Diversifying SCV with MTUM: yea or nay?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 9:59 am
Location: Hic sunt dracones

Diversifying SCV with MTUM: yea or nay?

Post by CULater »

Based on my reading re: factor investing (especially Larry's new book), I'd like to add SCV to my portfolio and am looking at IJS and VIOV for that purpose. According to Larry and others, momentum tends to be negatively correlated to value so that implies also adding a MOM asset. I'm hoping that such a combo might soften the inevitable drawdown risk of SCV somewhat and improve risk-adjusted returns. Not much out there except MTUM. Is a combo of SCV (IJS, VIOV) with MTUM worth undertaking? There isn't much historical data for MTUM upon which to base a decision using Portfolio Visualizer. Wondering if other factor investors have pondered the same sort of combination of SCV and MOM and what some good options might be. Not able to invest in DFA and am looking for the options available to the average investor like me.
On the internet, nobody knows you're a dog.
User avatar
GreatOdinsRaven
Posts: 572
Joined: Thu Apr 23, 2015 8:47 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by GreatOdinsRaven »

I do. Sure there are probably better ways to do it. This seemed good enough to me.
"The greatest enemies of the equity investor are expenses and emotions." -John C. Bogle, Little Book of Common Sense Investing. | | "Winter is coming." Lord Eddard Stark.
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by lack_ey »

It's a reasonable choice if that's what you want to do. Some caveats with using iShares Edge MSCI USA Momentum Factor ETF (MTUM): Momentum is historically not as large of an effect in large caps, and some of the pretty good performance we've seen from MTUM since inception is just the negative value (growth) bias. So far MTUM has actually subtracted more value exposure than added momentum exposure—or at least it's pretty close—based on the factor regression.

There are also some multifactor value/momentum (sometimes more) funds available, if that's what you actually want. Fewer in the small cap space but at least there's iShares Edge MSCI Multifactor USA Small-Cap ETF (SMLF). Some mutual funds too. AQR Small Cap Multi-Style is available in N shares (QSMNX) at some brokerages for small minimums but I wouldn't pay that kind of fees for access to that pretty vanilla strategy.

Also in a tax-advantaged account, you might consider a couple of Vanguard's quant team funds, Strategic Equity and Strategic Small-Cap Equity. They have some value and momentum exposure in mid and small caps. I took a look at them here:
viewtopic.php?f=10&t=206890&newpost=3173401
printer
Posts: 124
Joined: Sun Dec 14, 2014 11:37 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by printer »

IJS and VIOV have momentum loading near zero, according to portfoliovisualizer.com with its default settings.

My meager understanding: some funds make an attempt to be momentum-neutral even when momentum is not the primary aim of the fund. This happens with some value funds for example - if something is cheap, then maybe it became cheap recently, i.e., has negative momentum loading. So the fund may take steps to counteract this. How they do it, exactly, I don't know. It seems better to do it for individual stocks rather than with the fund in aggregate.

The upshot is that you might easily decide that even if you are concerned about momentum, you don't need to do anything other than find a fund that does this and is stable enough that you're confident that it'll continue doing it.

You might review the idea "momentum tends to be negatively correlated to value so that implies also adding a MOM asset" and see whether you think it is logically correct.
User avatar
David Jay
Posts: 14586
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Diversifying SCV with MTUM: yea or nay?

Post by David Jay »

CULater wrote:According to Larry and others, momentum tends to be negatively correlated to value...
Are you sure Larry said they are negatively correlated? I can understand low correlation, but negative correlation is a different animal.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Diversifying SCV with MTUM: yea or nay?

Post by larryswedroe »

The best way to accomplish this IMO is to use value funds that screen out for negative momentum stocks, and also are more patient in selling high MOM stocks (giving priority in the algorithms to the ones that don't have as strong a MOM signal). That way you don't have the issue of one fund buying a stock that your other fund is selling. Now it won't get you positive MOM exposure. AQR's multi-style fund on other hand uses a ranking system to give + loadings on value, MOM and profitability. That avoids the problem I just mentioned. Note that will tend to lower value exposure in return for MOM exposure--diversifying your sources of risk which IMO is a good idea, and also for those who it matters to reduces the risk of that dreaded disease of tracking error regret which can cause those subject to it to abandon their plans.

Hope that helps
Larry
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by betablocker »

I'm thinking through this exact issue. Sounds like mtum just dilutes value rather than gaining momentum. It also seems that if you use a multi factor fund that you would have to adjust your other scv holdings. As a multi factor the new fund would add additional value exposure along with the momentum exposure wouldn't it?
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by betablocker »

larryswedroe wrote:The best way to accomplish this IMO is to use value funds that screen out for negative momentum stocks, and also are more patient in selling high MOM stocks (giving priority in the algorithms to the ones that don't have as strong a MOM signal). That way you don't have the issue of one fund buying a stock that your other fund is selling. Now it won't get you positive MOM exposure. AQR's multi-style fund on other hand uses a ranking system to give + loadings on value, MOM and profitability. That avoids the problem I just mentioned. Note that will tend to lower value exposure in return for MOM exposure--diversifying your sources of risk which IMO is a good idea, and also for those who it matters to reduces the risk of that dreaded disease of tracking error regret which can cause those subject to it to abandon their plans.

Hope that helps
Larry
If you add a multi style fund, then you'd need to readjust your other scv holdings using a tool like portfolio visualizer to make sure you still have the appropriate value loading?
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Diversifying SCV with MTUM: yea or nay?

Post by larryswedroe »

beta
yes you should look at the total portfolio's loadings.
Larry
betablocker
Posts: 522
Joined: Mon Jan 11, 2016 12:26 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by betablocker »

At some point shouldn't we just all own multi factor funds exclusively?
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by lack_ey »

David Jay wrote:
CULater wrote:According to Larry and others, momentum tends to be negatively correlated to value...
Are you sure Larry said they are negatively correlated? I can understand low correlation, but negative correlation is a different animal.
Yes, the data is pretty clear. See for example "Value and Momentum Everywhere."

Note that this is between momentum and value factors. If you're talking long-only momentum and value funds, then of course they're significantly positively correlated as both primarily have market beta exposure and this overcomes the negative correlation (generally) between the smaller momentum and value loads.
wije
Posts: 220
Joined: Sun Jun 07, 2015 8:35 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by wije »

David Jay wrote:
CULater wrote:According to Larry and others, momentum tends to be negatively correlated to value...
Are you sure Larry said they are negatively correlated? I can understand low correlation, but negative correlation is a different animal.
On portfoliovisualizer.com, MTUM and IJS have .72 correlation. Although that isn't as high as the correlation between growth and value stocks, it still seems too high to serve a strong diversification function IMHO.
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Diversifying SCV with MTUM: yea or nay?

Post by larryswedroe »

wije
The logic is pretty simple for the NEGATIVE correlation. The way you get to be value is to have POOR relative performance and the way you get to be growth is to have GOOD relative performance.

The following is from the table on correlations in Your Complete Guide to Factor Based Investing
Factor Market Beta Size Value Momentum Profitability Quality
Market beta 1.00 0.29 –0.27 –0.17 –0.27 –0.52
Size 0.29 1.00 0.01 –0.12 –0.22 –0.53
Value –0.27 0.01 1.00 –0.20 0.09 0.04
Momentum –0.17 –0.12 –0.20 1.00 0.08 0.30
Profitability –0.27 –0.22 0.09 0.08 1.00 0.74
Quality –0.52 –0.53 0.04 0.30 0.74 1.00

So when value does well MOM TENDS TO do poorly, obviously the correlation isn't -1, and vice versa.

Best wishes
larry
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by matto »

wije wrote:
David Jay wrote:
CULater wrote:According to Larry and others, momentum tends to be negatively correlated to value...
Are you sure Larry said they are negatively correlated? I can understand low correlation, but negative correlation is a different animal.
On portfoliovisualizer.com, MTUM and IJS have .72 correlation. Although that isn't as high as the correlation between growth and value stocks, it still seems too high to serve a strong diversification function IMHO.
That's because MTUM and IJS are both long only.

MTUM = MOM + BETA
IJS = VAL + BETA

0.72 COR(MTUM, IJS)
0.72 = COR(MOM + BETA, VAL + BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + COR(BETA, BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + 1
-.28 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA)
User avatar
David Jay
Posts: 14586
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Diversifying SCV with MTUM: yea or nay?

Post by David Jay »

lack_ey wrote:
David Jay wrote:
CULater wrote:According to Larry and others, momentum tends to be negatively correlated to value...
Are you sure Larry said they are negatively correlated? I can understand low correlation, but negative correlation is a different animal.
Yes, the data is pretty clear. See for example "Value and Momentum Everywhere."

Note that this is between momentum and value factors. If you're talking long-only momentum and value funds, then of course they're significantly positively correlated as both primarily have market beta exposure and this overcomes the negative correlation (generally) between the smaller momentum and value loads.
Yup, I was thinking of (long) funds. I see the difference.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by lack_ey »

matto wrote:That's because MTUM and IJS are both long only.

MTUM = MOM + BETA
IJS = VAL + BETA

0.72 COR(MTUM, IJS)
0.72 = COR(MOM + BETA, VAL + BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + COR(BETA, BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + 1
-.28 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA)
You're missing some coefficients there. MTUM and IJS don't have momentum and value loadings of 1. They also aren't purely those things plus beta.

Also better do the calcs in terms of covariance and make sure everything is scaled by the proper sqrt(var1 * var2).


larryswedroe wrote:The following is from the table on correlations in Your Complete Guide to Factor Based Investing
table was a bit hard to read so here:

Code: Select all

Factor         Market Beta     Size      Value     Momentum Profitability Quality
Market beta         1.00       0.29      –0.27      –0.17       –0.27      –0.52
Size                0.29       1.00       0.01      –0.12       –0.22      –0.53
Value              –0.27       0.01       1.00      –0.20        0.09       0.04
Momentum           –0.17      –0.12      –0.20       1.00        0.08       0.30
Profitability      –0.27      –0.20       0.09       0.08        1.00       0.74
Quality            –0.52      –0.30       0.04       0.30        0.74       1.00
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by matto »

lack_ey wrote:
matto wrote:That's because MTUM and IJS are both long only.

MTUM = MOM + BETA
IJS = VAL + BETA

0.72 COR(MTUM, IJS)
0.72 = COR(MOM + BETA, VAL + BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + COR(BETA, BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + 1
-.28 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA)
You're missing some coefficients there. MTUM and IJS don't have momentum and value loadings of 1. They also aren't purely those things plus beta.

Also better do the calcs in terms of covariance and make sure everything is scaled by the proper sqrt(var1 * var2).
I'm missing plenty of things, but adding more precision won't change the result :)
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by lack_ey »

matto wrote:
lack_ey wrote:
matto wrote:That's because MTUM and IJS are both long only.

MTUM = MOM + BETA
IJS = VAL + BETA

0.72 COR(MTUM, IJS)
0.72 = COR(MOM + BETA, VAL + BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + COR(BETA, BETA)
0.72 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA) + 1
-.28 = COR(MOM, VAL) + COR(MOM, BETA) + COR(VAL, BETA)
You're missing some coefficients there. MTUM and IJS don't have momentum and value loadings of 1. They also aren't purely those things plus beta.

Also better do the calcs in terms of covariance and make sure everything is scaled by the proper sqrt(var1 * var2).
I'm missing plenty of things, but adding more precision won't change the result :)

Yeah, I just got freaked out a bit seeing equals signs without disclaimers when so much is not accounted for and the equations are not meant to be taken literally. :shock:

Somebody might actually think that's precise otherwise.
garlandwhizzer
Posts: 3565
Joined: Fri Aug 06, 2010 3:42 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by garlandwhizzer »

My two cents worth. Back to the original question is SCV + the MTUM (large cap long only) MOM fund a good pairing? If you are a firm believer in factor investing and have no interest in plain beta I think one can make a case for this approach. MOM looks great on backtesting, robust premiums, but the problem is that that high turnover is necessary to harvest it, that is to say increased costs and friction. This is not so expensive to do in the large cap long only space where MTUM plays, but it becomes very expensive in the relatively illiquid small cap space especially if you use a long/short approach which is entirely impractical on a after cost basis in SC space. That, I believe, is why so few funds seek MOM in the SC space with long only and none that I know of that seek it in a long/short strategy in the SC space. Much of the MOM premium comes from shorting negative MOM SC stocks which often are circling the drain and none of this is harvestable on an after cost basis. The MOM premium is negatively correlated to value making it attractive theoretically but when you look at how much of that premium can be harvested on an after cost basis, most funds choose to forego shorting (too expensive) which cuts out at least half the premium, and also to forego small caps (too expensive with a high turnover strategy even in long only). Hence you're down to 25% of the premium that is maximally possible to harvest and then you have to subtract how much of MOM exposure subtracts from your value premium in a strictly multi factor approach. The more stringent your criteria for harvesting MOM, the more loaded you are with MOM, aiming for the sky with 25% of the full premium as a ceiling, the more it subtracts value exposure. In order to construct a multi-factor portfolio where MOM doesn't destroy value you have to go luke-warm into both rather than concentrate in deep value and strong MOM which is self annihilating. So ultimately if you do it well which MTUM does, you're looking at a very small fraction of that theoretical MOM premium.

So after costs, if you harvest any alpha at all above beta you're doing well. MTUM does this better than most but since its inception more than 3 years ago it has outperformed VUG, the Vanguard Growth ETF (which is an appropriate comparison because it plays in the same LCG sandbox) by a whisker and during much of that time VUG outperformed it. The difference between these two funds is a statistical wash for the full life of MTUM. This is why in my view there aren't many MOM funds, especially so in the MC and SC space, and a reason why factor research looks so promising on academic backtesting but there remains a wide gulf between these rosy results and the after cost returns of the funds and efts that seek to exploit it. Some factor funds perform better than others, but of course in any large group of results in nature there tends to be a random distribution, a Gaussian curve with left tails and right tails. If a strategy is sound we must compare not the extreme outlier in MOM funds, but the mean MOM fund result to verify whether MOM or multi-factor funds in general as a strategy produce consistent outperformance on an after cost basis. I am not aware that this has to date been established. If so, please let me know with after cost average real fund results.

Garland Whizzer
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Diversifying SCV with MTUM: yea or nay?

Post by larryswedroe »

Garland that's good analysis. I would add this, that the higher costs in small is why pure indexing approaches don't work well, you have to be a patient, algorithmic trader who lives with random tracking error.
That's true for many of the anomalies that have high turnover as most of them are in the small cap space due to limits to arbitrage
Larry
User avatar
nedsaid
Posts: 19275
Joined: Fri Nov 23, 2012 11:33 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by nedsaid »

larryswedroe wrote:Garland that's good analysis. I would add this, that the higher costs in small is why pure indexing approaches don't work well, you have to be a patient, algorithmic trader who lives with random tracking error.
That's true for many of the anomalies that have high turnover as most of them are in the small cap space due to limits to arbitrage
Larry
Wow, Garland. I second Larry that this was good analysis.

Pretty much what you hit on is that Growth and Value take turns outperforming each other. I suppose you would get a diversification benefit with a Large Growth vs a Small Value strategy, one zigs while the other zags. It sounds to me that the Vanguard Growth ETF is a good proxy for the iShares Momentum Factor ETF. At least in the large cap space, Growth and Momentum seem to be about equal.

It is interesting that Paul Merriman and Rick Ferri both advised investors to pair Large Value with the S&P 500 and Small Value with the Small Value Index. I think they felt that a pairing of Value vs. Growth in both the Small and Large-Cap space would mostly cancel out any factor effect you were trying to achieve. Garland made similar comments.

It is weird because when I read Larry's article on Factor Investing in ETF Magazine, momentum had better performance and persistence than Value. My guess is that most of this excess performance of momentum over value is in the Small-Cap space which is hard for investors to capture. As Garland said, the momentum effect was diluted in the Large Cap space, diluted enough that you may as well go with a good growth fund. So it makes sense for the small investor to focus on the Small and Value premiums and not get too fancy with momentum.
A fool and his money are good for business.
Theoretical
Posts: 1546
Joined: Tue Aug 19, 2014 10:09 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by Theoretical »

I think if you were going to go all out for momentum, you'd do something like a Beta+quality+momentum strategy in the mid-large domains with maybe a smalller slice of a MOM neutral SCV like IJS or the DFA funds to lessen the negative Small and value loads.

Momentum and AQR's (sector neutral) definition of low beta would be interesting to target.
larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Re: Diversifying SCV with MTUM: yea or nay?

Post by larryswedroe »

It's important to understand, which is why we have an appendix in Your Complete Guide to Factor Based Investing, that there is diminishing marginal utility of adding factors once you get beyond size and value. Has to be that way as beta, size and value explain more than 90% of the differences in returns of diversified portfolios. As you add other factors the exposure to the new factor you gain likely is coming from another factor. Like adding MOM will attentuate the exposure to value. Now that doesn't mean there isn't value in further diversification, especially for factors with negative correlation. That helps reduce volatility and also tracking error risk.
Larry
DaufuskieNate
Posts: 605
Joined: Wed May 28, 2014 11:53 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by DaufuskieNate »

While we are talking about diversifying with momentum, it's not solely a choice between small cap and large cap equities. Pervasive factors "work" in a variety of asset classes. AQR has funds which seek to harvest momentum in commodities, currencies and fixed income in addition to equities. The Managed Futures funds focus on time-series momentum. The Style Premia funds focus on cross-sectional momentum.
garlandwhizzer
Posts: 3565
Joined: Fri Aug 06, 2010 3:42 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by garlandwhizzer »

Just my two cents worth again. When it comes to factors I think MOM is robust theoretically but very hard to harvest in an after cost fashion. If you're a factor investor I believe that value particularly in the SC space is a more harvestable alternative for long term investors. It, too, has its challenges. Constructing a SCV portfolio such that you harvest the small plus value premiums without overcoming those gains with strong negative MOM and negative alpha seems to me something of an art. To be successful you have to dilute pure deep value selectively by getting rid of junk, screening for negative profitability, negative alpha, and for negative MOM. Tracking error and cost can pose challenges. It's a tough tightrope to walk particularly as more and more trying to walk it. Ultimately I believe that for both behavioral and risk reasons, value, if properly done which many SCV funds don't do, will likely provide modestly improved results for those with sufficient patience to hold on through its sometimes long periods of underperformance (which is likely a minority of investors). Simply ruling out lottery type SCG stocks which value screens do reliably would be expected in the long run to improve results over SCB for long term buy and hold investors.

Garland Whizzer
User avatar
nedsaid
Posts: 19275
Joined: Fri Nov 23, 2012 11:33 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by nedsaid »

garlandwhizzer wrote:Just my two cents worth again. When it comes to factors I think MOM is robust theoretically but very hard to harvest in an after cost fashion. If you're a factor investor I believe that value particularly in the SC space is a more harvestable alternative for long term investors. It, too, has its challenges. Constructing a SCV portfolio such that you harvest the small plus value premiums without overcoming those gains with strong negative MOM and negative alpha seems to me something of an art. To be successful you have to dilute pure deep value selectively by getting rid of junk, screening for negative profitability, negative alpha, and for negative MOM. Tracking error and cost can pose challenges. It's a tough tightrope to walk particularly as more and more trying to walk it. Ultimately I believe that for both behavioral and risk reasons, value, if properly done which many SCV funds don't do, will likely provide modestly improved results for those with sufficient patience to hold on through its sometimes long periods of underperformance (which is likely a minority of investors). Simply ruling out lottery type SCG stocks which value screens do reliably would be expected in the long run to improve results over SCB for long term buy and hold investors.

Garland Whizzer
Garland, what you are describing, tightrope and all, sounds a lot like active management. The more things you screen for, the smaller the universe of stocks, and the more the process looks active.
A fool and his money are good for business.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by stlutz »

Since inception, MTUM has trailed its benchmark by an average of .21% per year. It has an ER of .15%. So, it's only losing .06% per year to other factors. I personally wouldn't find that unacceptable if I wanted a momentum fund in my portfolio. I don't think any others have real life holding costs that low.

As a rules-based fund, if MTUM continues to get larger it does face issues of front-running. That would not show up in the tracking vs. the index, but is a potential additional cost. As a mid-to-large cap fund this cost should not be high, but it is something to be aware of. I have seen an analysis which claims that USMV (the minimum variance ETF) does lose several basis points of return each year to front-running.

Of all the factors, momentum is the most robust--both in terms of performance an in terms of issues with underlying data. If one really is into all of the factor stuff, it seems worth including.

If I use a mix of 75% IJS and 25% MTUM, I get non-zero loading on all of the factors, so they aren't really cancelling each other out: https://www.portfoliovisualizer.com/fac ... tion2_1=75
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Diversifying SCV with MTUM: yea or nay?

Post by stlutz »

One other note, if I do a comparative factor analysis of VUG vs. MTUM, their loadings are substantially different from each other. https://www.portfoliovisualizer.com/fac ... ssetType=1

If momentum is what one really wants, then buy momentum stocks as opposed to a fund that has a fair number of momentum stocks in it among other things.
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Diversifying SCV with MTUM: yea or nay?

Post by lack_ey »

stlutz wrote:If I use a mix of 75% IJS and 25% MTUM, I get non-zero loading on all of the factors, so they aren't really cancelling each other out: https://www.portfoliovisualizer.com/fac ... tion2_1=75
Okay, so over that period,

Code: Select all

Portfolio              Mkt    Size   Value Momentum Quality Alpha (annual)
75% IJS, 25% MTUM     1.05    0.75    0.20    0.12    0.36   -0.16%
100% IJS              1.07    1.08    0.30    0.07    0.50   -0.58%
Is that actually an improvement (ignoring the alpha, assuming that's at least mostly random)?

That said, it may be better to use the daily regressions (not available there for portfolios) given that 3 years doesn't provide a lot of samples to do regressions on monthly data and the above estimates may be significantly off.

Also, as I noted in my thread I'm yet again shamelessly linking too, Vanguard has some reasonably cheap quant funds that are deliberately factor loading (albeit using non-standard measures that overlap with but are not the same as academic factor definitions). Let me throw that in there too, though note the similar mid/small cap fund had lower loadings and both are more or less running the same strat, so I wouldn't count on consistency and similar size of those tilts in the future.

Individually:

Code: Select all

Ticker                 Mkt    Size   Value Momentum Quality Alpha (annual)
IJS                   1.08    1.02    0.32    0.06    0.44   -0.11%
MTUM                  0.96   -0.15   -0.22    0.26   -0.10    1.37%
VSTCX                 1.12    0.80    0.14    0.21    0.13   -0.84%
https://www.portfoliovisualizer.com/fac ... ssetType=1
Well, that didn't change all that much for IJS.

But I think it's a fair question whether -0.15 size, -0.22 value, -0.10 quality is a price worth paying for 0.26 momentum, particularly given that this is primarily coming via large caps. If there are interaction effects and momentum in large caps is correlated with momentum in small caps but the latter effect is more positive on average then alpha for MTUM in the future may be biased downwards if anything. MTUM looks good now since inception but that's only 3.5 years and part of what looks good is just the alpha so far. We should probably be nervous about relying on alpha from factor funds. Or any fund.
User avatar
Topic Author
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 9:59 am
Location: Hic sunt dracones

Re: Diversifying SCV with MTUM: yea or nay?

Post by CULater »

So, what do you think of these loadings and which ETF do you think generated them using Portfolio Visualizer AQR Factor data? Doesn't look too bad to me -- what do you think?

Market__Size__Value__MOM__Quality
0.84...-0.24...0.21...0.24....0.23
On the internet, nobody knows you're a dog.
Post Reply