Twenty year anniversary - Fama-French paper

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Twenty year anniversary - Fama-French paper

Post by Robert T » Sun Aug 18, 2013 5:57 am

.
It has been 20 years since the Fama-French paper on Common Risk Factors in the Returns on Stocks and Bonds. An excellent paper in my view, and has been the single most influential journal article on my personal investment decisions (which together with the data on Ken French's website, has unusual practical application for individual investors compared to most other financial journal articles).

The paper identified five 'risk factors' - market, size, value, term and default risks, and showed that differing exposure to these factors explained most of the historical return premiums across a spectrum of 'funds'. These findings have been much debated - some saying they are return anomalies (not risks) and they will be arb'ed away as money flows into these segments of the market, while others say, as per Fama-French, that these factors 'proxy' risks that are rewarded with premium returns.

Its been 20 years since the article was published - a period where the annualized returns have again showed a premium of equities over bonds, small over large, value over market, intermediate term over short-term, and corporate over government (using DFA and Vanguard funds - results below). Will this continue? This was the same question asked 20 years ago. There are obviously no guarantees and there have been fairly long periods where this pattern did not occur. Personally I side with the risk story - premium returns for higher risk - and just as we expect equity returns above bond returns (in compensation for higher risk), I think the same is true of size, value, term, and default risks. Time will tell.

Annualized return: July 1993-June 2013 (growth of $1 over this period in parenthesis)

Stocks
Vanguard TSM.........= 8.8%....(5.5)
DFA Large Value......= 9.8%....(6.4)
DFA Small Cap.......= 10.6%....(7.5)
DFA Small Value.....= 12.4%...(10.4)
DFA Micro.............= 11.9%....(9.5)

Bonds
DFA One-year............................=3.7%....(2.1)
DFA Intermediate Gov..................=5.9%....(3.2)
Vanguard Intermediate Inv. Grade*..=6.2%....(3.3)

*Barclays Intermediate Credit from July 1992-Nov 1992, the Vanguard fund thereafter.

Robert
.

Calm Man
Posts: 2917
Joined: Wed Sep 19, 2012 9:35 am

Re: Twenty year anniversary - Fama-French paper

Post by Calm Man » Sun Aug 18, 2013 11:39 am

Robert, being the cynic that I am...
I know that a great deal of money has been made by DFA and the advisers that work with it as a result of the FF work. Did this paper happen to precede or closely follow the launch of the DFA funds???

dodonnell
Posts: 413
Joined: Wed Oct 29, 2008 6:48 pm

Re: Twenty year anniversary - Fama-French paper

Post by dodonnell » Sun Aug 18, 2013 12:10 pm

Calm Man wrote:Robert, being the cynic that I am...
I know that a great deal of money has been made by DFA and the advisers that work with it as a result of the FF work. Did this paper happen to precede or closely follow the launch of the DFA funds???
David Booth founded DFA in 1981 ... I believe originally in an apartment in Brooklyn.
Originally, the focus was squarely on SCV for institutional purposes.
He was Eugene Fama's graduate research assistant in the 60s/70s at what is now ... the Chicago Booth School of Business.
Last edited by dodonnell on Sun Aug 18, 2013 12:13 pm, edited 1 time in total.

User avatar
baw703916
Posts: 6679
Joined: Sun Apr 01, 2007 1:10 pm
Location: Seattle

Re: Twenty year anniversary - Fama-French paper

Post by baw703916 » Sun Aug 18, 2013 12:10 pm

I'm not sure what that difference that makes. While DFA manages funds that are based on the three factor model, and there are many advisors whose clients invest in those funds, you don't have to have any DFA funds, or an advisor, to achieve a tilted portfolio. I don't.

The paper did a statistical analysis, and made a conclusion that, if it is valid, would have to hold going forward as well. Namely, that small stocks and value stocks tend to have higher returns over the long term, because they carry extra risks compared to equities in general. The data from the last 20 years seem to support this.
Most of my posts assume no behavioral errors.

Beagler
Posts: 3442
Joined: Sun Dec 21, 2008 7:39 pm

Re: Twenty year anniversary - Fama-French paper

Post by Beagler » Sun Aug 18, 2013 12:11 pm

Morningstar lists 2/19/1993 as the first trade date for DFA Large Cap Value (DFLVX). (Open to correction if others have a more accurate date.)
“The only place where success come before work is in the dictionary.” Abraham Lincoln. This post does not provide advice for specific individual situations and should not be construed as doing so.

User avatar
momar
Posts: 1359
Joined: Sun Nov 13, 2011 12:51 pm

Re: Twenty year anniversary - Fama-French paper

Post by momar » Sun Aug 18, 2013 12:15 pm

If only they had the insight to look at the esoteric metric profitability, which no one would have ever suspected of being related to a company's health or stock performance.
"Index funds have a place in your portfolio, but you'll never beat the index with them." - Words of wisdom from a Fidelity rep

User avatar
baw703916
Posts: 6679
Joined: Sun Apr 01, 2007 1:10 pm
Location: Seattle

Re: Twenty year anniversary - Fama-French paper

Post by baw703916 » Sun Aug 18, 2013 12:20 pm

momar wrote:If only they had the insight to look at the esoteric metric profitability, which no one would have ever suspected of being related to a company's health or stock performance.
Now if we could only identify those companies for which profitability will be persistent going forward...
Most of my posts assume no behavioral errors.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Twenty year anniversary - Fama-French paper

Post by Robert T » Sun Aug 18, 2013 2:38 pm

Calm Man wrote:Robert, being the cynic that I am...
I know that a great deal of money has been made by DFA and the advisers that work with it as a result of the FF work. Did this paper happen to precede or closely follow the launch of the DFA funds???
Calm Man,

The paper preceded the launch of the DFA value funds listed. The first draft of the Journal of Financial Economics paper linked in the OP was received by the Journal in July 1992 (and no doubt there were earlier forms of this work on which the paper was based). The DFA large value fund was launched 2/19/1993, the DFA small value fund was launched 3/2/1993.

For the micro cap fund, it was based on earlier work on the size premium by Rolf Banz http://perrittmutualfunds.com/media/Ban ... ffects.pdf . Again, first draft received by the Journal of Financial Economics in June 1979, the DFA micro cap fund was launched 12/23/1981, the DFA small cap fund was launched 3/19/1992.

So I give DFA its due - their approach is research based. Now I agree, some advisors/advocates of DFA funds go over board and sometimes says DFA adds 1-3% in alpha beyond factor exposure due to patient trading, momentum screens, block trading, exclusion of utilities etc. etc. But these strategies seem to simply change a fund's exposure to the three equity factors in OP article, rather than add alpha.

Robert
.
Last edited by Robert T on Sun Aug 18, 2013 3:34 pm, edited 2 times in total.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Twenty year anniversary - Fama-French paper

Post by Robert T » Sun Aug 18, 2013 2:54 pm

baw703916 wrote:I'm not sure what that difference that makes. While DFA manages funds that are based on the three factor model, and there are many advisors whose clients invest in those funds, you don't have to have any DFA funds, or an advisor, to achieve a tilted portfolio. I don't.
I completely agree. I set up two portfolios in the M* tracker - one with Vanguard/iShares/Bridgeway funds (called the 'ETF' portfolio) the other with DFA Tax-Managed funds. Both portfolios have the same estimated value and size loads across US:EAFE:EM markets, and the same term and default loads (if anything the DFA portfolio has a higher beta exposure due to exclusion of utilities, and a slightly higher term load). Since inception, which was 11-06-08 - almost 5 years, the annualized returns of the two portfolios:

Annualized return since 11-06-08

ETF = 13.6%
DFA TM = 13.6%

The same annualized returns over this period from the same estimated factor exposure (no 1-3% alpha so far). While there have been difference in year-on-year returns, the annualized returns over the past 5 years or so have been the same.

Will be interesting to see if there will be a return difference going forward with inclusion of the profitability factor in the DFA fund screens.

FWIW - I don't use DFA funds for my retirement investments (I get my target factor exposure from non-DFA funds), although I do use DFA funds in a 529 plan.

Best,

Robert
.

SnowSkier
Posts: 264
Joined: Fri Mar 16, 2012 12:21 pm
Location: Austin

Re: Twenty year anniversary - Fama-French paper

Post by SnowSkier » Sun Aug 18, 2013 6:19 pm

Robert T wrote:I set up two portfolios in the M* tracker - one with Vanguard/iShares/Bridgeway funds (called the 'ETF' portfolio) the other with DFA Tax-Managed funds. Both portfolios have the same estimated value and size loads across US:EAFE:EM markets, and the same term and default loads (if anything the DFA portfolio has a higher beta exposure due to exclusion of utilities, and a slightly higher term load). Since inception, which was 11-06-08 - almost 5 years, the annualized returns of the two portfolios:

Annualized return since 11-06-08
ETF = 13.6%
DFA TM = 13.6%

The same annualized returns over this period from the same estimated factor exposure (no 1-3% alpha so far). While there have been difference in year-on-year returns, the annualized returns over the past 5 years or so have been the same.
...
Robert, what do your 'ETF' sample portfolio target allocations look like nowadays? I wasn't sure if the sample portfolios listed in "collective thoughts" were up to date or not.

If I'm interpreting correctly, it looks like:
23% VOE (VG mid cap value)
09% IJS (iShares S&P600 small cap value)
05% BRSIX (Bridgeway ultra small)
23% EFV (iShares EAFE value)
05% SCZ (iShares EAFE small)
10% VWO (VG Emerging)
25% IEI (iShares 3-7yr Treasury)

Can you please clarify/correct or confirm?

Many thanks for All your contributions!

Calm Man
Posts: 2917
Joined: Wed Sep 19, 2012 9:35 am

Re: Twenty year anniversary - Fama-French paper

Post by Calm Man » Sun Aug 18, 2013 6:56 pm

Robert, thank you for providing the dates of the research paper and launch of the funds. It supports what I expected as I have done a good deal of work consulting for pharmaceutical companies. The timing is not chance but well planned in advance of the fund launch. I am not disputing the research simply proposing that the research was not done out of indifferent academic interest but ultimately it was all about money, surprise, surprise. And one poster asks why it matters? It matter a lot. because in 1993 it was all retrospective research which people who are familiar with reser4ach know is hypothesis-generating, meaning it doesn't prove anything. if it has proven subsequently to be valid so be it but it could have been the other way as well.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Twenty year anniversary - Fama-French paper

Post by Robert T » Sun Aug 18, 2013 7:48 pm

SnowSkier wrote:Robert, what do your 'ETF' sample portfolio target allocations look like nowadays? I wasn't sure if the sample portfolios listed in "collective thoughts" were up to date or not.

If I'm interpreting correctly, it looks like:
23% VOE (VG mid cap value)
09% IJS (iShares S&P600 small cap value)
05% BRSIX (Bridgeway ultra small)
23% EFV (iShares EAFE value)
05% SCZ (iShares EAFE small)
10% VWO (VG Emerging)
25% IEI (iShares 3-7yr Treasury)

Can you please clarify/correct or confirm?
That's it, it has not changed. Vanguard recently changed from tracking MSCI MidCap Value to CRSP MidCap Value. The Russell MidCap Value index is closer to the the MSCI MidCap Value index than CRSP MidCap Value - so could use iShares Russell MidCap Value (IWS) instead of VOE - but think both are fine.

Robert
.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Twenty year anniversary - Fama-French paper

Post by Robert T » Sun Aug 18, 2013 8:48 pm

Calm Man wrote:Robert, thank you for providing the dates of the research paper and launch of the funds. It supports what I expected as I have done a good deal of work consulting for pharmaceutical companies. The timing is not chance but well planned in advance of the fund launch. I am not disputing the research simply proposing that the research was not done out of indifferent academic interest but ultimately it was all about money, surprise, surprise. And one poster asks why it matters? It matter a lot. because in 1993 it was all retrospective research which people who are familiar with reser4ach know is hypothesis-generating, meaning it doesn't prove anything. if it has proven subsequently to be valid so be it but it could have been the other way as well.
Calm Man,

I understand what you say, but in this case I don't think DFA is like the example you gave of pharmaceutical companies (i.e. having a product then finding someone to do research to support its launch). I think for the small cap and value research, DFA looked at the research then determined if a product could be developed to replicate the findings with live funds (transaction costs included) - yes they saw a market for it. The same could be said for the recent research on the profitability premium by Novy-Marx. This was first picked up in a product by AQR capital (Asness), then DFA - but both first looked at the research. As an additional bonus, Ken French makes public the raw data from his research so we can replicate their results and make our own determination of the robustness of the findings. The Fama-French research using historical data shed light on how investors (collectively) have responded to risk (providing a more complete explanation than the CAPM). If we believe that investors historical behavior towards risk will on average be the same/similar going forward then higher risks (e.g. stocks over bond, value over market etc) should be rewarded with higher returns (or no-one would take these risks). Over the last 20 years this has proved to be the case - but obviously no guarantees.

Now there are also less obvious research linked products in DFA's line-up e.g. their 'extended quality' corporate bond funds which were likely introduced to meet investor demand, rather than derived from research. And they are a for profit company with a 500 billion AUM target (as I understand). Their core funds with lower tilts can absorb larger fund inflows without as much market impact and some of the earlier products such as the WV529 plan shifted from a larger tilted portfolio to a less of a tilt through use of core funds. In my view, at least part of this was driven to manage larger fund inflows. So I am not oblivious to their profit motives.

Robert
.

emtl97
Posts: 2
Joined: Wed Sep 18, 2013 10:40 am

Re: Twenty year anniversary - Fama-French paper

Post by emtl97 » Fri Oct 04, 2013 3:51 pm

Hi Robert, hope this is an on topic question. I was where to find the data for DEF and TERM factors from the original paper? I was able to get the Equity factors from Kenneth French's site. Also what software do people use to run FF regressions? I'm thinking of doing this in R.

Thank you.
Last edited by emtl97 on Wed Oct 16, 2013 6:46 pm, edited 1 time in total.

berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Twenty year anniversary - Fama-French paper

Post by berntson » Fri Oct 04, 2013 4:07 pm

R works well and has the significant advantage of being multi-platform.

User avatar
LadyGeek
Site Admin
Posts: 47440
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Twenty year anniversary - Fama-French paper

Post by LadyGeek » Fri Oct 04, 2013 4:11 pm

Welcome! If you want to know how to do FF regression in R, you've come to the right place. Take a look in the wiki: Fama-French three-factor model analysis. The section under R has the source code and methodologies you need.

This is a collaborative effort between this forum and our sister Canadian site (Financial Webring Forum). You should read the referenced wiki article first: finiki:Multifactor investing - a comprehensive tutorial.

Both of these articles have extensive forum discussion threads, check under "External links." You are welcome to post in the the discussion threads if you want to ask questions or provide comments. Or, start a new thread.

Sorry, I can't answer about the DEF and TERM factors.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

emtl97
Posts: 2
Joined: Wed Sep 18, 2013 10:40 am

Re: Twenty year anniversary - Fama-French paper

Post by emtl97 » Wed Oct 16, 2013 8:42 pm

Thank you LadyGeek for the welcome and useful information. It looks like our man Fama has won the Nobel, so good timing for this thread.

I didn't find the TERM and DEF factors from the original paper so if anyone on here knows where I can get that data give me a shout.

Anyways, question for everyone, seeing that HML, SMB, UMD (momentum) maket factors all have close to zero correlation, seems a good idea to diversify among those? So what's the cheapest way to diversify among those? Seems like most Value and Momentum funds will be long market factor still quite a bit.

berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Twenty year anniversary - Fama-French paper

Post by berntson » Thu Oct 17, 2013 7:08 pm

emtl97 wrote: I didn't find the TERM and DEF factors from the original paper so if anyone on here knows where I can get that data give me a shout.
I'm trying to figure this out too. Drop me a line if you find an easy way to get the data.
emtl97 wrote: Anyways, question for everyone, seeing that HML, SMB, UMD (momentum) market factors all have close to zero correlation, seems a good idea to diversify among those? So what's the cheapest way to diversify among those? Seems like most Value and Momentum funds will be long market factor still quite a bit.
Two great places to start are Robert T's collective thoughts (http://www.bogleheads.org/forum/viewtop ... =10&t=7353) and the introduction to multi-factor investing (http://www.finiki.org/wiki/Multifactor_ ... e_tutorial).

AQR has funds targeting UMB, but they are expensive and only available to large investors or through an advisor. The main thing is to watch out for negative momentum (which deep value funds can have in spades). I don't know of any other reasonable UMD funds.

Otherwise, you should be able to do all the HML and SMB tilting you need with Vanguard funds and maybe a couple other from iShare or PowerShares. My own portfolio uses VOE, VBR, EFV, and VSS. I also really like PXF for international value. I posted a thread earlier on why investors should avoid microcap funds (http://www.bogleheads.org/forum/viewtop ... &p=1819332). With the run-up in dividend stocks, I would be cautious about DGS and DLS, two popular international SV on these boards.

User avatar
Ketawa
Posts: 1933
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Twenty year anniversary - Fama-French paper

Post by Ketawa » Thu Oct 17, 2013 7:55 pm

berntson wrote:
emtl97 wrote: I didn't find the TERM and DEF factors from the original paper so if anyone on here knows where I can get that data give me a shout.
I'm trying to figure this out too. Drop me a line if you find an easy way to get the data.
ClosetIndexer has a set, you might want to try him: http://www.bogleheads.org/forum/viewtop ... 0&t=103816

Post Reply