Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

For investors outside the US. Personal investments, personal finance, investing news and theory.
Sister forums: Canada, Spain (en español)
---------------
Post Reply
Topic Author
ilarion
Posts: 3
Joined: Tue May 12, 2020 1:25 am

Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by ilarion » Wed May 13, 2020 2:49 am

Hi, I am a new investor, but this is not an analysis of my portfolio. I am invested in this ETF and have ran FRA on it many times in portfoliovisualizer. However, I get different returns with different settings and I don't know which to trust. This is part of my "forever" portfolio so I want to be sure about it. I want to know the theoretical premiums of this ETF, but again, do not know which analysis to trust.

Here is a link to PortfolioVisualizer, where I have used 3-factor model on this ETF and using international developed markets:
https://www.portfoliovisualizer.com/fac ... ssion=true

Can I calculate the theoretical premiums like this?
SMB: 0.36 * (SMBpremium) + 0.64 * (HMLpremium).
Is there any consensus in bogleheads, which premium values should I use in this calculation?
Should I use five or four factor model, and which values of premiums are considered to be okay?

Thank you.

muffins14
Posts: 196
Joined: Wed Oct 26, 2016 4:14 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by muffins14 » Wed May 13, 2020 11:13 am

If you were to assume the premia were zero but the volatility was as observed over the last 10 years, could you run some analysis to make sure you would hit your financial goals?

To me, that seems to be the more conservative approach, with the upside that if the premia do occur over your investing timeline, then you will realize the benefits from your strategy. The bad alternative would be to assume premia that are too large (compared to the unknowable future reality), and have your plan's success be dependent upon hitting those expectations.

thermalfama
Posts: 39
Joined: Sat Nov 26, 2011 9:38 pm

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by thermalfama » Wed May 13, 2020 3:24 pm

VVL.TO is an actively managed value fund. Have you considered the disadvantages of actively managed funds? I invest in IWVL, iShares Edge MSCI World Value Factor UCITS ETF USD (Acc), which targets a very similar set of companies, but is passively managed. If I recall correctly, VVL is much more heavy on financials than IWVL.
As for your question regarding factors, sorry I can't answer. I'm not sure I would worry too much what the output is from the factor analysis tool, since it cannot predict the future, and the active managers in VVL are free to change their decision making any time. Plus, I think VVL had an inception date only a few years back, which would make its past pretty useless in predicting future factor returns.

M1sup
Posts: 1
Joined: Thu May 14, 2020 2:02 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by M1sup » Thu May 14, 2020 2:39 am

I did a lot of research about this fund as I am also interested in investing in it.

A few points I found:

1.yes, the fund is active, but Vanguard states this is for the purpose that it can be rebalanced every day and have a higher factor loading this way

1.they have rules / a model how they determine the underlying securities: they rate all stocks in the ftse index and weigh them according to their value, then they take the top 30%

3.this way the fund has significant factor loadings for size and value(I checked by uploading the data to portfoliovisualizer, Fama French 5 factors, but also with the other facfors you get quite a positive loading)

4.you have small cap value, which adds diversification as a world-index has no small cap, also small cap value has other benefits

5.the fund is diversified with over 1200 securities compared to the few that are in the Ishares world value, also the loading of value is less in the ishares value fund and it has no small cap value. I found the factor loading of the Ishares value fund to be not really that great, it is significant for value , but does it make a difference for the amount of diversification you give up (398 holdings)?

6.one thing I also considered is the turnover ratio of the fund, according to Vanguard it is 37% at the moment, which is okay, I couldn't find earlier numbers though. You can look up the costs for turnover in a document of them, as it is EU-regulation that all costs are stated. At 37% its costs about 0.22% extra per year for the trading involved. But even if it should go higher you can see that even the active momentum fund with 192% turnover only costs 0.39% extra. Which seems okay for me.

I am looking forward to opinions to this :happy

Valuethinker
Posts: 40006
Joined: Fri May 11, 2007 11:07 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by Valuethinker » Thu May 14, 2020 3:39 am

thermalfama wrote:
Wed May 13, 2020 3:24 pm
VVL.TO is an actively managed value fund. Have you considered the disadvantages of actively managed funds? I invest in IWVL, iShares Edge MSCI World Value Factor UCITS ETF USD (Acc), which targets a very similar set of companies, but is passively managed. If I recall correctly, VVL is much more heavy on financials than IWVL.
As for your question regarding factors, sorry I can't answer. I'm not sure I would worry too much what the output is from the factor analysis tool, since it cannot predict the future, and the active managers in VVL are free to change their decision making any time. Plus, I think VVL had an inception date only a few years back, which would make its past pretty useless in predicting future factor returns.
I believe that the Vanguard Fund in question is "active" in that it uses quantitative strategies to maximize exposure to value factor, but seeking to minimize costs like holding very illiquid shares.
The Fund pursues an actively-managed investment strategy whereby the investment manager has discretion over the composition of the Fund’s portfolio holdings. The Fund will seek to achieve its investment objective by investing primarily in equity securities.

In seeking to achieve the Fund’s investment objective, the investment manager uses a proprietary quantitative model to evaluate an investment universe comprised of large, mid and small cap equity securities from developed markets across the world, which includes a diverse representation of companies, market sectors and industry groups. This investment universe is drawn primarily from equity securities included in the FTSE Developed All Cap Index and the Russell 3000 Index (the “Indices”) (the “Investment Universe”).

While the investment manager may adopt constraints, depending on market conditions, as to the extent to which the Fund’s holdings deviate from the constituents of the Indices on a regional, country, sector, industry group and individual stock level, the extent to which the portfolio holdings may deviate from the Indices may be significant. The investment manager has discretion to apply, change, and remove any such constraints, which may limit the extent to which the Fund can outperform the Indices.

The investment manager’s quantitative model implements a rules-based active approach that aims to assess the factor exposures of securities, favouring equity securities which, when compared to other securities in the Investment Universe, have lower prices relative to their fundamental measures of value (which measures may include price-to-book or price-to-earnings ratio, estimated future earnings and operating cash flow). This measure (the “Value Factor”) has been shown to be a component of long-run stock market returns.
This does not seem, to me, to be too far from a quant indexation strategy. If you look at the country and sector weights the country weights are quite close to actual percentages for those countries on a market cap weighted basis.

I hold this one, and the ishares equivalent IWVL in a different account.

The ishares one is much more aggressively tilted - something like 20% underweight USA and double-weighted in Japan. In terms of performance against world indices, VVL is likely to be much closer, I believe - meaning both upside and downside.

VVL has underperformed the equivalent Vanguard global index fund by something like 30% since I bought it in December 2018 ;-). Gulp. Well, stay the course ;-).
Last edited by Valuethinker on Thu May 14, 2020 3:43 am, edited 1 time in total.

Valuethinker
Posts: 40006
Joined: Fri May 11, 2007 11:07 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by Valuethinker » Thu May 14, 2020 3:41 am

ilarion wrote:
Wed May 13, 2020 2:49 am
Hi, I am a new investor, but this is not an analysis of my portfolio. I am invested in this ETF and have ran FRA on it many times in portfoliovisualizer. However, I get different returns with different settings and I don't know which to trust. This is part of my "forever" portfolio so I want to be sure about it. I want to know the theoretical premiums of this ETF, but again, do not know which analysis to trust.

Here is a link to PortfolioVisualizer, where I have used 3-factor model on this ETF and using international developed markets:
https://www.portfoliovisualizer.com/fac ... ssion=true

Can I calculate the theoretical premiums like this?
SMB: 0.36 * (SMBpremium) + 0.64 * (HMLpremium).
Is there any consensus in bogleheads, which premium values should I use in this calculation?
Should I use five or four factor model, and which values of premiums are considered to be okay?

Thank you.
I do not think there is any consensus as to what are the "right" factor loadings.

You probably read Larry Swedroe's various pieces on the web? He's thought about this a lot. I don't necessarily agree with him in all cases, but it's worth reading everything he has written on the subject (he used to post here, but does not do so much any more).

Topic Author
ilarion
Posts: 3
Joined: Tue May 12, 2020 1:25 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by ilarion » Mon May 18, 2020 5:38 am

Many thanks to every replier! I got lots of valuable info. My favorite was the turnover: 1-2 weeks ago morningstar reported it to be 72%, but here I heard it was lower. Do you happen to have a link to the turnover caused expenses, which you said to be in range of 0,2-0,3%?

Indeed even though this is an 'active' fund, in reality it follows rules based strategies very close to other passive factor funds. I would not worry about that.

Good tip was to assume premia to be 0, and consider if this is still worthy investment for me. And I would say not, honestly. The reason is that dividends get taxed, and in my country i can not deduct them later. Even though ireland domiciled cuts it half. This fund has currently high div yield. I calculate it to add 0,6% extra fees, where as market index fund would add me 0,3%. So i need the premia to be larger than div yield tax / 2 (to compare the disadvantage to market index) + turnover costs. I do believe that the premia will be at least that, if not more, but ofc, cant be sure.

Angst
Posts: 2307
Joined: Sat Jun 09, 2007 11:31 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by Angst » Mon May 18, 2020 9:22 am

US investors should know that this ETF is from Vanguard Canada. Although it's investment rules and implementation appear to be in keeping with the Vanguard factor ETFs provided in the US, we don't have access to it. I wish we did...

https://www.vanguardcanada.ca/advisors/products/en/detail/etf/9795/equity

Code: Select all

Country      Region     Fund

Code: Select all

US           N.America  64.6%
Japan        Pacific     8.4%
UK           Europe      5.1%
France       Europe      3.5%
Korea        Pacific     3.3%
Canada       N.America   3.0%
Germany      Europe      2.2%
Italy        Europe      1.6%
Hong Kong    Pacific     1.2%
Netherlands  Europe      1.0%
Spain        Europe      1.0%
Sweden       Europe      0.8%
Switzerland  Europe      0.8%
Poland       Europe      0.7%
Austria      Europe      0.6%
Australia    Pacific     0.5%
Belgium      Europe      0.5%
Denmark      Europe      0.4%
Finland      Europe      0.2%
Ireland      Europe      0.2%
Norway       Europe      0.2%
Singapore    Pacific     0.2%
Israel       Middle East  —
New Zealand  Pacific      —
Portugal     Europe       —

Topic Author
ilarion
Posts: 3
Joined: Tue May 12, 2020 1:25 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by ilarion » Mon May 18, 2020 9:54 am

Yes, vvl.to is the canadian version. I used that ticker because it can be used in PV.

However, EU investors have the same fund to be invested with eur or gbp.

For example i invest with ticker VGVL.

IFYOUCAN
Posts: 29
Joined: Mon Dec 24, 2018 10:48 pm

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by IFYOUCAN » Wed May 20, 2020 3:05 am

The 30% turnover could be due to recent bear market, why would they cut off certain stocks if they present even better value when the bear strikes, expect a higher turnover when the fund goes way up in value.

actuallyxy
Posts: 55
Joined: Wed May 24, 2017 6:25 am

Re: Factor regression analysis on Vanguard Global Value Factor ETF (vvl.to)

Post by actuallyxy » Thu May 21, 2020 9:25 pm

Hi,

It is interesting to look at the Vanguard and iShares ETFs in Morningstar, and compare them with a market cap weighted global fund.

The Vanguard ETF sticks to the global country allocations, but drifts to a significantly different sector allocation.

The iShares ETF sticks very closely to the global sector allocation, but drifts to a very different country allocation.

Post Reply