Updating a slice and dice portfolio with Multi-factor funds

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Updating a slice and dice portfolio with Multi-factor funds

Post by betablocker »

Has anyone taken a look at adapting a typical slice and dice portfolio based on the research showing that multi-style/factor funds are more efficient than buying factors through separate funds. I've been reading Larry's new book and playing with portfolio visualizer but I'm interested if anyone else has done research on the topic. The only thing I would have added to Larry's book are model portfolios like the one's in The Only Guide You'll Ever Need for a Winning Investment Strategy or in Rick Ferri's All About Asset Allocation.

p.s. I already understand all the comments people will have about keeping it simple with the 3 fund portfolio is and about reversion to the mean, etc.
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by larryswedroe »

beta
You can own a VERY simple two fund portfolio that is multi-factor and gets you more of a risk parity strategy
I use Bridgeway Omni Small Value for US and can use DFA World Ex-US for the international which includes EM.
There are obviously other funds that can be used to create tilted 2 or 3 fund portfolios
Theoretical
Posts: 1546
Joined: Tue Aug 19, 2014 10:09 pm

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by Theoretical »

Robert T did a nice update on his classic portfolio: viewtopic.php?t=184501 towards the RAFI/Russell Fundamental way of doing things.
SpaceCowboy
Posts: 1006
Joined: Sun Aug 12, 2012 12:35 am

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by SpaceCowboy »

AQR has a set of multi-factor funds that you can use. I own the small-cap one QSMLX. Unfortunately, it's performance has been underwhelming to date. VBR has outperformed it over my ownership period.
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by larryswedroe »

Re AQR multifactor funds. Here's basically the story with relatively poor performance. First, momentum underperformed value and so if benchmark is value index you get negative tracking error. Second, the way AQR ranks stocks means it tends to buy the stocks say that are more the "Bs" in value, momentum and profitability rather than the stocks that are "As" in one of the three as if were a component fund. Over time the Bs outperform the As, but in about 1/3 of years the As outperform. This is one of them. So you had two things working against the fund at same time. Both should be random events.
Larry
Topic Author
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by betablocker »

Wouldn't it be better to use the AQR multi style funds Domestic large and small and international and emerging large and small? That would increase your exposure to momentum and profitability rather than DFA and Bridgeway. You get more factor diversification that way I would think. I'll run it in portfolio visualizer.
vesalius
Posts: 794
Joined: Tue Jul 13, 2010 7:00 pm
Location: Texas

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by vesalius »

There are no AQR multistyle small intl or emerging variants as far as I can tell. Only large offered currently, well I think they may be total market style, but regardless they are both overwhelmingly dominated by large.

ETF variants: all the caveats about liquidity and trading apply. Many are quite small. Some are surprisingly low fee, especially the Goldman Sachs, but also the Deutsche X-trackers and John Hancock funds.

LV
LRGF iShares Edge MSCI Multifactor USA ETF
GSLC Goldman Sachs ActiveBeta U.S. Large Cap Equity ETF
JHML John Hancock Multifactor Large Cap ETF (DFA run)
DEUS Deutsche X-trackers Russell 1000 Comprehensive Factor ETF

SV
SMLF iShares Edge MSCI Multifactor USA Small-Cap ETF
DESC Deutsche X-trackersRussell 2000 Comprehensive Factor ETF

Intl LV
INTF iShares Edge MSCI Multifactor Intl ETF
GSIE Goldman Sachs ActiveBeta International Equity ETF
DEEF Deutsche X-trackers FTSE Developed Ex US Comprehensive Factor ETF

Intl SV
ISCF iShares Edge MSCI Multifactor Intl Small-Cap ETF

Emerging V
EMGF iShares Edge MSCI Multifactor Emerging Markets ETF
GEM Goldman Sachs ActiveBeta Emerging Markets Equity ETF
DEMG Deutsche X-trackers FTSE Emerging Comprehensive Factor ETF
SpaceCowboy
Posts: 1006
Joined: Sun Aug 12, 2012 12:35 am

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by SpaceCowboy »

Many of the multi-factor ETFs just have no liquidity. I really think it's important to set a threshold for Market Cap/AUM when looking at these ETFs.
I've used DLS for years as an International Small Cap Value fund, and have been surprised at its spreads and lack of liquidity despite a market cap of over $1 billion
User avatar
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by Robert T »

.
Three points worth making:

(1) Multifactor funds allow investors to potentially get greater combined factor tilts than a component approach. Value funds are multifactor funds in that they target more than one factor (with exposure to market + value), as are small value funds (with exposure to market + value + size).
  • For example: It’s not possible (or very difficult) to get as large a market, size, and value load of a small value fund with a combination of a large/midcap value fund and a small cap fund. Using DFA funds to illustrate – a 20:80 combination of DFA US large value:US Micro matches the size load of DFA small value (0.79) but only with about half the value load (0.35 vs. 0.67). https://www.portfoliovisualizer.com/fac ... ssetType=1 In this case, using small cap value allows for a greater combined portfolio tilt than the ‘component approach’ leading to higher expected return. https://www.portfoliovisualizer.com/bac ... tion3_2=80 . Note the higher returns are not from additional alpha from multifactor funds, but simply from greater factor exposure.

    The more factors you add the smaller the incremental benefit (given trade-offs, e.g. adding momentum lowers the value tilt) - as Fama earlier indicated. To illustrate – adding ‘profitability’ to the DFA smaller cap and value tilted portfolio only added 0.2-0.3 percent in additional returns, and slightly lowered volatility.

    1975-2012: Annualized return (%) / Standard Deviation
    DFA Core 2 without profitability 14.0 / 17.4
    DFA Core 2 with profitability 14.3 / 17.6

    DFA Vector without profitability 15.6 / 19.5
    DFA Vector with profitability 15.8 / 19.3

    DFA Targeted value without profitability 17.6 / 21.8
    DFA Targeted value with profitability 17.8 / 21.7

    These results are derived from the DFA Matrix Book 2013 (where the simulated series from 1975 excluded profitability sorts) and the DFA Matrix Book 2015 (where the simulated series from 1975 included profitability sorts). DFA Core 2 = Dimensional US Adjusted Market 2 Index, DFA Vector = Dimensional US Adjusted Market Value Index.

    I would just note that some academic studies suggest much larger incremental gains e.g. Novy-Marx’s data (averaging the top 3 decile sorts) over the same time period 1975-2012 suggest much larger incremental gains with annualized returns from ‘Value’ = 15.2 percent and for ValueProfit = 17.9 percent, a 2.7 percent difference. The DFA simulations (which show smaller return differences) are probably closer to what’s implementable. I think the same is likely true for any incremental improvements from multi-factor funds additionally including momentum.
(2) Choose the factors you want exposure to and your associated portfolio factor load targets. For some investors this may mean just beta, for others it may be size and value, for others it may include quality and momentum. Your choice should be based on some fundamental understanding of the underlying factors and likelihood of them persisting.

(3) Once you have done (2) then select funds to achieve your factor load targets, and assess the role multifactor funds can play in achieving your factor load targets. As Bernstein indicated in his book ‘Rationale Expectations”: “"...tilt is tilt, and these core funds have a small transactional advantage over a component approach. This is because fewer companies have to be bought and sold as they cross the borders between the older four-corner pigeonholes. This small advantage has to be weighed against the loss of control over the fine-tuning of value/small loadings and of the benefit of rebalancing with the four corners."
  • As illustrated in an earlier post:

    Jan. 2007 to August 2016

    DFA Component Portfolio
    21% DFA US Large Company
    11% DFA Tax-Managed US Marketwide Value
    68% DFA Tax-managed US Targeted Value

    100% DFA Vector

    Both portfolios had same market, size, and value load of 1.06, 0.45, and 0.29 over this period.

    Annualized returns (%) / SD / Sharpe Ratio
    DFA Component = 5.97% / 19.75 / 0.36
    DFA Vector = 5.83% / 19.63 / 0.36
So for this level of factor market, size, and value exposure there is little difference. If you want larger tilts to quality and momentum then perhaps multi-factor funds that also include these exposures provide some advantages (as per first point).

Robert
.
Topic Author
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by betablocker »

Robert T. I forgot to write back in November to thank you for this analysis. It was great and I've revisited it several times as I'm planning for some changes in my asset allocation. I think I'm comfortable now with using a modified component approach i.e. using DFA and Bridgeway to get some factor combination and screening but not going all in on multi-factor. I got some interesting results building all AQR multi factor portfolios in portfolio visualizer. In the end the funds are too new for me to make and large changes.

My biggest remaining concern is around how to incorporate momentum into a small and value tilted portfolio. As I understand it given their low correlation owning value and momentum in the same fund makes the most sense (avoiding selling and buying the same equities in two different funds) and that simply screening for momentum (as DFA does) doesn't really capture the factor. The simplest approach would be to just include something like MOM in my portfolio and give it some allocation. Are there alternatives that I should consider. Thanks
XBTC
Posts: 14
Joined: Sat Jan 21, 2017 6:50 am

Re: Updating a slice and dice portfolio with Multi-factor funds

Post by XBTC »

betablocker wrote: My biggest remaining concern is around how to incorporate momentum into a small and value tilted portfolio. As I understand it given their low correlation owning value and momentum in the same fund makes the most sense (avoiding selling and buying the same equities in two different funds) and that simply screening for momentum (as DFA does) doesn't really capture the factor. The simplest approach would be to just include something like MOM in my portfolio and give it some allocation. Are there alternatives that I should consider. Thanks
I am curious at what conclusion you arrived in the meantime? I am thinking about replacing the large cap portion of a SCV tilted portfolio with MTUM. I think this will avoid the problem of selling/buying of the same equities in different funds, as there is no intersection between a SCV fund and MTUM .

However I really am not sure about the following:

1) Ishares MTUM holds only 123 stocks. Is that enough for the whole large cap part of a portfolio? What risks could this cause?

2) Ishares MTUM has no constraints on sector allocation. So it could deviate from the S&P 500 arbitrarily which is kinda scary.

3) 1+2 would imply to maybe mix S&P 500 and MTUM for the large caps but then the exposure to size, value and momentum gets dilluted again...

Any thoughts on these points?
Post Reply