How Does DFA Compare To Vanguard ?

Comments

• msf

  March 2016

  "For example, when a new stock enters an index, a traditional index fund has to add it all at once. This can push the price up before the purchase is completed."
  
  Bzzzzt. Wrong answer.
  
  "The public at large has been fully aware of any changes scheduled to be made in an index fund portfolio before the index change is implemented. This is not to say, of course, that the manager of the index fund will make the change at the close of trading on the day that the index changes. Even the “manager” of the original SPDR had considerably greater flexibility than an exact timing match of index and fund changes. [Citation to SPDR 2001 prospectus.]"
  
  2002, Gary L. Gastineau (managing director at Nuveen), Equity Index Funds Have Lost Their Way.
  (It was easier to find than reading through prospectuses.)
  
  The AssetBuilder column continues:
  "DFA, by comparison, is working to characteristics, not names. So they build their portfolios differently. By doing so, they avoid the cost bumps of stock list-following index funds"
  
  This is correct, and does represent DFA's difference. It's not so much a matter of timing of purchases but in how they they replicate indexes - they can even substitute one stock for another (e.g. Pepsi for Coke). The objective is to track the index, not beat it, as would be the case for a quant fund using the same stock substitution technique.
• heezsafe

  March 2016 edited March 2016

  @msf Is this how DFA mitigates tracking error accumulation? (and have you ever seen DFA data for tracking error of their funds? I would bet they do calculate it, so they can monitor if they are avoiding it, but perhaps they keep the data in-house?)
  
  Also, FWIW, this from pdf docs relating to index changes to be made to the "J Hancock Dimensional Indexes" on certain dates:

  The date used for calculating index holdings, new index weights, and share adjustments for a given Enhanced Redistribution will change from the Data Date to the Redistribution Date, which will generally be one or two days prior to the announcement for an Enhanced Redistribution.
• msf

  March 2016

  I really don't know much more about DFA's funds - I've read through their literature from time to time to understand why they don't consider their funds index funds. It seems to amount to greater flexibility in various dimensions (such as the stock substitutions).
  
  My impression is that they manage their equity pseudo-index funds much like bond index funds - they identify securities that have the same or similar attributes to those in the index (IMHO easier done with bonds that are quantifiable as to maturity, duration, rate, risk, etc.). Then they use these securities to build a fund that approximates the index.
  
  Dimensional (like many indexers) calls rebuilding an index anew a Reconstitution. The quote you gave describes how DFA indexes (as opposed to funds) manage intra-reconstitution events (such as a merger); Dimensional calls this an Enhanced Redistribution process. It's not a description of how a fund tracks an index.
  
  The Data Date is a periodic date (typically last day of each month) that is (or was) used for index calculations. It appears that for enhanced redistributions (tweaking the index due to external events), data from a more current date will be used. So instead of using data from the end of the last month to calculate securities and weights, current data will be used. ("Current" meaning a couple of days before the change is announced.)
  
  Here's a sample Dimensional methodology page. You can find info on all the John Hancock Dimensional Indexes here.
  
  
• davidrmoran

  March 2016

  Graph DFA LC and SC since 1/1/99 against FCNTX and FLPSX and take a look. Can't be true, right? So saith all the wise.
• Tony

  March 2016

  Another take on DFA vs. Vanguard: http://www.altruistfa.com/dfavanguard.htm
• msf

  March 2016

  Altruist Financial Advisors seems to have some curious notions regarding Vanguard:
  
  "For those who do not have access to DFA, we suggest they limit their bond holdings to domestic bond funds." A blanket rejection of VTIBX without even mentioning it.
  
  (That fund was started in 2013, and the page has information from at least 2014, so this is not a matter of the fund not existing at the time of the writing.)
  
  " However, the existence of an ETF share class makes [VTMGX] quite a bit more capital-gains tax-efficient than it would otherwise be."
  
  Vanguard has always stated that the tax value of ETFs (at least for cap-weighted index funds) is overrated (i.e. they are not "quite a bit more capital-gains tax-efficient'). The data bear this out not only generally but specifically for this fund.
  
  The fund was created in 2014 as the result of a merger between Vanguard's tax-managed foreign developed markets fund and its "regular" fund. See:
  http://mutualfundobserver.com/discuss/discussion/12700/what-happened-to-vdmix
  
  So I went back to look at cap gains distributions for the "regular" fund VDMIX (no ETF share class, not tax-managed):
  
  Per share cap gains distributions (from the 2013 prospectus:
  2013: - (six months ended April 30)
  2012: - (full year)
  2011: -
  2010: -
  2009: -
  2008: $0.005
  
  To put that last figure in perspective, total distributions/share were $0.387.