Small Cap Value heads Rejoice !!!

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 10:06 am

YRT70 wrote:
Tue Nov 26, 2019 8:22 am
305pelusa wrote:
Tue Nov 26, 2019 8:17 am
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
Yeah I noticed your post on that. IIRC you decided for a 0.2 loading on size and value?

So why did you choose 0.2 and not 0.3 for example?
I decided on 0.3/0.3. Its a good compromise between the deep tilt I wanted, but not so much that I have to start using more expensive and uncommon funds. For reference, the DFA Core 1 and 2 target around 0.1/0.1 and 0.2/0.2 respectively. Larry has said in the past he tends to start most clients at around 0.3/0.3. And Robert T uses 0.2/0.4. So 0.3/0.3 seems overall reasonable and I went with it

Random Walker
Posts: 4229
Joined: Fri Feb 23, 2007 8:21 pm

Re: Small Cap Value heads Rejoice !!!

Post by Random Walker » Tue Nov 26, 2019 10:11 am

Most long only SV funds have a loading on market factor of about 1. In the effort to diversify across factors and move towards risk parity, I like the idea of heavily loading on SV funds with deep exposure and concomitantly increasing safe bond exposure. Big fan of Larry Swedroe’s Black Swan Book.

Dave

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 10:24 am

305pelusa wrote:
Tue Nov 26, 2019 10:06 am
YRT70 wrote:
Tue Nov 26, 2019 8:22 am
305pelusa wrote:
Tue Nov 26, 2019 8:17 am
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
Yeah I noticed your post on that. IIRC you decided for a 0.2 loading on size and value?

So why did you choose 0.2 and not 0.3 for example?
I decided on 0.3/0.3. Its a good compromise between the deep tilt I wanted, but not so much that I have to start using more expensive and uncommon funds. For reference, the DFA Core 1 and 2 target around 0.1/0.1 and 0.2/0.2 respectively. Larry has said in the past he tends to start most clients at around 0.3/0.3. And Robert T uses 0.2/0.4. So 0.3/0.3 seems overall reasonable and I went with it
Interesting. I'm playing around with PV now.

Say we only focus on US, would it be possible to figure out an allocation between SCV and TSM that comes closest to giving the smoothest expected performance? I'm not sure if I'm formulating this right but I think you understand.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 10:35 am

Random Walker wrote:
Tue Nov 26, 2019 10:11 am
Most long only SV funds have a loading on market factor of about 1. In the effort to diversify across factors and move towards risk parity, I like the idea of heavily loading on SV funds with deep exposure and concomitantly increasing safe bond exposure. Big fan of Larry Swedroe’s Black Swan Book.

Dave
I like that portfolio too but some factors drive me away from it: especially the possibility that SCV will continue to underperform TSM for a longer period is a reason I don't want to have 100% SCV. I want something milder.

And then there's the local situation here (Netherlands): expected returns on government bonds are negative, they include risky countries like Italy and new upcoming tax regulations make the expected return on bonds even lower.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 10:37 am

YRT70 wrote:
Tue Nov 26, 2019 10:24 am
305pelusa wrote:
Tue Nov 26, 2019 10:06 am
YRT70 wrote:
Tue Nov 26, 2019 8:22 am
305pelusa wrote:
Tue Nov 26, 2019 8:17 am
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
Yeah I noticed your post on that. IIRC you decided for a 0.2 loading on size and value?

So why did you choose 0.2 and not 0.3 for example?
I decided on 0.3/0.3. Its a good compromise between the deep tilt I wanted, but not so much that I have to start using more expensive and uncommon funds. For reference, the DFA Core 1 and 2 target around 0.1/0.1 and 0.2/0.2 respectively. Larry has said in the past he tends to start most clients at around 0.3/0.3. And Robert T uses 0.2/0.4. So 0.3/0.3 seems overall reasonable and I went with it
Interesting. I'm playing around with PV now.

Say we only focus on US, would it be possible to figure out an allocation between SCV and TSM that comes closest to giving the smoothest expected performance? I'm not sure if I'm formulating this right but I think you understand.
I'm not super sure what you mean by smoothest performance. You meab highest Sharpe or lowest volatility?

Two things about PV:
1) It only has data since fund inception. Using actual index data will some times double or triple the amount of data to look at. You're probably getting sick of me repeating that haha.
2) It only shows you the PAST. So playing around will tell you what gave the "smoothest" performance (whatever that is). But knowing if it is EXPECTED to requires some assumptions about future returns, premiums and volatilities.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 11:09 am

305pelusa wrote:
Tue Nov 26, 2019 10:37 am
YRT70 wrote:
Tue Nov 26, 2019 10:24 am
305pelusa wrote:
Tue Nov 26, 2019 10:06 am
YRT70 wrote:
Tue Nov 26, 2019 8:22 am
305pelusa wrote:
Tue Nov 26, 2019 8:17 am
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
Yeah I noticed your post on that. IIRC you decided for a 0.2 loading on size and value?

So why did you choose 0.2 and not 0.3 for example?
I decided on 0.3/0.3. Its a good compromise between the deep tilt I wanted, but not so much that I have to start using more expensive and uncommon funds. For reference, the DFA Core 1 and 2 target around 0.1/0.1 and 0.2/0.2 respectively. Larry has said in the past he tends to start most clients at around 0.3/0.3. And Robert T uses 0.2/0.4. So 0.3/0.3 seems overall reasonable and I went with it
Interesting. I'm playing around with PV now.

Say we only focus on US, would it be possible to figure out an allocation between SCV and TSM that comes closest to giving the smoothest expected performance? I'm not sure if I'm formulating this right but I think you understand.
I'm not super sure what you mean by smoothest performance. You meab highest Sharpe or lowest volatility?
I need to get more clear on that myself. You wrote this in another thread: "An investor who expects SV to outperform in the long run but is unwilling to tolerate the excessive tracking error and additional volatility produced by a full tilt."

That really applies to me. So basically I'm looking to reduce excessive outperformance and (greatly) reduce tracking error regret. How to quantify that, I'm not sure.
Two things about PV:
1) It only has data since fund inception. Using actual index data will some times double or triple the amount of data to look at. You're probably getting sick of me repeating that haha.
I've read you say that before but not in the context of PV. I can't use index data in PV or can I?
2) It only shows you the PAST. So playing around will tell you what gave the "smoothest" performance (whatever that is). But knowing if it is EXPECTED to requires some assumptions about future returns, premiums and volatilities.
Yes I realise that. That's why I asked if it would be useful to set up an MC tool to do this.

Oh and one more question: what settings on PV do you recommend to check factor loadings? FF or AQR, 3/4/5, QMJ, rolling period etc. Thanks by the way.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 11:32 am

YRT70 wrote:
Tue Nov 26, 2019 11:09 am
That really applies to me. So basically I'm looking to reduce excessive outperformance and (greatly) reduce tracking error regret. How to quantify that, I'm not sure.
Haha it applies to all of us! One could quantify it by noting the historical pervasiveness of the factors and calculating potential tracking error. I think that's asking too much of the (limited) data us retail investors can access. And it's only the past; future will look different any ways.

Settling on the desired factor exposure is, IMO, the most arbitrary and difficult part. I think 0.3/0.3 is as deep as I'd go as a retail investors with only some basic knowledge of factors. And 0.1/0.1 is as low as I'd bother tilting. So perhaps split the difference and use 0.2/0.2? That's consistent with what some DFA funds target so it's not unreasonable. If you like value/have more faith on value than size, maybe a little more value.
YRT70 wrote:
Tue Nov 26, 2019 11:09 am
I've read you say that before but not in the context of PV. I can't use index data in PV or can I?
No, you can't. You need to use Excel and download index data.
YRT70 wrote:
Tue Nov 26, 2019 11:09 am
Yes I realise that. That's why I asked if it would be useful to set up an MC tool to do this.

Oh and one more question: what settings on PV do you recommend to check factor loadings? FF or AQR, 3/4/5, QMJ, rolling period etc. Thanks by the way.
Once I've chosen my allocation, I might use a MC to see how it might do and figure out savings rates. But I would never use the MC to figure out what allocation to use. The difference is key. I know whatever expected return assumptions I make will be wrong. So picking allocation based on it is an exercise in futility.

So I use index data to figure out exposure and pick a portfolio from there. At no point am I looking at historical or future expected returns to determine the "smoothest" portfolio. IMO, it's data mining.

I like FF's factors.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 11:54 am

305pelusa wrote:
Tue Nov 26, 2019 11:32 am
YRT70 wrote:
Tue Nov 26, 2019 11:09 am
That really applies to me. So basically I'm looking to reduce excessive outperformance and (greatly) reduce tracking error regret. How to quantify that, I'm not sure.
Haha it applies to all of us!
Not really right? People going with the Larry style portfolio (only SCV & interm. bonds) are choosing to accept relatively larger potential tracking error regret I thought.
305pelusa wrote:
Tue Nov 26, 2019 11:32 am
Once I've chosen my allocation, I might use a MC to see how it might do and figure out savings rates. But I would never use the MC to figure out what allocation to use. The difference is key. I know whatever expected return assumptions I make will be wrong. So picking allocation based on it is an exercise in futility.

So I use index data to figure out exposure and pick a portfolio from there. At no point am I looking at historical or future expected returns to determine the "smoothest" portfolio. IMO, it's data mining.

I like FF's factors.
Ok got ya. I'll use the FF factors. When you talk about 0.3 and 0.3, is that done with 3, 4 or 5 factors? With 5 factors I'm not coming there. And should I leave it on 36 months rolling periods?

60 IJS/40 VTI gives 0.49 on size and 0.28 on value with FF 4 factor model.

MotoTrojan
Posts: 7602
Joined: Wed Feb 01, 2017 8:39 pm

Re: Small Cap Value heads Rejoice !!!

Post by MotoTrojan » Tue Nov 26, 2019 12:12 pm

305pelusa wrote:
Tue Nov 26, 2019 8:17 am
YRT70 wrote:
Tue Nov 26, 2019 7:56 am
How have you guys decided on how much you want to tilt?

I believe tilting is a good idea but I have hard time deciding on how much tilt I prefer. I saw some people have 90% TSM 10% IJS for their US allocation but that seems like such a small tilt that it's hardly worth it. 100% IJS for US stocks feels to aggressive for me personally even though more diversification across factors makes sense.

Is there anything I can do to get clearer on my desired tilt? Maybe run MC simulations?

I've played around with back testing but since inception of IJS it doesn't seem to matter much how much one tilts, assuming the holding time is long enough.

So how have you guys settled on your amount of tilting?
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
I am one of the more naive types (pick a number and preferred index) myself with a current allocation of 27% of US in SCV (20% total portfolio). I prefer the S&P600 variant but unfortunately hold some of the Vanguard CRSP in my 401k. For International SCV I hold as much as I can in my Roth now via the new Avantis fund, and would have no qualms about 100% of my International exposure being in SCV for even deeper diversification; currently International is 25-30% of equities.

305pelusa do you mind sharing your overall allocation or pointing me to a place where you have before?

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 12:20 pm

MotoTrojan wrote:
Tue Nov 26, 2019 12:12 pm
305pelusa wrote:
Tue Nov 26, 2019 8:17 am
YRT70 wrote:
Tue Nov 26, 2019 7:56 am
How have you guys decided on how much you want to tilt?

I believe tilting is a good idea but I have hard time deciding on how much tilt I prefer. I saw some people have 90% TSM 10% IJS for their US allocation but that seems like such a small tilt that it's hardly worth it. 100% IJS for US stocks feels to aggressive for me personally even though more diversification across factors makes sense.

Is there anything I can do to get clearer on my desired tilt? Maybe run MC simulations?

I've played around with back testing but since inception of IJS it doesn't seem to matter much how much one tilts, assuming the holding time is long enough.

So how have you guys settled on your amount of tilting?
I decided how much factor exposure I wanted, and then chose whatever funds got me that total factor exposure at the cheapest price.

Tilting as a % of assets is very nebulous. 20% of US equity in SCV doesn't tell you anything unless you know the fund. If it's VBR? Maybe won't do much. If it's SLYV? That might be about right. If it's RZV, that might be a lot more than you intended. And so on. When I read Bernstein's book and see he recommends, say 50% of US on SCV, I just shake my head. That honestly tells you nothing.
I am one of the more naive types (pick a number and preferred index) myself with a current allocation of 27% of US in SCV (20% total portfolio). I prefer the S&P600 variant but unfortunately hold some of the Vanguard CRSP in my 401k. For International SCV I hold as much as I can in my Roth now via the new Avantis fund, and would have no qualms about 100% of my International exposure being in SCV for even deeper diversification; currently International is 25-30% of equities.

305pelusa do you mind sharing your overall allocation or pointing me to a place where you have before?
My current allocation is far from my desired factor exposure. I'm currently leveraged as part of a Lifecycle Investing approach. I believe time diversification far outweighs asset diversification or factor exposure.
viewtopic.php?f=10&t=274390

My desired endgame equity portfolio is:
21% VIOV
20% VB
12% MTUM
7% IUSV
14% IVLU
11% FNDC
8% FNDE
7% EMGF

It targets 0.3/0.3 with neutral momentum. The factor exposure is as evenly spread out domestically/Int/EM as I felt reasonable with ERs.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 12:30 pm

YRT70 wrote:
Tue Nov 26, 2019 11:54 am
305pelusa wrote:
Tue Nov 26, 2019 11:32 am
YRT70 wrote:
Tue Nov 26, 2019 11:09 am
That really applies to me. So basically I'm looking to reduce excessive outperformance and (greatly) reduce tracking error regret. How to quantify that, I'm not sure.
Haha it applies to all of us!
Not really right? People going with the Larry style portfolio (only SCV & interm. bonds) are choosing to accept relatively larger potential tracking error regret I thought.
305pelusa wrote:
Tue Nov 26, 2019 11:32 am
Once I've chosen my allocation, I might use a MC to see how it might do and figure out savings rates. But I would never use the MC to figure out what allocation to use. The difference is key. I know whatever expected return assumptions I make will be wrong. So picking allocation based on it is an exercise in futility.

So I use index data to figure out exposure and pick a portfolio from there. At no point am I looking at historical or future expected returns to determine the "smoothest" portfolio. IMO, it's data mining.

I like FF's factors.
Ok got ya. I'll use the FF factors. When you talk about 0.3 and 0.3, is that done with 3, 4 or 5 factors? With 5 factors I'm not coming there. And should I leave it on 36 months rolling periods?

60 IJS/40 VTI gives 0.49 on size and 0.28 on value with FF 4 factor model.
First off, there's no "Larry portfolio". The one quoted is just an example he gives of the theory Dave said in this thread; add factor exposure to reduce equity allocation.

In that vein, how much exposure you add is simply a question of how much tracking error and faith you have in factors. We all want factor exposure with little or no tracking error (market underperformance). So it's a balancing act and it's all relative.

I use 0.3/0.3 in a 3 Factor + MOM analysis (the FF 4F in PV). That's because the frames of reference I have (DFA Core funds, Larry and Robert T) did it back before 5F.

The numbers you quote look about right. Welcome to your first finding when using factor regression instead of just % allocation: Funds like SLYV are smaller than they are valuable so if you want even size and value exposure, you'll find you need to add LCV (like the IUSV and IVLU I added in my previous post). Or use something like VOE which is more even in exposure.

You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 12:56 pm

305pelusa wrote:
Tue Nov 26, 2019 12:30 pm
First off, there's no "Larry portfolio".
Well Larry calls it the "Larry portfolio" in his book "Reducing the risk of black swans". And yes I'm aware of the ideas behind it and that he doesn't recommend it to everyone.
In that vein, how much exposure you add is simply a question of how much tracking error and faith you have in factors. We all want factor exposure with little or no tracking error (market underperformance). So it's a balancing act and it's all relative.

I use 0.3/0.3 in a 3 Factor + MOM analysis (the FF 4F in PV). That's because the frames of reference I have (DFA Core funds, Larry and Robert T) did it back before 5F.

The numbers you quote look about right. Welcome to your first finding when using factor regression instead of just % allocation: Funds like SLYV are smaller than they are valuable so if you want even size and value exposure, you'll find you need to add LCV (like the IUSV and IVLU I added in my previous post). Or use something like VOE which is more even in exposure.

You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".
Thanks that is helpful.

MotoTrojan
Posts: 7602
Joined: Wed Feb 01, 2017 8:39 pm

Re: Small Cap Value heads Rejoice !!!

Post by MotoTrojan » Tue Nov 26, 2019 1:00 pm

305pelusa wrote:
Tue Nov 26, 2019 12:30 pm


You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".
I guess we will find out if us willy-nilly folks notice any subtleties in our returns too :sharebeer .

Interesting stuff for sure. Perhaps something like the new Avantis AVUV would give you better small AND value exposure all in one; I haven't run any regressions but it sure looks a good bit more value'y than the S&P600.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 1:13 pm

MotoTrojan wrote:
Tue Nov 26, 2019 1:00 pm
305pelusa wrote:
Tue Nov 26, 2019 12:30 pm


You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".
I guess we will find out if us willy-nilly folks notice any subtleties in our returns too :sharebeer .

Interesting stuff for sure. Perhaps something like the new Avantis AVUV would give you better small AND value exposure all in one; I haven't run any regressions but it sure looks a good bit more value'y than the S&P600.
No clue. FWIW:
1) Avantis is a new company. That has a host of Black Swan vulnerabilities. Do they know how to manage a company? Taxation? What about fraud? All unlikely but I'll stick to Vanguard/iShares/Schwab.
2) Looked like the funds had an active management component (security selection). Huh no thanks.
3) I think most are getting convinced by the appeal to authority about being ex-DFA. I frankly couldn't care less.
4) God knows what exposures they'll get.

These are all uncompensated risks. Just beware broski.

MotoTrojan
Posts: 7602
Joined: Wed Feb 01, 2017 8:39 pm

Re: Small Cap Value heads Rejoice !!!

Post by MotoTrojan » Tue Nov 26, 2019 1:29 pm

305pelusa wrote:
Tue Nov 26, 2019 1:13 pm
MotoTrojan wrote:
Tue Nov 26, 2019 1:00 pm
305pelusa wrote:
Tue Nov 26, 2019 12:30 pm


You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".
I guess we will find out if us willy-nilly folks notice any subtleties in our returns too :sharebeer .

Interesting stuff for sure. Perhaps something like the new Avantis AVUV would give you better small AND value exposure all in one; I haven't run any regressions but it sure looks a good bit more value'y than the S&P600.
No clue. FWIW:
1) Avantis is a new company. That has a host of Black Swan vulnerabilities. Do they know how to manage a company? Taxation? What about fraud? All unlikely but I'll stick to Vanguard/iShares/Schwab.
2) Looked like the funds had an active management component (security selection). Huh no thanks.
3) I think most are getting convinced by the appeal to authority about being ex-DFA. I frankly couldn't care less.
4) God knows what exposures they'll get.

These are all uncompensated risks. Just beware broski.
Much appreciated thoughts. In general I planned to dabble in their AVDV (International SCV) but stick with S&P600 funds for domestic SCV exposure to not put all eggs in one basket.

It was not my understanding that there was direct security selection, more-so just non-index driven screening methods, but I may be wrong. Either way none of these types of holdings would be in taxable, so that drastically reduces the black swan risks to nearly zero. I'll keep an eye out for other international small-value products but thus far I haven't seen any from the major players that look super appealing; DLS looked the best but the ER is pretty hefty.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 1:36 pm

305pelusa wrote:
Tue Nov 26, 2019 12:20 pm
My desired endgame equity portfolio is:
21% VIOV
20% VB
12% MTUM
7% IUSV
14% IVLU
11% FNDC
8% FNDE
7% EMGF

It targets 0.3/0.3 with neutral momentum. The factor exposure is as evenly spread out domestically/Int/EM as I felt reasonable with ERs.
When I select FF 4 factor in PV I see no option of doing a 'whole world' regression. Do you do it separately for US, Int and EM?

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 1:48 pm

YRT70 wrote:
Tue Nov 26, 2019 1:36 pm
305pelusa wrote:
Tue Nov 26, 2019 12:20 pm
My desired endgame equity portfolio is:
21% VIOV
20% VB
12% MTUM
7% IUSV
14% IVLU
11% FNDC
8% FNDE
7% EMGF

It targets 0.3/0.3 with neutral momentum. The factor exposure is as evenly spread out domestically/Int/EM as I felt reasonable with ERs.
When I select FF 4 factor in PV I see no option of doing a 'whole world' regression. Do you do it separately for US, Int and EM?
Yes, separately. Some ETFs are too new to use PV and some have a fund that tracks the same thing that's older. Here are the substitutions to get more representative data:

EMGF->SFENX
FNDE->SFENX
FNDC->SFILX
IVLU->SFNNX

That'll give you data since 2013. I'd be curious what you get. It's too short of a time period to make any conclusions but might be a fun exercise.

rascott
Posts: 1266
Joined: Wed Apr 15, 2015 10:53 am

Re: Small Cap Value heads Rejoice !!!

Post by rascott » Tue Nov 26, 2019 1:55 pm

305pelusa wrote:
Tue Nov 26, 2019 1:13 pm
MotoTrojan wrote:
Tue Nov 26, 2019 1:00 pm
305pelusa wrote:
Tue Nov 26, 2019 12:30 pm


You wouldn't notice that kind of subtlety if you just willy-nilly "sliced and diced".
I guess we will find out if us willy-nilly folks notice any subtleties in our returns too :sharebeer .

Interesting stuff for sure. Perhaps something like the new Avantis AVUV would give you better small AND value exposure all in one; I haven't run any regressions but it sure looks a good bit more value'y than the S&P600.
No clue. FWIW:
1) Avantis is a new company. That has a host of Black Swan vulnerabilities. Do they know how to manage a company? Taxation? What about fraud? All unlikely but I'll stick to Vanguard/iShares/Schwab.
2) Looked like the funds had an active management component (security selection). Huh no thanks.
3) I think most are getting convinced by the appeal to authority about being ex-DFA. I frankly couldn't care less.
4) God knows what exposures they'll get.

These are all uncompensated risks. Just beware broski.
It's actually American Century.....Avantis is just a new brand name for them. So wouldn't be too concerned with #1.

No idea on the other points either, as I'll wait for a long while before considering these funds.

pdavi21
Posts: 1296
Joined: Sat Jan 30, 2016 4:04 pm

Re: Small Cap Value heads Rejoice !!!

Post by pdavi21 » Tue Nov 26, 2019 2:03 pm

Yikes, I thought factor regression was losing popularity. If you are going to compare performance to factor loadings, the factor loadings shouldn't be derived from performance...just my opinion.
"We spend a great deal of time studying history, which, let's face it, is mostly the history of stupidity." -Stephen Hawking

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 2:12 pm

305pelusa wrote:
Tue Nov 26, 2019 1:48 pm
YRT70 wrote:
Tue Nov 26, 2019 1:36 pm
305pelusa wrote:
Tue Nov 26, 2019 12:20 pm
My desired endgame equity portfolio is:
21% VIOV
20% VB
12% MTUM
7% IUSV
14% IVLU
11% FNDC
8% FNDE
7% EMGF

It targets 0.3/0.3 with neutral momentum. The factor exposure is as evenly spread out domestically/Int/EM as I felt reasonable with ERs.
When I select FF 4 factor in PV I see no option of doing a 'whole world' regression. Do you do it separately for US, Int and EM?
Yes, separately. Some ETFs are too new to use PV and some have a fund that tracks the same thing that's older. Here are the substitutions to get more representative data:

EMGF->SFENX
FNDE->SFENX
FNDC->SFILX
IVLU->SFNNX

That'll give you data since 2013. I'd be curious what you get. It's too short of a time period to make any conclusions but might be a fun exercise.
Ok here's my attempt at the US portion only. Value is only 0.14. Am I doing something wrong?
https://www.portfoliovisualizer.com/fac ... n4_1=11.67

Random Walker
Posts: 4229
Joined: Fri Feb 23, 2007 8:21 pm

Re: Small Cap Value heads Rejoice !!!

Post by Random Walker » Tue Nov 26, 2019 2:15 pm

YRT70 wrote:
Tue Nov 26, 2019 10:35 am
Random Walker wrote:
Tue Nov 26, 2019 10:11 am
Most long only SV funds have a loading on market factor of about 1. In the effort to diversify across factors and move towards risk parity, I like the idea of heavily loading on SV funds with deep exposure and concomitantly increasing safe bond exposure. Big fan of Larry Swedroe’s Black Swan Book.

Dave
I like that portfolio too but some factors drive me away from it: especially the possibility that SCV will continue to underperform TSM for a longer period is a reason I don't want to have 100% SCV. I want something milder.

And then there's the local situation here (Netherlands): expected returns on government bonds are negative, they include risky countries like Italy and new upcoming tax regulations make the expected return on bonds even lower.
I’d be curious to hear what Larry would say about that situation. Basically he usually says that the bonds are for safety, to cool off the portfolio volatility to one’s tolerable point. So how mild one goes can be controlled with bonds. I’d be a little surprised if negative rates affect that philosophy.

Dave

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 2:20 pm

Random Walker wrote:
Tue Nov 26, 2019 2:15 pm
YRT70 wrote:
Tue Nov 26, 2019 10:35 am
Random Walker wrote:
Tue Nov 26, 2019 10:11 am
Most long only SV funds have a loading on market factor of about 1. In the effort to diversify across factors and move towards risk parity, I like the idea of heavily loading on SV funds with deep exposure and concomitantly increasing safe bond exposure. Big fan of Larry Swedroe’s Black Swan Book.

Dave
I like that portfolio too but some factors drive me away from it: especially the possibility that SCV will continue to underperform TSM for a longer period is a reason I don't want to have 100% SCV. I want something milder.

And then there's the local situation here (Netherlands): expected returns on government bonds are negative, they include risky countries like Italy and new upcoming tax regulations make the expected return on bonds even lower.
I’d be curious to hear what Larry would say about that situation. Basically he usually says that the bonds are for safety, to cool off the portfolio volatility to one’s tolerable point. So how mild one goes can be controlled with bonds. I’d be a little surprised if negative rates affect that philosophy.

Dave
I did ask him, from what I can tell he didn't object to my move away from bonds into deposits. You might want to ask him yourself if you want to be sure.

If you knew your money would be expected to return -0.1% in bonds and 1% in deposits, would you still go with bonds? Why not use deposits to cool off the volatility?

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 2:27 pm

YRT70 wrote:
Tue Nov 26, 2019 2:12 pm
305pelusa wrote:
Tue Nov 26, 2019 1:48 pm
YRT70 wrote:
Tue Nov 26, 2019 1:36 pm
305pelusa wrote:
Tue Nov 26, 2019 12:20 pm
My desired endgame equity portfolio is:
21% VIOV
20% VB
12% MTUM
7% IUSV
14% IVLU
11% FNDC
8% FNDE
7% EMGF

It targets 0.3/0.3 with neutral momentum. The factor exposure is as evenly spread out domestically/Int/EM as I felt reasonable with ERs.
When I select FF 4 factor in PV I see no option of doing a 'whole world' regression. Do you do it separately for US, Int and EM?
Yes, separately. Some ETFs are too new to use PV and some have a fund that tracks the same thing that's older. Here are the substitutions to get more representative data:

EMGF->SFENX
FNDE->SFENX
FNDC->SFILX
IVLU->SFNNX

That'll give you data since 2013. I'd be curious what you get. It's too short of a time period to make any conclusions but might be a fun exercise.
Ok here's my attempt at the US portion only. Value is only 0.14. Am I doing something wrong?
https://www.portfoliovisualizer.com/fac ... n4_1=11.67
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Tue Nov 26, 2019 2:41 pm

305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
Interesting. Is there anything 'better' about value in EM? My portfolio does basically the opposite: it's more 'valuey' in US and less in EM.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 2:56 pm

YRT70 wrote:
Tue Nov 26, 2019 2:41 pm
305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
Interesting. Is there anything 'better' about value in EM? My portfolio does basically the opposite: it's more 'valuey' in US and less in EM.
Factor exposure is hard to obtain in EM. So I do my best with SFENX and EMGF, and complete as needed domestically. Again, I just want as much spread out exposure over the globe

User avatar
Topic Author
hdas
Posts: 1335
Joined: Thu Jun 11, 2015 8:24 am

Re: Small Cap Value heads Rejoice !!!

Post by hdas » Tue Nov 26, 2019 3:31 pm

305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession.
There are too many unsubstantiated assumptions in this sentence that make one head hurt. Do you have any references or data to support this mumbo?

Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Tue Nov 26, 2019 4:14 pm

hdas wrote:
Tue Nov 26, 2019 3:31 pm
305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession.
There are too many unsubstantiated assumptions in this sentence that make one head hurt. Do you have any references or data to support this mumbo?

Cheers :greedy
Here's an example:
viewtopic.php?p=4812255#p4812255

I don't have a reference or definitive data set that proves 6 years is not good enough. But having regressed various indices, I've personally found it almost never is good enough.
If you really think 6 years is indicative of future long term factor exposure, by all means be my guest. I think going to 20+ years makes for a more robust conclusion.

rkhusky
Posts: 7857
Joined: Thu Aug 18, 2011 8:09 pm

Re: Small Cap Value heads Rejoice !!!

Post by rkhusky » Tue Nov 26, 2019 6:27 pm

305pelusa wrote:
Tue Nov 26, 2019 4:14 pm
I think going to 20+ years makes for a more robust conclusion.
It's not clear that the average factor exposure for any particular fund over the next 20 years will be the same as for the past 20 years (except for TSM and the base portfolios that are used in the regression). The farther a fund is from the base portfolios, the more likely that the factor exposure will fluctuate.

Random Walker
Posts: 4229
Joined: Fri Feb 23, 2007 8:21 pm

Re: Small Cap Value heads Rejoice !!!

Post by Random Walker » Tue Nov 26, 2019 10:31 pm

YRT70 wrote:
Tue Nov 26, 2019 2:20 pm
Random Walker wrote:
Tue Nov 26, 2019 2:15 pm
YRT70 wrote:
Tue Nov 26, 2019 10:35 am
Random Walker wrote:
Tue Nov 26, 2019 10:11 am
Most long only SV funds have a loading on market factor of about 1. In the effort to diversify across factors and move towards risk parity, I like the idea of heavily loading on SV funds with deep exposure and concomitantly increasing safe bond exposure. Big fan of Larry Swedroe’s Black Swan Book.

Dave
I like that portfolio too but some factors drive me away from it: especially the possibility that SCV will continue to underperform TSM for a longer period is a reason I don't want to have 100% SCV. I want something milder.

And then there's the local situation here (Netherlands): expected returns on government bonds are negative, they include risky countries like Italy and new upcoming tax regulations make the expected return on bonds even lower.
I’d be curious to hear what Larry would say about that situation. Basically he usually says that the bonds are for safety, to cool off the portfolio volatility to one’s tolerable point. So how mild one goes can be controlled with bonds. I’d be a little surprised if negative rates affect that philosophy.

Dave
I did ask him, from what I can tell he didn't object to my move away from bonds into deposits. You might want to ask him yourself if you want to be sure.

If you knew your money would be expected to return -0.1% in bonds and 1% in deposits, would you still go with bonds? Why not use deposits to cool off the volatility?
Totally agree

Dave

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Wed Nov 27, 2019 10:51 am

305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
OK so I'm trying to learn more about factor portfolios so bare with me. I wonder why you went with almost the same amount of VB as VIOV. From what I've seen so far the S&P 600 SCV funds aren't that valuey to begin with, adding a lot of VB reduces the value loading a lot, especially with MTUM added (as indicated by the 0.14 loading). Is it only because you have more value loading in EM? Wouldn't it make more sense to get more VIOV and less VB, so the value loading increases?

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Wed Nov 27, 2019 11:06 am

YRT70 wrote:
Wed Nov 27, 2019 10:51 am
305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
OK so I'm trying to learn more about factor portfolios so bare with me. I wonder why you went with almost the same amount of VB as VIOV. From what I've seen so far the S&P 600 SCV funds aren't that valuey to begin with, adding a lot of VB reduces the value loading a lot, especially with MTUM added (as indicated by the 0.14 loading). Is it only because you have more value loading in EM? Wouldn't it make more sense to get more VIOV and less VB, so the value loading increases?
If I substituted VB for VIOV in that portfolio, then I'd end up with more overall value loading than I wanted (more than my target 0.3).

Uncorrelated
Posts: 259
Joined: Sun Oct 13, 2019 3:16 pm

Re: Small Cap Value heads Rejoice !!!

Post by Uncorrelated » Wed Nov 27, 2019 11:13 am

I wasted some time to run a mean variance optimizer on a combinations of factor funds. Although this type of analysis isn't known for it's robustness, it can be used the get a general idea. I used average factor premia data from Fama & French 1963~1991, halved the size and value premium, and added 50% to the term premium. Factor exposure was calculated using PV with a FF 3 factor model on the longest available time interval (IJS 0.8/0.4 size/value, IUSV 0.0/0.4, BOSVX 0.92/0.67). Correlations between factors were assumed to be zero.

total stock market + bonds.
Image
We see that the optimal term exposure is approximately half of the market exposure. Note that short term bonds can be synthesized using LTT + cash, in the cash-heavy region it would be more appropriate to choose intermediate term or short term bonds to capitalize on bet-against-beta effects in bonds.

total stock market + IJS + IUSV
Image
We see that the optimal factors are approximately TERM = HmL = SmB = 2 * MKT. No allocation to total stock market is chosen anywhere along the efficient frontier. Note that obtaining market exposure from small cap and value funds results in higher idiosyncratic risk than total stock market, but I have no idea how large the impact of this effect is.

total stock market + IJS + IUSV + BOSVX
Image
BOSVX makes IJS redundant, but the difference is small enough and possibly the result of parameter estimation errors.

Constraining the total stock market exposure to be at least as large as IUSV or BOSVX exposure results in a portfolio with equal parts of TSM, IUSV and BOSVX and reduces performance by around 0.07% annually:
Image

With shorting allowed, the optimizer shorts total stock market in favor of IUSV, even if the shorting premium is substantial (0.5% over the risk free rate). In this scenario, the optimizer heavily favors IUSV over IJS/BOSVX even though both have identical value loadings (resulting factor exposure is approximately TERM = SmB = 2 * HmL = 3 * MKT) This indicates that the fund selection is constrained primarily by access to the value factor. A fund with deeper value exposure would help.

tldr: picking equal parts of total stock market, small cap value and value is probably fine. Value is more important than size.

User avatar
whodidntante
Posts: 7114
Joined: Thu Jan 21, 2016 11:11 pm
Location: outside the echo chamber

Re: Small Cap Value heads Rejoice !!!

Post by whodidntante » Wed Nov 27, 2019 11:20 am

rkhusky wrote:
Tue Nov 26, 2019 6:27 pm
305pelusa wrote:
Tue Nov 26, 2019 4:14 pm
I think going to 20+ years makes for a more robust conclusion.
It's not clear that the average factor exposure for any particular fund over the next 20 years will be the same as for the past 20 years (except for TSM and the base portfolios that are used in the regression). The farther a fund is from the base portfolios, the more likely that the factor exposure will fluctuate.
That's true. I took a sizeable position in ISCF back when it had a useful value tilt. Now it looks more like a growth fund. :oops:

I guess those who target certain factor loadings and who wish to minimize trading would be better suited by funds that target a certain factor loading. However, that isn't reliable because factor regressions can only be calculated retroactively. I personally let my factor loadings drift, because you can't know what it is, you can only know what it was.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Wed Nov 27, 2019 11:24 am

305pelusa wrote:
Wed Nov 27, 2019 11:06 am
YRT70 wrote:
Wed Nov 27, 2019 10:51 am
305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
OK so I'm trying to learn more about factor portfolios so bare with me. I wonder why you went with almost the same amount of VB as VIOV. From what I've seen so far the S&P 600 SCV funds aren't that valuey to begin with, adding a lot of VB reduces the value loading a lot, especially with MTUM added (as indicated by the 0.14 loading). Is it only because you have more value loading in EM? Wouldn't it make more sense to get more VIOV and less VB, so the value loading increases?
If I substituted VB for VIOV in that portfolio, then I'd end up with more overall value loading than I wanted (more than my target 0.3).
It's only showing 0.14 in the regression I did. How did you come to 0.3? EM is only small portion of your portfolio.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Wed Nov 27, 2019 11:45 am

YRT70 wrote:
Wed Nov 27, 2019 11:24 am
305pelusa wrote:
Wed Nov 27, 2019 11:06 am
YRT70 wrote:
Wed Nov 27, 2019 10:51 am
305pelusa wrote:
Tue Nov 26, 2019 2:27 pm
You've done it correctly.True exposure cannot be judged from 6 years. That's barely long enough to encompass over market cycle. And it doesn't contain a recession. So the exact numbers should not be minded.

What we can conclude is:
1) There's a slight positive MOM tilt, to counteract some of the negative momentum from international value.
2) It is smaller than value-y. This is to counteract the large value EM holdings.

So in general, the signs and magnitudes of the exposure relative to each other seem sensible and as I intended them. But I wouldn't put any faith in the actual numbers.
OK so I'm trying to learn more about factor portfolios so bare with me. I wonder why you went with almost the same amount of VB as VIOV. From what I've seen so far the S&P 600 SCV funds aren't that valuey to begin with, adding a lot of VB reduces the value loading a lot, especially with MTUM added (as indicated by the 0.14 loading). Is it only because you have more value loading in EM? Wouldn't it make more sense to get more VIOV and less VB, so the value loading increases?
If I substituted VB for VIOV in that portfolio, then I'd end up with more overall value loading than I wanted (more than my target 0.3).
It's only showing 0.14 in the regression I did. How did you come to 0.3? EM is only small portion of your portfolio.
I used much more data. I regressed the actual underlying index data to get data back to around 1996 for all of these positions.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Wed Nov 27, 2019 11:50 am

305pelusa wrote:
Wed Nov 27, 2019 11:45 am
I used much more data. I regressed the actual underlying index data to get data back to around 1996 for all of these positions.
Ok, thank you.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Wed Nov 27, 2019 11:57 am

rkhusky wrote:
Tue Nov 26, 2019 6:27 pm
305pelusa wrote:
Tue Nov 26, 2019 4:14 pm
I think going to 20+ years makes for a more robust conclusion.
It's not clear that the average factor exposure for any particular fund over the next 20 years will be the same as for the past 20 years (except for TSM and the base portfolios that are used in the regression). The farther a fund is from the base portfolios, the more likely that the factor exposure will fluctuate.
I'm not sure what you mean by "clear". Nothing is certain of course but I think if a methodology, over a couple of decades, ended up achieving around X/Y factor exposures, I think it's reasonable to expect it will achieve a similar factor exposure over the long term future. Over the short term, things can vary. But the hope is that variation is random (not systematic) such that it kinda averages out.

Ex: I meant my US holdings to have around 0.3 value. These past 6 years, it had lower. I meant my EM part to have 0.3 value. These past 8 years, it had around 0.5. They don't quite cancel out, but it's at least in opposite directions. As long as the variations are kinda random, I'm hoping they'll roughly cancel out within holdings and across time.

rkhusky
Posts: 7857
Joined: Thu Aug 18, 2011 8:09 pm

Re: Small Cap Value heads Rejoice !!!

Post by rkhusky » Wed Nov 27, 2019 1:36 pm

305pelusa wrote:
Wed Nov 27, 2019 11:57 am
But the hope is that variation is random (not systematic) such that it kinda averages out.
Unless the fund managers decide to change their factor exposures or add some other factor or change the definition of a factor. Or if the publicity of factors over the past 5-10 years has fundamentally changed the interactions between some of the more non-independent factors. If one has a fund tracking one or two factors, like a value fund or a small value fund, then that is much more likely to have constant exposure, than if one is using a more complex 5-factor fund.

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Wed Nov 27, 2019 2:00 pm

rkhusky wrote:
Wed Nov 27, 2019 1:36 pm
Unless the fund managers decide to change their factor exposures or add some other factor or change the definition of a factor.
Well all of these funds track indices; if a fund manager thinks value should be defined some other way, that has no bearing on the fund. Now if the index provider itself changed its methodology, or the fund managers decided to follow some other index, I could see this. I'm not as concerned about this.
rkhusky wrote:
Wed Nov 27, 2019 1:36 pm
Or if the publicity of factors over the past 5-10 years has fundamentally changed the interactions between some of the more non-independent factors.
I'm not following.

Ex: Size is defined as the smallest 5 deciles minus the largest 5 deciles. Since the S&P 600, by methodology, invests in the companies in the small deciles, then I don't see how its popularity/publicity, etc could affect its factor exposure. It could certainly affect the SC premium. But as the small companies swell (as money goes looking for that factor), so does the definition of the factor itself change.

In other words: If the methodology of the fund invests with similar definitions/methodology as the factors themselves are defined, then I don't follow how popularity, interactions, etc have any bearing.

Could you explain?

rkhusky
Posts: 7857
Joined: Thu Aug 18, 2011 8:09 pm

Re: Small Cap Value heads Rejoice !!!

Post by rkhusky » Wed Nov 27, 2019 3:26 pm

305pelusa wrote:
Wed Nov 27, 2019 2:00 pm
In other words: If the methodology of the fund invests with similar definitions/methodology as the factors themselves are defined, then I don't follow how popularity, interactions, etc have any bearing.

Could you explain?
From what I understand, there are some interactions between value, momentum, quality, profitability, investability. If one had a fund that invested in a number of these, there could be unexpected variation over time. The factor decomposition procedure only works well when the factors are approximately independent (e.g. like the size and value factors).

Uncorrelated
Posts: 259
Joined: Sun Oct 13, 2019 3:16 pm

Re: Small Cap Value heads Rejoice !!!

Post by Uncorrelated » Thu Nov 28, 2019 7:34 am

Uncorrelated wrote:
Wed Nov 27, 2019 11:13 am
I wasted some time to run a mean variance optimizer on a combinations of factor funds.
I verified my results using simba's backtesting tool 1927 ~ 2018. According to my model IJS + IUSV + LTT (A) outperforms a 2-fund portfolio by 0.8% and a leveraged variant (B) that shorts large cap growth outperforms 2-fund by 1.1%, all with the same standard deviation. According to simba's backtesting spreadsheet A outperforms 2-fund by 0.79% and nearly identical standard deviation. B outperforms 2-fund by 1.01% with somewhat higher standard deviation (same sharpe ratio as A)

I also updated my model to a 6-factor model (FF5 + momentum, and correlations). In the 6-factor model IUSV dominates all other funds, usually selecting an asset allocation with an 80:20 split between IUSV and LTT and the rest cash.

When I add VFMF, this fund dominates all other funds by an extremely large margin. I have to discount VFMF' factor exposure by 2 and the momentum exposure by 4 before it even starts to consider other funds. I believe this is because the FF2013 factor data I used has similar premia for all different factors (except bond factors) and VFMF is the fund that has the most balanced factor exposure out of the fund selection.

User avatar
Topic Author
hdas
Posts: 1335
Joined: Thu Jun 11, 2015 8:24 am

Re: Small Cap Value heads Rejoice !!!

Post by hdas » Thu Nov 28, 2019 9:09 am

Uncorrelated wrote:
Thu Nov 28, 2019 7:34 am
VFMF is the fund that has the most balanced factor exposure out of the fund selection.
Check XSLV

Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

Uncorrelated
Posts: 259
Joined: Sun Oct 13, 2019 3:16 pm

Re: Small Cap Value heads Rejoice !!!

Post by Uncorrelated » Thu Nov 28, 2019 9:40 am

hdas wrote:
Thu Nov 28, 2019 9:09 am
Uncorrelated wrote:
Thu Nov 28, 2019 7:34 am
VFMF is the fund that has the most balanced factor exposure out of the fund selection.
Check XSLV

Cheers :greedy
Good catch. With XSLV added to my model a mix of XSVL + IUSV dominates VFMF. However, I'm not a fan of XSLV due to the low amount of holdings and sector concentration. I have more confidence in the selection procedure of VFMF (and IJS and IUSV) to deliver consistent factor tracking over time. Unfortunately there is not nearly enough data available to draw conclusions.

According to portfolio visualizer 'rolling regression' and 'factor clone' and R^2, XSLV has a substantial amount of time-varying factor exposure and unexplained risk. Even on ARQ's factor set.

Girya
Posts: 72
Joined: Wed May 29, 2019 7:57 am

Re: Small Cap Value heads Rejoice !!!

Post by Girya » Thu Nov 28, 2019 11:03 am

Hey everyone I've been reading this thread with great interest. I have a portfolio which is tilted small, value and most interestingly small international value. It was developed following a Swedroe model before I left my advisor, and frankly I haven't touched much. Indeed I opened up another taxable account specifically so I wouldn't mess with their allocation (thus making things super complicated); ergo the iShares funds.

I was hoping to post my asset allocation and stocks I have in there, and see what you all think, given the above. I'm not sure if with what I have (both in my own brokerage, and what was developed for me by BAM, AND in my 401K which is at a completely different custodian) meets an appropriate value load without being a "Larry" Portfolio (because I think that pulling allocation fully from TSM would be silly).

The above conversation is interesting but perhaps a touch over my head. I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?

Girya Family Allocation with only REITS
Min Target Max Funds I'm using for this
U.S. Total Market Equity 20 25 30 ITOT, DFSPX, DFUEX
U.S.Total Market Value 4 5 6 IUSV, VVIAX
U.S. Small/Mid Cap 2 5 7 DFMVX
U.S. Small-Cap Value 10 12.5 15 IJS, BOSVX, DFFVX

Equity: Developed Markets 7.5 10 12.5 IXUS
Equity: International Value 7.5 10 12.5 DISVX, DTMIX
Emerging Market Equity 5 6 8.3 DFCEX, DFESX

Domestic REITs 1.6 2.1 2.6 USRT, VGSLX
International REITs 1.1 1.4 1.8
Total Alternatives 0 3.5 4.4 A bunch of bad stuff I'm trying to get rid of (see recent posts)

Total Fixed Income 16 23 30 MUB, G Fund

User avatar
305pelusa
Posts: 1289
Joined: Fri Nov 16, 2018 10:20 pm

Re: Small Cap Value heads Rejoice !!!

Post by 305pelusa » Thu Nov 28, 2019 11:52 am

Girya wrote:
Thu Nov 28, 2019 11:03 am
The above conversation is interesting but perhaps a touch over my head. I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?
If the above is too complex and you don't understand it, it must definitely is not right for you. How about just keep it simple with the 3 Fund? It's an excellent choice.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Thu Nov 28, 2019 11:57 am

Girya wrote:
Thu Nov 28, 2019 11:03 am
I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?
Being able to handle tracking error regret is an important thing when it comes to a tilted portfolio. Given that you've concluded that alternatives are bad investments after a couple of years of underperformance, my feeling is that you may be best served with a normal TSM portfolio.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Thu Nov 28, 2019 12:02 pm

Uncorrelated wrote:
Thu Nov 28, 2019 7:34 am
When I add VFMF, this fund dominates all other funds by an extremely large margin. I have to discount VFMF' factor exposure by 2 and the momentum exposure by 4 before it even starts to consider other funds. I believe this is because the FF2013 factor data I used has similar premia for all different factors (except bond factors) and VFMF is the fund that has the most balanced factor exposure out of the fund selection.
Too bad we can't buy VFMF in the Netherlands. iShares has a bunch of multifactor funds available in Europe. Have you done any analysis on them? https://www.justetf.com/de-en/find-etf. ... ultifactor

Girya
Posts: 72
Joined: Wed May 29, 2019 7:57 am

Re: Small Cap Value heads Rejoice !!!

Post by Girya » Thu Nov 28, 2019 12:15 pm

305pelusa wrote:
Thu Nov 28, 2019 11:52 am
Girya wrote:
Thu Nov 28, 2019 11:03 am
The above conversation is interesting but perhaps a touch over my head. I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?
If the above is too complex and you don't understand it, it must definitely is not right for you. How about just keep it simple with the 3 Fund? It's an excellent choice.
Your post is taken in good heart, but this always bums me out when I ask a question like this on Bogleheads. I politely ask “hey I don’t understand this thing; can you all help me understand it” and what I get is some response like “well you don’t understand it switch to a 3 fund portfolio”.

Maybe for once could someone just take the time to, I don’t know, explain it?

Maybe I’ll learn something, maybe you’ll learn something. Just a suggestion.

Girya
Posts: 72
Joined: Wed May 29, 2019 7:57 am

Re: Small Cap Value heads Rejoice !!!

Post by Girya » Thu Nov 28, 2019 12:22 pm

YRT70 wrote:
Thu Nov 28, 2019 11:57 am
Girya wrote:
Thu Nov 28, 2019 11:03 am
I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?
Being able to handle tracking error regret is an important thing when it comes to a tilted portfolio. Given that you've concluded that alternatives are bad investments after a couple of years of underperformance, my feeling is that you may be best served with a normal TSM portfolio.
Well I’m fairness NO ONE understands alts so my intolerance for tracking error is shared. I’m seeking to understand small and value better so that I can stomach the tracking error, over time. I understand this is a multi year process.

YRT70
Posts: 404
Joined: Sat Apr 27, 2019 8:51 am

Re: Small Cap Value heads Rejoice !!!

Post by YRT70 » Thu Nov 28, 2019 12:26 pm

Girya wrote:
Thu Nov 28, 2019 12:22 pm
YRT70 wrote:
Thu Nov 28, 2019 11:57 am
Girya wrote:
Thu Nov 28, 2019 11:03 am
I want to believe in this approach but wonder if I have the mathematic and investment skills to truly do so, not to mention the stomach, and if I don't understand it, can it ever be right for me?
Being able to handle tracking error regret is an important thing when it comes to a tilted portfolio. Given that you've concluded that alternatives are bad investments after a couple of years of underperformance, my feeling is that you may be best served with a normal TSM portfolio.
Well I’m fairness NO ONE understands alts so my intolerance for tracking error is shared. I’m seeking to understand small and value better so that I can stomach the tracking error, over time. I understand this is a multi year process.
I found the videos by Ben Felix on the topic quite helpful to learn about it. https://www.youtube.com/watch?v=2MVSsVi1_e4

Post Reply