large/small and growth/value indexes instead of continuing to add to their sector fund lineup, as well as a myriad of active strategies?

PS--CRSP is developing a new suite of indexes that may help to reduce the reconstitution issues associated with retail indexes, while also offering more sensible and targeted value/growth exposure that Vanguard may eventually adopt.

Whose wealth do you think is driving the CRSP work (at the University of Chicago)? DFA CEO David Booth.

In an ironic twist of fate, David Booth (and indirectly DFA) may turn out to do more good for Vanguard than their own CEOs ever did!

I think it is great that a billionaire like Booth donates $300MM to his alma mater.

Still, no matter how philanthropic he is, I prefer my dollars to stay my dollars and to have my investments with an organization more fully aligned with my personal interests.
cheapskate
Posts: 926
Joined: Thu Apr 26, 2007 1:05 pm

Post by cheapskate »

Multifactor Advisor wrote:
Multifactor Advisor wrote:
Rick Ferri wrote:It is absurd to compare the simulated DFA fund performance to indexes that are not trying to capture deep value exposure. It's just absurd.
Deep value, relative value, GARP, whatever Wall Street lingo you want to apply, it all comes down to the asset class return. Either you are capturing all of it, or less than all of it. There's no way around it no matter what labels you want to use.

Also, just to clarify, my #s above are index to index, so they are perfectly comparable.
MF Advisor

You compare Russell and DFA. Can you post the S/V loadings for the offerings you compare ? Are the value loadings for the Russell 1000 and the DFA LV identical ? What about the S/V loadings for your other compares ?

Comparisons like these ignoring factor loadings make no sense. Outside of factor exposure, the only reason DFA would add value is because of securities lending, "patient trading" (aka market timing) and similar factors, each of which entails extra potential risk.

This is a bit like comparing simulated DFA US Vector returns with the S&P500 over the last 10 years (right after the bursting of the LG/TMT bubble) and claiming that DFA Vector is far superior to anything else out there.

Anytime there is a DFA thread out there, DFA advisors come out of the woodwork in full force hammering the table about how wonderful DFA is, how there is really no alternative to DFA etc. Your post is very similar to the aggressive marketing emanating from a prominent DFA shop in Southern California (even their acronym rhymes with DFA).

Interested readers can search for and read Robert T's many fantastic posts that compare DFA with non-DFA offerings and draw their own conclusions. And hear Rick's unbiased views on this issue. Contrary to what DFA advisors would like to claim, the options for DIY investors to tilt towards SV is expanding every month as ETF providers bring new products to market.

I've said this before. A better strategy for DFA advisors is to focus the marketing pitch on the many benefits advisors bring to the table (DFA access isn't a measurable benefit).
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T »

.
  • "Over the long haul, what matters is factor exposure and expense" - Bill Bernstein

    Reminds me of the following:

    “The truth is incontrovertible, malice may attack it, ignorance may deride it, but in the end; there it is.” - Winston Churchill

    Its factor exposure not this:

    Image
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

DFA more efficient for same factor exposure?

Post by Random Walker »

So if you can get the same exposure to small and value with less exposure to beta is that a better portfolio? Seems like it to me. Can that improved efficiency be quantified or approximated? Seems to me it is not just about factor exposure, but factor exposure in an efficient manner. Seems that VG wins on costs. Also seems that DFA wins on portfolio efficiency, tax efficiency, and efficiency of fund structure with core funds. A DFA portfolio will have less trading and rebalancing costs I believe. Everyone's thoughts appreciated.

Dave
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

dave

Post by larryswedroe »

Basically that is it. The higher the loading factors the less beta you need. That means you cut the fat tail risk. Also you cut total costs because bond investing cheaper than equity investing, especially if you build bond portfolios individually, avoiding manager fees. And then if DFA has added value by security lending and block trading and screens you end up with more benefits than just the tilting, and those are fee lunches--though they induce (random) tracking error.

Hope that is helpful
james22
Posts: 1992
Joined: Tue Aug 21, 2007 2:22 pm

Re: Need less beta exposure to get same tilts with DFA

Post by james22 »

Random Walker wrote:...to get the same tilts to small and value with VG funds, you need to take on more beta exposure than you would have to with DFA funds.
This the most meaningful difference to me.

Edit: I see the last two pasts just made this point.
User avatar
Munir
Posts: 3200
Joined: Mon Feb 26, 2007 3:39 pm
Location: Oregon

Re: DFA

Post by Munir »

EyeDee wrote:.
Here is Larry's post with URL's corrected so the post will show:
larryswedroe wrote:Here is a series of blog posts I did on DFA and how they compare to similar index funds that I think you will find helpful

http://moneywatch.bnet.com/investing/bl ... ntent;col1
http://moneywatch.bnet.com/investing/bl ... ntent;col1
http://moneywatch.bnet.com/investing/bl ... blog-river
http://moneywatch.bnet.com/investing/bl ... blog-river
I found Larry's articles quite helpful.

My question is whether being a retiree in a distribution phase is a significant factor to tilt one to Vanguard or DFA? Or is it not a relevant factor?
psteinx
Posts: 5801
Joined: Tue Mar 13, 2007 2:24 pm

Post by psteinx »

Multifactor Advisor wrote:So here are the historical DFA index advantages over the associated Russell indexes (which are the best we have for sufficiently long periods of time):

1) DFA Large Value index has outpaced the Russell 1000 Value by 1.2% per year since 1979.
DFA has not been around since 1979*. I think it should be very clear to readers of this forum that you are quoting backtested figures using an index that was derived AFTER the start of your time frame.

*Wikipedia link: http://en.wikipedia.org/wiki/Dimensional_Fund_Advisors
Mazz
Posts: 77
Joined: Tue Oct 19, 2010 12:51 pm

Post by Mazz »

Multifactor,

Thanks for showing those returns.

I looked at the performance of their funds relative to the benchmarks for 10 year periods and found similar results that you showed.

I personally don't care if a DFA fund is 15 basis points (0.15%) higher than Vanguard or an ETF, as long as the performance is better.
Even if that performance is only 1%.

A significant concern of mine was tax efficiency. I found the DFA funds to be more tax efficient than any of the other fund companies, including Vanguard.

I think part of that benefit is a result of the advisor gatekeeper business model. I do agree that paying a big premium (.75% or more) for access may be counter productive. (Although may still be worth it if the performance is 1-3% higher)

I am only paying between 1 and 2 basis points to access the DFA funds, so I view the access fee to be insignificant.

I just recently switched over to DFA, so it is premature to say how much better I have done. Hopefully I won't be disappointed. If so, I can always switch back to ETFs or Vanguard funds.
afan
Posts: 8191
Joined: Sun Jul 25, 2010 4:01 pm

Post by afan »

Is there any room in the DFA system for someone like me who knows that he could never give trading authority to anyone? I have tried, and I just cannot do it. I could consider having an advisor who did exactly that, provide advice, for a fee. I could not tell the advisor to make the actual rebalancing transactions and let me know about it. I count holding an index fund as different, since the investment advisor, the people doing the trading, and the custodian are three separate groups, and the advisor has very little discretion of what to do. That discretion would never extend to actually getting their hands on the funds.

In the case that NY Times columnist Ron Lieber avoided being a fraud victim when his financial advisor turned rogue, the scam was a person who had legitimate access to trade then forged authorization to make distributions from the accounts. I found myself wondering "why would anyone let the advisor actually make the trades? The clients could do that themselves. In fact, why not send the advisor a list of securities held, whether in taxable or tax favored accounts, along with cost basis, but no account number, or (not relevant for DFA) even who holds the assets?" Under such a model the advisor could offer investment suggestions, which the client could follow, or not, but if there were a crook at the advisory firm, they would not even know where to turn to attempt to get their hands on the money.

Anyone know whether any DFA-accepted advisors would function in the manner described above? Truly advice-only, with no ability to do anything to the assets?
caklim00
Posts: 2420
Joined: Mon May 26, 2008 10:09 am

Post by caklim00 »

Mazz wrote:I am only paying between 1 and 2 basis points to access the DFA funds, so I view the access fee to be insignificant.
:?: 1 - 2 bp? How is this even possible unless you have like $10M+? WVa 529 which is around 35 bp is the lowest I've even seen for someone with less than $1M.
InertiaMan
Posts: 267
Joined: Tue Oct 16, 2007 10:26 am

Post by InertiaMan »

Perhaps we should step back a bit from the perpetual debate over whether DFA funds are good, great or the best, and directly address a key piece of Stemikger's original post: the comparison to Madoff.

I sincerely hope that every person on Bogleheads can agree that there is absolutely no validity in comparing DFA to Madoff.

We can debate the merits of factor exposure and 20bp differences in ER between DFA, Vaguard, etc. But DFA is no doubt a legitimate, professionally run organization with a fully transparent strategy. Your risks with DFA are exactly the risks of the asset classes you hold with them. The risk of DFA being a Ponzi scheme is as close to zero as you get.

Madoff, on the other hand, had a 100% opaque strategy (which turned out to be no strategy) run by a closely held, tiny group of people.

You're comparing a sociopath (Madhoff) to DFA, arguably one of the "safest" investment companies (in terms of their management of funds in the interest of their clients, not the risk exposure of the asset classes you choose). Sure there is one "similarity" noted that Madoff was popular with some audience at some point in time, and DFA is popular with another audience. Sure, both those audiences are investors. But I suspect there isn't much overlap between the audiences.

In terms of acting in their clients interest, there is a big spectrum between blatant criminals like Madoff on one end, and investor-owned Vanguard on the other. DFA should be placed very close to Vanguard on that spectrum, and merits NO comparison to Madoff.
InertiaMan
Posts: 267
Joined: Tue Oct 16, 2007 10:26 am

Post by InertiaMan »

caklim00 wrote:
Mazz wrote:I am only paying between 1 and 2 basis points to access the DFA funds, so I view the access fee to be insignificant.
:?: 1 - 2 bp? How is this even possible unless you have like $10M+? WVa 529 which is around 35 bp is the lowest I've even seen for someone with less than $1M.
You could pay $2000 to Evanson on a $5M account and reach 2bp. But I'm guessing there is a zero missing in Mazz' comment.

Mazz, another statement you made that raises concern is that you emphasize tax efficiency, but then you talk about moving back and forth between funds as if it is no big deal. Tax efficiency differences among good index and DFA funds are relatively insignificant in comparison to the potentially huge tax consequences of fickle movements of your assets in and out of fund choices.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

cheapskate: I'm gonna disagree with you on substance. I'll leave the DFA fund/advisor hatred alone.

Of course DFA funds register higher small/value factor loads than many retail indexes. "Loadings" just measure how full a dose of the asset class return a particular strategy captures. And we know that DFA does it better than retail indexes. All things being equal, a higher loading is preferable to a lower loading, because it implies higher expected returns within the asset class.

Does that mean I think factor exposure is useless? No. Factor exposure does help you differentiate between two similar index funds in the same asset class (the higher the loading the better). It can also helps you approximate expected asset class and portfolio returns based on your assumption of future stock/bond, small/large, and value/growth risk/return premiums.

But saying that all portfolios with the same factor loadings are apples to apples is taking factor exposure too far. Simply put: if you decide to develop your asset allocation based on the goal of a particular portfolio factor load, you've decided you are comfortable with X% in the market or some large growth-like portfolio, with X-1% in non market asset classes (large value, mid value, small value, small market, micro market, etc.). Two portfolio's are only comparable to the extent they have approximately the same overall asset allocation structure.

Somehow, the benefit of additional "factor loads" by using DFAs small and value strategies has magically turned into a "handicap". The conversations sound like this:

"my retail portfolio can only get 0.X value exposure based on the less than optimal structure of my retail value index funds that I use for non-market exposure. But you, as an investor with DFA access, if you want to compare yourself to me, you must reduce the use of your higher powered small cap/value funds vs. the market, or it's not fair! And if you don't comply with my request, well that is dishonest."

Or, even better: "I've developed my retail index portfolio, and it beats a similar DFA blend. How similar are the two? We'll, the DFA blend does have a noticeable bit less in small and value funds, and more in the market, but that's only fair! DFAs small and value funds capture more of the asset class return, so using them in the same percentages would change the outcome (and honestly, my goal all along was to show that I can beat a DFA fund mix!)"

But in reality, both of us should decide on a certain (lets say identical) allocation to market asset classes and non market asset classes(ie. large value, small, and small value). Then pick the non-market asset classes that most fully capture the desired asset class return. At the end of the day, those funds that most fully deliever the asset class return (most "valuey" value funds and "smallest" small funds) will allow that portfolio to come out ahead.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

 

 
Last edited by Multifactor Advisor on Thu Dec 16, 2010 11:30 pm, edited 1 time in total.
mountainmoney
Posts: 112
Joined: Sat Aug 16, 2008 2:12 pm
Location: Colorado

Post by mountainmoney »

The return difference between a perfectly executed DFA strategy and a perfectly executed non-DFA (but otherwise Boglehead) strategy is going to show up entirely to the right side of the decimal point.

Even assuming an investor could execute an investment plan reasonably well (which may be good enough) the reason to pay an advisor if a consumer of financial advice so chooses is to get your asset allocation right in the first place and to have a muse to help you stick to it - in both frothy and in depressing times. And to help you coordinate and integrate your investment plan with your tax plan, education plan, estate plan, philanthropic plan and over all life plan. Mistakes in any one of these areas can show up on the left side of the decimal point. As a percentage of AUM is just a convenient way pay for all of this.
"It is better to be vaguely right than precisely wrong." JM Keynes
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T »

Multifactor Advisor wrote:With all this emphasis on regression factors while completely ignoring the obvious reality of actual returns, I'm reminded of another Churchill quote:

“A lie gets halfway around the world before the truth has a chance to get its pants on.”
When I looked at this earlier this week - actual returns over the last (almost) 8 years (2003-2010 YTD) for portfolios with the same estimated factor loads (as reflected in this earlier thread http://www.bogleheads.org/forum/viewtop ... collective):

....................................................Annualized Returns...........Standard Deviation
Actual (non-DFA) Portfolio.......................11.3..............................19.3
DFA TM Portfolio......................................11.3..............................19.0

Robert

FWIW - I use non-DFA funds in retirement portfolio and DFA funds (WV529) in college education portfolio to target desired factor exposure (at lowest cost).
.
boglety
Posts: 60
Joined: Sun Jun 15, 2008 1:44 am

Post by boglety »

If I may ask how do DFA do after tax compared to ETFs? I have been discussing with an adviser about my situation.

No trading fee with ETF (100 free trades a year at WellsTrade)
No capital gains tax with ETFs

DFA has a transaction fee every time I purchase the mutual fund and capital gains despite the fact I have only bought and do not sell shares. 30/40 years down the road it is difficult for DFA to beat the adviser fee, transaction costs and taxes.

I am hoping if I switch to DFA that the small valuely tilt will increase my returns but there is no free lunch...

Any advice or thoughts?
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

Robert T wrote:When I looked at this earlier this week - actual returns over the last (almost) 8 years (2003-2010 YTD) for portfolios with the same estimated factor loads (as reflected in this earlier thread http://www.bogleheads.org/forum/viewtop ... collective):

....................................................Annualized Returns...........Standard Deviation
Actual (non-DFA) Portfolio.......................11.3..............................19.3
DFA TM Portfolio......................................11.3..............................19.0

Robert

FWIW - I use non-DFA funds in retirement portfolio and DFA funds (WV529) in college education portfolio to target desired factor exposure (at lowest cost).
.
Again, the emphasis on factor loads produces very different actual asset allocations. The ETF portfolio holds NO large US stocks, while 17% of the entire DFA allocation is large value (DFA TM Marketwide Value). The Int'l side of the ETF portfolio has 0% large growth and 5% in Int'l small, yet the DFA blend has 17% in Int'l large growth (DFA Large Cap Int'l) and 0% in Int'l small. Quite surprising, actually, that the DFA mix is able to match an ETF portfolio which a much heavier weighting to small cap and value asset classes.

Now, what if we actually compare apples to apples asset classes? Dividing the retail US Mid Value fund between DFAs Large and Small Value funds (TM Marketwide Value and TM Targeted Value), matching Bridgeway Ultra Small with DFA Micro, and then match the EAFE Value and EAFE Small with DFA TM Int'l Value and Int'l Small? This adds +0.7% per year over the proposed DFA blend in the above link. Far less than I'd expect, until I realized that US mid value stocks,which were the single largest asset class in the ETF mix, had unexpectedly high returns over this stretch, exceeding both large and small value for the period (and overcorrecting for the previous 8 years where the Russell Mid Value index return of +12.4% underperformed both the DFA Marketwide (LV) index return of +13.9 and the DFA Targeted Value (SV) index return of +14.5%).

But wait. Can we apply a bit of asset class common sense to the DFA portfolio, given we have a full suite of strategies at our disposal? I think we can.

1) Lets recognize that we already own all the necessary US micro cap stocks we want in our US small value fund (TM targeted value) and consolidate.
2) Lets also recognize, per Rex Sinquefield's timeless study on international diversification, that int'l small value stocks (not "small blend") offer the highest expected returns and diversification and make the switch from Int'l Small to Int'l Small Value.
3) Finally, value and size are also beneficial diversifiers if one is going to invest in emerging markets. So lets not waste our emerging markets allocation on the lowest expected returning large growth stocks (via DFA Emerging Markets). Instead, lets use the marketwide value fund (DFA Emerging Markets Value).

Making just these 3 simple and common sense changes, we've again moved the dial forward and picked up an additional 0.6% per year. That brings the total advantage to 1.3% per year over the ETF portfolio listed in the link above. Less than I'd expect, but I wouldn't ignore it. Just how different are the propsed DFA blend in the link above and my common sense alternative? I'll leave the regressions to someone with more idle time than I, and mention only that the correlation of their annual returns since 12/02 were +0.998--so in effect, they were identical.

As you see, what matters most is sensible asset allocation structure within the confines of a general risk tolerance while using the best investment strategies available to you. Use asset allocation and common sense as a compass, and regression analysis around the edges.
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T »

Multifactor Advisor wrote:That brings the total advantage to 1.3% per year
The way you have added – what you call a “total advantage” – of 1.3% is by changing the factor exposure of the portfolio. The first 0.7% by adding more value exposure (to domestic and intl), and the second 0.6% by adding more value exposure to domestic (micro to targeted value), to intl (from small to small value), and to EM (from market to value and smaller cap).

Again its factor exposure driving return differences (not DFA alpha). Greater risk (factor) exposure leading to higher expected returns (and over the past 8 years, higher realized returns). So in my view this is not an apples to apples comparison leading to a “1.3% advantage”, its due to high risk (factor) exposure. Consistent with the Fama-French research.

There is a lot more noise in an asset class focus (just compare differences in index returns within asset classes) than in a factor load focus. With the former there is (perhaps conveniently) more room to claim a 1-3% “total advantage”, which seems to disappear when focusing on the latter.

Now I don’t disagree with arguments that use of DFA funds allows for a more even factor tilt across global markets, or allows to lower beta through use of funds with higher value and small cap tilts (although the alternatives are increasing), what I don’t see is evidence to support the claim that DFA adds 1-3% in alpha beyond factor exposure.

Robert
.
asamir
Posts: 75
Joined: Thu Feb 22, 2007 7:51 pm

Asset class allocation

Post by asamir »

Asset class allocations (large/medium/small, value/growth) are labels used to imperfectly describe factor exposure. Portfolios with the same factor exposure tilts are comparable, regardless of the "lack of US large growth" or similar. The notion of "asset class comparison" is faintly ridiculous when your objective is to achieve the cheapest possible exposure to the desired factors.

Bottom line: of course DFA funds perform better than indices based on given asset classes: those indices don't have the same factor exposure.

Comparing e.g. a DFA large value fund to a US large value index makes no sense: the DFA fund is different by design. And unsurprisingly, if you replicate the factor exposure with non-DFA funds you end up with similar performance.

DFA funds are very very valuable because they provide a time efficient mechanism of securing desired factor exposures. They are cost-effective.

Suggesting that they are somehow better because they provide 1-3% more yield than comparable indices is like suggesting that they're better than lottery tickets and is simply disingenuous.

Anthony
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

You don't need look far to figure out why factor loadings should be a minor consideration within the asset (class) allocation process and not a reliable way to compare various portfolios of asset classes based on different indexes/funds.

1) The Russell Mid Value index and the DFA US Core 1 index have had the same 0.1 size loading since 1986. But the Mid Value Index has a weighted average market cap of approximately $7B, the DFA US Core 1 fund has a weighted average market cap of aproximately $47B.

2) The MSCI 1750 Small Value index and the DFA US Vector index have had the same 0.4 size loading since 1992. But the MSCI index has a weighted average market cap of approximately $2B, the DFA US Vector Fund has a weighted average market cap of approximately $20B.

In fact, the US Vector fund has a weighted average market cap more than double the size of the Russell Mid Value index, but a size loading 4X higher than the index.

3) Since 1995, CRSP 10 (the smallest of the 10 deciles) has a size loading of only 0.85, which is larger that CRSP 9 and 8, and approximately the same as CRSP 7 (.96, .88, and .82 respectively).

Half the time, the factor loadings don't even remotely correspond to the actual portfolio characteristics on the asset class level. Now we are supposed to use them exclusively for portfolio allocations? I guess if it leads you to the results you desire (regardless if portfolio comparisons contain extreme asset allocation disparities), I can see how it would make sense to some.

Just for the heck of it, I checked on the most efficient way to develop a portfolio of DFA funds in the same stock/bond and regional allocations as the ETF listed above with a consistent 0.2/0.4 size and value load in each region. The allocation is as follows:

19% TA US Core 2, 15% US TM Marketwide Value, 4% US TM Targeted Value, 17% TA World xUS Core, 8% TM Int'l Value, 4% Int'l Small Value, 7% EM Value, 25% Intermediate Government

It has had a +12.2% compound return from 1/03 through 10/10, approximately 1% above the ETF blend and the much more market oriented DFA component blend.

In order to achieve a particular factor target (if you want to use that as your asset allocation guide), using DFA you can afford to hold much more of your allocation in large/market oriented asset classes resulting in a much less extreme portfolio. The ETF portfolio has 60% of its equity allocation in mid value/small value/small/micro asset classes that represent extreme deviations from the market, while the DFA blend has only 10% of its equity allocation in non-Core/large value portfolios. And while "market tracking error" hasn't reared its ugly head in the better part of a decade, the painful experience of 84-90 and 95-99 reminds us that the ability to hold a much less extreme (in terms of the amount dedicated to remote small value/small/micro asset classes) portfolio and achieve the same or higher returns is a major benefit.

So, to conclude from an earlier post of mine, the investment goal should be to develop an overall asset allocation, and then attempt to achieve the highest factor loadings within each asset class (for example, the 0.6 value tilt for DFA US Large Value/TM Marketwide Value is preferrable to the 0.4 value tilt for Russell 1000 Value or MSCI 750 Value). Of course, if you wish to skip the multifactor regressions (as 99.9% of investors should do), then just check the average P/E or P/B of the associated index, and choose the lowest possible in concert with the highest diversification. For small value or small blend asset classes, you also want the lowest market cap possible, and diversification is much more important.
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T »

Multifactor Advisor wrote:You don't need look far to figure out why factor loadings should be a minor consideration within the asset (class) allocation process and not a reliable way to compare various portfolios of asset classes based on different indexes/funds.
Well – it depends on your starting point. If you believe the Fama-French research that its factor exposure that matters, not “asset class” exposure, then the focus should be on structuring portfolios to achieve the desired factor exposure to match portfolio expected return to needed return (and to ability and willingness to take risk). This is more effectively done through a factor exposure focus which in my opinion should be given major (not minor) consideration (with asset classes just representing different combinations and levels of factor exposure). If however, you believe that its ‘asset classes’ that matter (however you wish to define them), and not factor exposure then I could see how giving a major consideration to asset classes (and not factor exposure) makes sense. Now estimating factor exposure (through factor loads) is not perfect, but its better than current alternatives IMO.

For example - lets take an ‘asset class’ approach for sake of argument – following your guidance. The investor goes to Morningstar and wants to invest in smallcap value. So under the ‘asset class’ smallcap value the investor sees the following funds: DFA US Vector and Rydex Small Cap Pure Value. Within the same ‘asset class’ the market cap varies from $20bn to $0.5bn. Perhaps an extreme example – but this illustrates my earlier point that there is a lot more noise in an asset class focus than a factor load focus.

FWIW - from January 1995 to October 2010 here are the size loads for the 1-10 deciles using the available data from Ken French’s website. Over this period size loads increase with each incrementally smaller size decile apart from decile 9 and 10. Doesn't look too wildly off to me.

Code: Select all

January 1995 to October 2010

Size loads

Decile  1     -0.28
        2      0.10
        3      0.33
        4      0.38
        5      0.45
        6      0.68
        7      0.83
        8      0.93
        9      1.12
Decile 10      1.12

Data source: Ken French website
Robert
.
Multifactor Advisor
Posts: 266
Joined: Wed Aug 18, 2010 12:42 pm

Post by Multifactor Advisor »

Robert:

1) The original Fama/French research isolated high, medium, and low book to market (ie. value, blend, and growth) stocks in the large and small cap asset classes (as defined by CRSP 1-5 and CRSP 6-10) in the US and only the large cap segment of the value/growth spectrum in the International markets (due to a lack of small cap style data availability at the time). So the initial FF research was soley asset class based. Their initial research didn't say "diversified portfolios with 0.4 to 0.7 HmL coefficients have....., or diversified portfolios with 0.6 to 1.0 SmB coefficients have......). The "model" came along after simply as a way to think about where various asset classes sat in multifactor space. Not as an attempt to replace any consideration about asset classes.

2) On the CRSP factor loads, I double checked, not sure why we get different results, but I am still showing an increasing size load as we go from CRSP 10 to CRSP 7. (our averages for 1-5 are closer but not identical).

3) All I am saying is: lets use some common sense here. Does the $40B+ Core 1 Index (or Marketwide Value Index) have the same 0.1 exposure to size dimension as the $7B Russell Mid Value index? Should the $20B DFA Vector Index have the same size loading as the $3B Vanguard Small Value Index? Of course not. The model is not working here. The model is helpful, but an abundance of mid cap stocks invalidates the portfolio outputs. Here's another example: the Russell Mid Growth index has had a size tilt of 0.3 since '86, vs. 0.1 for Russell Mid Growth (yet both have identical weighted average sizes). If 0.2 difference doesn't seem large, realize that a Total Stock Index, with an approximate $60B weighted average market cap, has only a 0.2 greater size load than the $7B Russell Mid Cap Index--a $50B difference!

4) Using a sensible asset class framework (in which large cap stock portfolios are defined as weighted average market caps above $25B, mid caps are in the $3B to $15B range, and small caps are $2B and below) leads you to much more precision than the approximate $20B to $30B potential differences that I highlight above using only factor loads.
For example - lets take an ‘asset class’ approach for sake of argument – following your guidance. The investor goes to Morningstar and wants to invest in smallcap value. So under the ‘asset class’ smallcap value the investor sees the following funds: DFA US Vector and Rydex Small Cap Pure Value. Within the same ‘asset class’ the market cap varies from $20bn to $0.5bn. Perhaps an extreme example – but this illustrates my earlier point that there is a lot more noise in an asset class focus than a factor load focus.
Using the Vector fund as an example, there is no reason (other than avoiding common sense) one would categorize DFVEX as a small value fund (M* calls it a mid blend/value, which is strange when you realize that the largest single holding in the Russell Mid Cap Index is $15B and smaller than the weighted average holding of Vector). I have no idea how comparable these factor loads are, but a 20% S&P, 30% DFA Large Value, 20% DFA Small, 30% DFA Small Value mix has the same asset class structure (w/o as much in midcaps), same approximate average portfolio size and price to book, and a 99% historical correlation of annual returns since 1928. If you can calculate a multifactor regression, this asset class view should be obvious.

5) Here is a caveat I would add to even the most ardent factor follower: the more midcap oriented your portfolio is, the less relevant the model will be. If you (referring to anyone reading) like it, great. Just realize it isn't perfect, some additional common sense considerations would be beneficial. Finally, calling someone who doesn't follow it at the behest of all other considerations is not unethical, as Rick Ferri mistakenly said above. There are very good reasons (as I've outlined above) why the model may have limited practicality.
User avatar
abuss368
Posts: 27850
Joined: Mon Aug 03, 2009 2:33 pm
Location: Where the water is warm, the drinks are cold, and I don't know the names of the players!
Contact:

Post by abuss368 »

I believe Arnold Schwarzenegger owns a portion of DFA.
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T »

.
(1) On Fama-French - their 1993 seminal paper (linked below) looked at 5 size based portfolios by 5 value-growth portfolios (for total of 25). Not just 2 size based portfolios.
http://faculty.fuqua.duke.edu/~charvey/ ... n_risk.pdf Here’s a extract from that seminal paper:
  • “If the five factors capture the cross-section of average returns, they can be used to guide portfolio selection. The exposures of a candidate portfolio to the five risk factors can be estimated with a regression of the portfolio’s past excess returns on the five explanatory returns. The regression slopes and the historical average premiums for the factors can then be used to estimate the (unconditional) expected return of the portfolio. “
(2) On decile 1-10 factor loads, I downloaded these yesterday so not sure why there's a difference. FWIW the table below shows the size loads for the 25 portfolio from the FF paper linked above.

Code: Select all

Here are the size loads from the FF paper:
From table 6: July 1963 - December 1991

           Low     2      3      4    High

Small     1.46   1.26   1.19   1.17   1.23
2         1.00   0.98   0.88   0.73   0.89
3         0.76   0.65   0.60   0.48   0.66
4         0.37   0.33   0.29   0.24   0.41
Large    -0.17  -0.12  -0.23  -0.17  -0.05

(3) On size load variation – if I'm not mistaken, I think you were advocating that the best way to capture the size premium was to avoid decile 1. i.e. that the size premium is significantly non-linear with small differences in size premium from decile 2 and above. These small differences may be showing up in small size load differences you mention.

(4) On precision of portfolio characteristics – I have not found a more precise approach to matching portfolio characteristics to expected returns than through estimates of factor exposure (again while not perfect it has lower noise than an asset class approach).

(5) On common sense – I agree there should be common sense added to all decisions, but do not think that saying the earlier DFA portfolio had a 1.3% “total advantage” when the reason was additional [risk] factor exposure (in my opinion its common sense to say it was not an ‘advantage’ but due to higher risk exposure [your earlier comparison was between DFA portfolios]).

Robert
.
lorneabramson
Posts: 64
Joined: Tue Jul 29, 2008 10:07 am
Location: Burlingame, California

DFA's Financial Engines (using DFA funds)

Post by lorneabramson »

Post Reply