RAFI Funds
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
RAFI Funds
There has been some discussion of these funds so I thought people might find today's blog of interest.
http://moneywatch.bnet.com/investing/bl ... blog-river
http://moneywatch.bnet.com/investing/bl ... blog-river
- RaleighStClaire
- Posts: 733
- Joined: Mon Jun 18, 2007 2:08 pm
- Location: Tilting
Larry, this is interesting. I thought maybe the recent outperformance had more to do with the style drift of the RAFI indexes to tilt more towards growth when growth is outperforming.
How do other indexes that reconstitute quarterly compare with this natural value drift? Is it as expected? How often do the big indexes reconstitute anyways? MSCI, S&P, Russell?
How do other indexes that reconstitute quarterly compare with this natural value drift? Is it as expected? How often do the big indexes reconstitute anyways? MSCI, S&P, Russell?
Where's that red one gonna go?
- Random Musings
- Posts: 6756
- Joined: Thu Feb 22, 2007 3:24 pm
- Location: Pennsylvania
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
few thoughts
When growth outperforms value RAFI should be expected to better because they LOSE exposure to the underperforming asset class over the year. This is EXACTLY what happened during the period in question (was seasonal value premium, totally random of course).
When value is outperforming RAFI would be expected to underperform an index that reconstitutes daily (as DFA does) or monthly (like Bridgeway does). That is one reason we prefer the fund families of DFA and Bridgeway---more constant exposure to the asset class we want exposure to
FWIW, all the analysis we have done shows there is nothing special about the funds. IT is all exposure to loadings with small benefit from using multiple "screens" vs. one. Bridgeway uses four as well and we have carefully looked at their data and found some benefit as your pick up some diversification benefit as the four factors, while highly correlated (like all above .9) are not perfectly correlated.
Hope above is helpfu
When value is outperforming RAFI would be expected to underperform an index that reconstitutes daily (as DFA does) or monthly (like Bridgeway does). That is one reason we prefer the fund families of DFA and Bridgeway---more constant exposure to the asset class we want exposure to
FWIW, all the analysis we have done shows there is nothing special about the funds. IT is all exposure to loadings with small benefit from using multiple "screens" vs. one. Bridgeway uses four as well and we have carefully looked at their data and found some benefit as your pick up some diversification benefit as the four factors, while highly correlated (like all above .9) are not perfectly correlated.
Hope above is helpfu
Re: few thoughts
Larry, are you just referring to the Ultra-Small Market fund (BRSIX) when you mentioned Bridgeway, or are there other funds you like? Everything aside from BRSIX and their large-cap index appear to be actively managed with fairly high turnover. I know they managed institutional accounts, but I was under the impression they were just variations of their retail mutual funds (or perhaps a separate share class of the same fund).larryswedroe wrote:That is one reason we prefer the fund families of DFA and Bridgeway---more constant exposure to the asset class we want exposure to.
I'm just curious.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
scott
I was specifically referring to the new SV fund that is also a tax managed fund. It is on our recommended and approved list.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
calijim
First, momentum and value are negatively related. So yes the RAFI funds by only reconstituting annually do pick up some of that momentum effect, but you have the other side of that in that you are losing exposure the value effect
Second, DFA at least avoids the negative momentum side by not buying stocks that fall into the buy range until the negative momentum ceases. And they have the buy and HOLD ranges that help capture the positive momentum and for TM or TA funds they won't sell until have LT gains and have bit wider hold ranges.
Third, while FF doesn't explain the source of the risks premiums, the list of papers I have posted several times certainly makes attempt at explaining the sources of the premiums as risk stories.
BTW-there was not in any way to be a negative post on RAFI funds, but to show how serendipity can play a role in returns and you need as investor to fully understand WHY a fund is performing the way it is, so you know if the "success" is random or by design. In this case the "success" was at least IMO clearly a random outcome as there is not evidence that I am aware of that there is seasonality in the value premium. Though there was in this period.
Fourth, I don't buy the argument at all that the most overpriced stocks have to be the large ones. Clearly that can be wrong. Why cannot the small stocks be the most mispriced? In fact we have great evidence that it can be the case as we had a small cap bubble in the early 80s.
Hope that is helpful
Second, DFA at least avoids the negative momentum side by not buying stocks that fall into the buy range until the negative momentum ceases. And they have the buy and HOLD ranges that help capture the positive momentum and for TM or TA funds they won't sell until have LT gains and have bit wider hold ranges.
Third, while FF doesn't explain the source of the risks premiums, the list of papers I have posted several times certainly makes attempt at explaining the sources of the premiums as risk stories.
BTW-there was not in any way to be a negative post on RAFI funds, but to show how serendipity can play a role in returns and you need as investor to fully understand WHY a fund is performing the way it is, so you know if the "success" is random or by design. In this case the "success" was at least IMO clearly a random outcome as there is not evidence that I am aware of that there is seasonality in the value premium. Though there was in this period.
Fourth, I don't buy the argument at all that the most overpriced stocks have to be the large ones. Clearly that can be wrong. Why cannot the small stocks be the most mispriced? In fact we have great evidence that it can be the case as we had a small cap bubble in the early 80s.
Hope that is helpful
Re: scott
[BOSVX] OMNI TAX-MANAGED SMALL-CAP VALUE FUND is this the fund you are referring to?larryswedroe wrote:I was specifically referring to the new SV fund that is also a tax managed fund. It is on our recommended and approved list.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
bridgeway
Yes, that's the one.
Re: scott
Appears this fund has more in common with the DFA fund than just frequency of rebalancing. From prospectus:vesalius wrote:[BOSVX] OMNI TAX-MANAGED SMALL-CAP VALUE FUND is this the fund you are referring to?larryswedroe wrote:I was specifically referring to the new SV fund that is also a tax managed fund. It is on our recommended and approved list.
Purchase and Sale of Fund Shares: All investments are subject to approval by the Adviser. The Fund is generally available for investment only by institutional clients, clients of approved registered investment advisors, clients of financial institutions and a limited number of certain other investors as approved by the Adviser.
-
- Posts: 266
- Joined: Wed Aug 18, 2010 12:42 pm
Larry: good article. My research actually is simpler than yours: the 3 quarters starting on January of 2009 explain all of the outperformance of the RAFI strategies (I looked primarily at the US 1000) over other index funds. These 9 months consisted of the worst performance for the momentum factor since 1931 (momentum was -56% for the 9 months ending in September of 2009). Because the RAFI annual reconstution process plows headlong into stocks that have fallen in price (i.e. have a lower price relative to their "economic footprint"), a bout of negative momentum did/will significantly benefit the funds relative to indexes (like Russell/Vanguard) without negative momentum exposure.
As a matter of fact, from a regression standpoint since '06, the Powershares US Large Cap Fundamental Index ETF looks almost identical to the Russell 3000 Value index in terms of its beta, size, and value exposure. The main difference: momentum. While the Russell 3000 Value index has 0% exposure to momentum, the RAFI 1000 has a statistically significant -0.16% per month momentum exposure.
How does this impact returns? From 1927-2010, momentum had an annualized positive return of +7% per year. An index with -0.16 exposure to momentum should be expected to underperform their size/value factors by over 1.1% per year.
Put another way, the RAFI mousetrap has to be so adept at identifying mispriced securities not detected by size/value exposure that it can overcome a -1% per year expected loss due to mechanical rebalancing. That is a pretty big hurdle in a largely efficient market.
Of course Schwab could simply deviate from RAFI's buy/sell instructions for a few months to avoid negative momentum (at the cost of tracking error), but they aren't likely to do that based on their rock bottom expense structure.
To look at this a bit more closely, consider the following index/ETF total returns over 3 distinct periods. As you can see, all of the RAFI outperformance came from this severe bout of negative momentum. Over time we'd expect momentum to be positive and penalize the RAFI strategies:
As a matter of fact, from a regression standpoint since '06, the Powershares US Large Cap Fundamental Index ETF looks almost identical to the Russell 3000 Value index in terms of its beta, size, and value exposure. The main difference: momentum. While the Russell 3000 Value index has 0% exposure to momentum, the RAFI 1000 has a statistically significant -0.16% per month momentum exposure.
How does this impact returns? From 1927-2010, momentum had an annualized positive return of +7% per year. An index with -0.16 exposure to momentum should be expected to underperform their size/value factors by over 1.1% per year.
Put another way, the RAFI mousetrap has to be so adept at identifying mispriced securities not detected by size/value exposure that it can overcome a -1% per year expected loss due to mechanical rebalancing. That is a pretty big hurdle in a largely efficient market.
Of course Schwab could simply deviate from RAFI's buy/sell instructions for a few months to avoid negative momentum (at the cost of tracking error), but they aren't likely to do that based on their rock bottom expense structure.
To look at this a bit more closely, consider the following index/ETF total returns over 3 distinct periods. As you can see, all of the RAFI outperformance came from this severe bout of negative momentum. Over time we'd expect momentum to be positive and penalize the RAFI strategies:
Code: Select all
1/06 - 12/08
RAFI 1000 -27.3%
Russell 3000 Value = -22.8%
Momentum = +35.9%
Code: Select all
1/09 - 9/09
RAFI 1000 +39.7%
Russell 3000 Value = +15.0%
Momentum = -55.5%
Code: Select all
10/09 - 12/10
RAFI 1000 = +21.4%
Russell 3000 Value = +21.1%
Momentum = +11.6%
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
Sammy
The fund also has things in common with RAFI, in that it uses four fundamental factors instead of one as does DFA. But like DFA Bridgeway value fund is TM, unlike RAFI
Too bad both the DFA small value and Bridgeway small value seem to require a small individual investor to give up about 1% to an advisor to gain admission. RAFI still looks like a reasonable or near equivalent alternative under those circumstances.
I wish these funds would just slap significant early withdrawal fees and penalties on individual investors as a way to stop frequent traders and hot money, but still allow buy and hold types access.
I wish these funds would just slap significant early withdrawal fees and penalties on individual investors as a way to stop frequent traders and hot money, but still allow buy and hold types access.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
vesalius
The problem with those penalties (which in theory are good ideas) are that it makes it more difficult/expensive to rebalance or TLH for other investors, imposing penalties on them, so it doesn't work in practice.
Now ETF versions would work, but they really only work with pure indices, not ones that do patient/block trading strategies or TM strategies.
Now ETF versions would work, but they really only work with pure indices, not ones that do patient/block trading strategies or TM strategies.
Re: vesalius
ETFs should work well as a vehicle for any passive fund, and the more passive, the better. As long as a position change can be broken up over time (to be less predictable), front-running should be minimized.larryswedroe wrote:Now ETF versions would work, but they really only work with pure indices, not ones that do patient/block trading strategies or TM strategies.
Additionally, ETFs exist for actively managed funds, and also for less active non-market-cap index funds such as RAFI. If anything, public-index funds are less suited for non-whole-market ETFs (e.g., a small cap index or value index), since their portfolio changes are more predictable.
- RaleighStClaire
- Posts: 733
- Joined: Mon Jun 18, 2007 2:08 pm
- Location: Tilting
Larry -
How can DFA and Bridgeway reconstitute so frequently w/o incurring a much higher operating expense ratio? Wouldn't this be a drawback compared to funds that only reconstitute annually?
How can DFA and Bridgeway reconstitute so frequently w/o incurring a much higher operating expense ratio? Wouldn't this be a drawback compared to funds that only reconstitute annually?
"Life can only be understood backward; but it must be lived forward." ~ Søren Kierkegaard |
|
"You can't connect the dots looking forward; but only by looking backwards." ~ Steve Jobs
- RaleighStClaire
- Posts: 733
- Joined: Mon Jun 18, 2007 2:08 pm
- Location: Tilting
Not indexes. They use new money to go into the areas where the fund needs to invest.Lbill wrote:Larry -
How can DFA and Bridgeway reconstitute so frequently w/o incurring a much higher operating expense ratio? Wouldn't this be a drawback compared to funds that only reconstitute annually?
Where's that red one gonna go?
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
few thoughts
As I said ETFs work BEST with pure indices. How can an ETF replicate something whose holdings change due to opportunistic trading or TM strategies?
DFA and Bridgeway reconstitute daily but also have buy and HOLD ranges. They use cash flows and divs to rebalance the fund. So turnover is kept down.
DFA and Bridgeway reconstitute daily but also have buy and HOLD ranges. They use cash flows and divs to rebalance the fund. So turnover is kept down.
BOSVX has ER of 1.0%. By comparison, IJS has ER of 0.25% and VBR of 0.14%. Seems a bit expensive to me. What about RZV with an ER of 0.35%?
"Life can only be understood backward; but it must be lived forward." ~ Søren Kierkegaard |
|
"You can't connect the dots looking forward; but only by looking backwards." ~ Steve Jobs
Re: few thoughts
Well, actually:larryswedroe wrote:As I said ETFs work BEST with pure indices.
But anyways,Now ETF versions would work, but they really only work with pure indices...
It depends on how opportunistic the fund is. If it's about picking up new opportunities intraday, yes, it might not be the most optimal solution. However, I'm not sure why a fund couldn't simply do opportunistic trading and update the basket for the next day. I presume the opportunistic trading wouldn't be a substantial part of the account, so it shouldn't cause the fund to deviate much from the basket for the day.How can an ETF replicate something whose holdings change due to opportunistic trading or TM strategies?
With regards to TM, I don't see why it wouldn't work well for TM strategies, assuming the goal of TM is to minimize capital gains and/or dividends. If the fund has enough trading, capital gains would be minimized by redemptions. VTI and VWO haven't had capital gains for many years. Dividends could be avoided, of course by under-weighting high dividend stocks.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
ftrobin
ETFs must exactly match the fund to allow for creation and redemption--and they trade all day.Simply cannot do that with fund that doesn't index. Block trading goes on all day every day and TM also involves drifting from the construction rules. I just don't see how it can work.
BTW-DFA is a profit making institution and IMO if they thought issuing ETF versions would generate more profits I would assume they would do it. I believe they see the issues above as problems
Now you could set up an ETF that would follow "index" rules, using the definitions that DFA uses. But don't see how it could run the same way with daily reconstruction of buy and hold ranges and the other issues above (but I am not an expert on ETFs, perhaps Rick Ferri can address it). The ETFs I am familiar with all have fixed rules and annual reconstitution so we know what is in them all the time
BTW-DFA is a profit making institution and IMO if they thought issuing ETF versions would generate more profits I would assume they would do it. I believe they see the issues above as problems
Now you could set up an ETF that would follow "index" rules, using the definitions that DFA uses. But don't see how it could run the same way with daily reconstruction of buy and hold ranges and the other issues above (but I am not an expert on ETFs, perhaps Rick Ferri can address it). The ETFs I am familiar with all have fixed rules and annual reconstitution so we know what is in them all the time
Re: ftrobin
I believe that ETFs can employ a statistical sampling for a creation unit -- they don't require that unit match perfectly. For example, on February 15th VTI only 1355 stocks were required for a creation unit. The underlying fund, however, has 3380 stocks. I'm sure the result is similar for BND and VSS with their thousands of illiquid issues. I remember reading a Vanguard interview on the topic, but can't dig it up through Google at the moment. I know we download basket requirements each day from ETF providers, so I'm pretty sure the creation unit rules can change from day to day.larryswedroe wrote:ETFs must exactly match the fund to allow for creation and redemption--and they trade all day.
There is a competing interest in that DFA is able to generate scarcity and intrigue for their product by making them solely accessible via advisers or possibly 401ks. I don't know how much this factor should be weighted, but it should be considered.BTW-DFA is a profit making institution and IMO if they thought issuing ETF versions would generate more profits I would assume they would do it. I believe they see the issues above as problems.
Don't forget about actively managed ETFs.The ETFs I am familiar with all have fixed rules. and annual reconstitution so we know what is in them all the time
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
ftrobin
The quote scarcity value is really only a value for advisors, not DFA which would make far more money if they opened it up to more people including the public, which they are doing in many ways already. So that issue is one I personally don't buy, but that is just my opinion.
If there really is not an operational issue perhaps we will see them. Keep in mind that the block trading and TM strategies add value though and they can hold anywhere from 1/2 to 2x the market cap weighting of an asset, so sampling really doesn't seem to cut it, but perhaps I am wrong.
If there really is not an operational issue perhaps we will see them. Keep in mind that the block trading and TM strategies add value though and they can hold anywhere from 1/2 to 2x the market cap weighting of an asset, so sampling really doesn't seem to cut it, but perhaps I am wrong.
-
- Posts: 14
- Joined: Wed Mar 02, 2011 8:38 am
Re: scott
Larry - is this recommended and approved list posted somewhere for all to view?larryswedroe wrote:I was specifically referring to the new SV fund that is also a tax managed fund. It is on our recommended and approved list.
-
- Posts: 16022
- Joined: Thu Feb 22, 2007 7:28 am
- Location: St Louis MO
rob
Sorry but it is not. It is a list we use for our clients and also as a RECOMMENDED approved list for the about 125 RIA firms we have strategic relationships with, providing them with investment advice and other services/
-
- Posts: 14
- Joined: Wed Mar 02, 2011 8:38 am
It was nice to see this thread pop up. The article was a good read. The five Schwab RAFI funds are where I am moving our holdings. We have several accounts for us and the kids at Schwab plus I can purchase Schwab funds for a $19.95 transaction fee from my Hewitt 401K self directed brokerage account. These funds end up matching well with my regular 401K contribution choices, as I will still be contributing funds but will not be able to afford the transaction fees.
BTW, now if only someone could explain what the heck one buys when one purchases the ETF RALS. I know what long is. I know what short is. But I don't understand the mechanism of constructing a set of weights mixing the two together.
Also, I now refer to the RAFIs as RAEFWS. (Economic Footprint Weighting Strategy - ). But I am still happy to be moving to them. I am about a third of the way there.
David
BTW, now if only someone could explain what the heck one buys when one purchases the ETF RALS. I know what long is. I know what short is. But I don't understand the mechanism of constructing a set of weights mixing the two together.
Also, I now refer to the RAFIs as RAEFWS. (Economic Footprint Weighting Strategy - ). But I am still happy to be moving to them. I am about a third of the way there.
David