DFA vs. Vanguard

I’m interested to hear people’s opinions on how DFA compares to the bigger index fund players. I know Vanguard isn’t the only one that offers index funds/ETF’s, but they’re the most notorious. We could just as easily compare to iShares.

Admittedly, I know little about how index funds or ETF’s are constructed or the indices that they track. I don’t know if VFINX pays S&P a fee to track their index. I don’t know if they have to rebalance the fund every time S&P reconstitutes the index, or how the index is rebalanced. (From the untrained eye, it seems like rebalancing a market-value weighted index would somehow defeat the purpose.)

If anybody has any real insight into how the two differ, I’d be delighted to hear it.

No real insight here, but are you looking to begin slangin’ DFA to your clients? Many indy RIAs use them.

^Maybe someday in the future. For the next five years, I’m focused on tax compliance and planning, and hope to get a valuation business off the ground.

Maybe in five to seven years, I can start doing a financial planning/investment business. And I’m seriously thinking about using DFA. I don’t know if you have to be an RIA and have a clearing broker like Pershing, or if they have their own B-D. I’ll cross that bridge when I get there.

With this thread, I’m really more interested in finding out what people think about the differences between the DFA Large-cap blend fund (whatever it is) and the typical S&P 500 index fund. Why pick one over the other?

Lower fees. There is such thing as a free lunch when it comes to fund selection.

Vanguard pioneered the implementation of the index fund, and it basically created ETFs of their index funds so as not to lose market share to people who wanted to do their investing with ETFs as opposed to MFs.

When you are indexing with a long term time horizon, it’s pretty much about fees vs. tracking error. I haven’t run the statistics on tracking error effects here, but last I checked is that Vanguard is the winner in the fees department. iShares might do better in the tracking error department (which is actually less important in the long term, as long as it isn’t huge). I’m not sure where DFW lands.

But the biggest value if you are a buy-and-hold type is that these strategies keep you from messing with your stuff. I’ve seen very believable research that says that people who follow a passive strategy often end up underperforming the index by way more than fees and tracking error. What happens is that people buy in to buy-and-hold when the market is good, then switch to cash at the gloomiest moment. Effectively they are market timing an index, but always buying at the top and selling at the bottom… ooops! I have occasionally made this very same mistake, which is why I now stick to a model.

DFA is a quant cult, not an index shop. RIAs have to go to a “camp” for a few days on the DFA campus and then sign a pledge to turn over $25mm of their assets in the first year. Seriously.

Their products are designed to be used together, not on a one-off basis. If you go DFA you go all the way. Right until they blow up like every other quant shop eventually does.

vanguard is the king of low fee

DFA = enhanced indexing vs. Vanguard = pure indexing. DFA is driven by a bunch of academics and quants that capture the risk/return premium from tilting a portfolio towards small cap value equities, rebalancing away from high P/E’s, etc. I researched this a while ago for work, and I believe DFA funds do well for domestic equity, but EM equity, and other areas they were closely matching the returns of a regular index fund albeit with slightly higher fees. I have seen firms that mix DFA in with Vanguard and other mutual funds, but I’m not sure if DFA frowns upon that.

The biggest advantage of dfa for an ria is that they get to offer an “institutional” product to your clients as opposed to Vanguard, which they can purchase themselves. So there’s a barrier to entry which makes it easier to sell.

Dfa doesn’t track an index, BC the index you pick is arbitrary Anyways. So they don’t pay an index provider a licensing fee. Vanguard just switched from Russell and Msci to crsp and FTSE to lower their costs and reconstitution costs. Honestly, there’s a ton of literature on this if you do a search. It’s a debate as old as schewser vs stalla and CFA vs mba. Dfa prolly won’t take you because you don’t have any clients tho.

Also, sweep uses the word quant fund really loosely, the products are really well diversified and plain vanilla IMHO and the turnover is low. Sweep is right about it being a cult though, most fund co’s would be excited to have 1 fund in your portfolio, dfa expects almost 100‰ of your book.

Depends on your perspective. In the mutual fund world they’re hard core quant. Compared to someone like DE Shaw, yes, they’re pretty plain vanilla.

If I were to ever start my own RIA shop, I wouldn’t be concerned with tracking error. I think most individuals only care how their portfolio performs on its own accord–not whether it deviates from a commercial index by .72% (or what have you).

The tracking error is a small correction, but it eats away at your returns over time. What happens is that after 10 years, the market goes up X%, and the fund goes up Y% where Y is less than X by more than can be explained by your fees, and then your clients may want to know why, since you told them you are just tracking the market.

There is no alpha because it’s an index fund, but because it doesn’t track the index perfectly, the net effect over time is a gradual eating of returns (because being down 20% and up 20% leaves you with less money than just being down 10% then up 10%, just on a smaller scale).

If you are hoping to capture some alpha, tracking error is necessary, but if you are just tracking an index, why lose even a little bit of return if it doesn’t benefit either your client or you???

The main reasons to go with a higher tracking error index or ETF is if there are differences in transaction costs, though these tend to be more important if you have a strategy that rebalances or does some kind of tactical adjustments frequently.

Tracking error is important for portfolio construction purposes. There are still plenty of active managers with low tracking error. They just choose to take security level bets instead of sector bets. If you don’t pay any attention to tracking error, you’re likely to end up using several star fund managers running concentrated portfolios. They may all have good performance, but you never know where they’re placing their bets. Sooner or later you’ll find they’re all betting on the same thing and when it doesn’t work you’re clients are going to have a bad time.

As an RIA you need to build well diversified portfolios. You can’t do that without have a good core group of managers. Then you build around those positions with low tracking error, concentrated portfolios.

Edit: I didn’t bother reading all the posts. Obviously I’m talking about tracking error from an active management standpoint. bchad is nicely covering why tracking error is important from an indexing perspective. That’s also why you should never buy passive fixed income instruments.

@Bchad - I think there is a fault in your assumption. You assume that a tracking error ALWAYS leads to lower return than the index, even in a passive fund. I don’t think that’s the case. EG - if Dow Jones decides to replace Citigroup with Traveler’s in the DJIA, then the DIA has to sell C and buy TRV. A similar, non-indexed passive fund would not have to make such a trade. It is not “forced” to sell low and buy high to track the index. In this case, it would have tracking error, but that would be a good thing.

@STL - Regarding all managers investing in the same stocks–I would agree with you if I had a dozen managers managing my large-cap domestic part of my portfolio. However, if I identify ONE fund that I really like (say, Davis New York Venture fund), then you don’t have that problem. I know what he likes to invest in, and since he’s my only domestic LC manager, I don’t have to worry about other domestic LC managers duplicating his positions.

And you say that you need to build well-diversified portfolios. That’s a given, but it’s not useful. Do you mean that you need to hold all stocks in the index? That I should invest in all 6,500 companies in the US stock market? Or can I just buy a mutual fund that has 30 stocks? We’ve all seen the graph that after 20 or so stocks, the additional diversification benefits become almost zero.

Let me know what you think. I appreciate all the feedback.

Owning hundreds to thousands of stocks in a portfolio with low turnover and tracking error? Lots of the research they use orginated in the 70’s and 80’s… its pretty old school and simple, its 97% the same thing as Vanguard. The shit I’m more worried about are bond funds that are $10B collateral, short 15B NP bond futures, long $19B NP swaptions, short soverign debt and long 1 share of AGG.

What company are you a wholesaler for? Totally cool, if you dont want to say tho.

Your B-chad example is correct, it costs money to reconstitute on the same day as an Index. A company like DFA will own a stock that’s not in the index, and when its added, they’ll sell it to someone like Vanguard. They earn a little liquidity premium for this. The question is, does it cover the difference in expense ratios. DFA also tilts away from the market more than vanguard, so some of the extra returns come from taking more risk, not patient trading.

Uh, maybe hold 30 stocks from every sector in every asset class is you really want to be diversified. The 20-30 stock rule means you get the same variance, returns will likely look nothing like eachother unless you match sector weights.

Strictly speaking, Greenman’s example is of a company trying to generate alpha by deliberately parting from the index and trading around it at clever times. Tracking error without alpha does eat returns over time.

Alpha is of course the justification for taking on higher tracking error in active trading (this is just at the very low extreme of activity); you’ll want alpha > (TE^2)/2 (approximately) in order to be large enough to counteract the tracking error drag on perofrmance, and I guess if you are comparing Vanguard and other funds, you’d really just want alpha to be larger than one half the difference of the squares of the tracking errors (haven’t checked the algebra, but that’s my off-the-top-of-my-head guess).

It still doesn’t change the fact that higher tracking error will eat away at returns over time, the only real question is whether you believe there really is alpha generated and if it is enough to compensate for that.

So yes, if you have some kind of alpha generating process, and it’s large enough, it makes sense to have more tracking error.

Moreover, most index fund tracking errors are measured in basis points, and so the hurdle is probably pretty small before an some kind of alpha process makes sense.

If Vanguard has a tracking error of 6 bps, and DFA or iShares has a tracking error of 14 bps, then you you need an alpha difference from trading of only 0.8 of a basis point annually to justify the higher tracking error. That’s a pretty low hurdle, as long as you believe that the alpha is really there.

You’ll also recall that I said tracking error is “less important in the long term as long as it isn’t huge.” So this is all splitting hairs pretty finely.

Low tracking error on index funds is probably more important for marketing than for choosing funds, the difference fees are almost certainly going to have way more of an effect than the difference in tracking errors, at least for cap-weighted indexes, which have the advantage of self-reblacing without trading except for when companies are added or dropped from the index.

Again, you’re looking at it from a hedge fund perspective. Every 40 Act mutual fund that’s labeled “quant” is going to have around 1000 names if it’s large cap or 400-800 names if it’s small cap. Generally they are lower tracking error but that doesn’t mean they don’t produce alpha. They just take small stock bets based on their black box calculations.

And…I’m not saying.

Buying one large cap manager is not diversified. At the very least you’ll see guys buy a style pure large growth and large value manager, then maybe add something like Yacktman or Fairholme to add alpha. The “core and explore” approach is pretty standard these days. Think about just the equity portion of your asset allocation. You’ll see lots of advisors putting 80% towards style pure funds to fill out the Morningstar boxes (mainly growth and value; blend doesn’t get much love), and use the remaining 20% on things like Ivy Asset Strategy, Blackrock Global Allocation, PIMCO All Asset, etc.

If Davis NY Venture was your clients’ only large cap holding, you’re having tough conversations today on why their 3 and 5 year ranking is in the 79th percentile.

No prob, I was just wondering so I could see how your firm views markets and invests, nothing degrogitory (sp?). I dont have any affiliation w/ DFA… just know everything about them. There are no back box calculations in their funds… they really just provide liquidity to the market when there’s a motivated buyer or seller and try to caputure a liquidity premium, its not as intense as you think.

Anyways Greenie. Go to a TAMP (Turnkey Asset Management Program) like Buckingham Asset Management, they split your fees with you but take care of your trading and operations, I guarantee you have no clue what your getting yourself into when it comes to starting one of these. Unless you already have $100m commited, your better off sharing your revenue with a TAMP and focus on growing your business or else you’ll have zero assets and be working on operations/compliance all day.