Avantis Funds: what's so special about them?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
JSPECO9
Posts: 458
Joined: Thu Nov 24, 2016 11:34 pm

Re: Avantis Funds: what's so special about them?

Post by JSPECO9 »

comeinvest wrote: Thu Mar 30, 2023 5:38 pm
rkhusky wrote: Thu Mar 30, 2023 9:23 am There is a less than perfect correlation between LV, LG, SV, and SG. So, for maximum diversity one should own all of them. TSM is the easy way to do that. Although one could also use four funds at various proportions.
"So, for maximum diversity one should own all of them." - That is not what the factor "experts" are trying to make us believe. They want to convince us that "value" is a source of return, and "growth" is a source of negative return. In the long run, or in the sense of "expected" returns, of course; not realized returns.

Rebalancing the individual components would theoretically create some excess return, if we believe in mean reversion. I'm sure there is a study quantifying this.
I don't think that's true, maybe someone with more expertise can correct me. Growth stocks are still risky and warrant a risk premium over bonds. It's just value stocks are riskier and warrant a premium over growth stocks.

Someone feel free to correct me.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

JSPECO9 wrote: Thu Mar 30, 2023 9:12 pm
I don't think that's true, maybe someone with more expertise can correct me. Growth stocks are still risky and warrant a risk premium over bonds. It's just value stocks are riskier and warrant a premium over growth stocks.

Someone feel free to correct me.
I think that's right. I've never seen growth characterized as having negative expected return. It's just that value is expected to earn a premium relative to growth. I suspect both would be expected to earn a positive premium over t bills because both are substantially exposed to market risk
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Avantis Funds: what's so special about them?

Post by comeinvest »

km91 wrote: Thu Mar 30, 2023 9:56 pm
JSPECO9 wrote: Thu Mar 30, 2023 9:12 pm
I don't think that's true, maybe someone with more expertise can correct me. Growth stocks are still risky and warrant a risk premium over bonds. It's just value stocks are riskier and warrant a premium over growth stocks.

Someone feel free to correct me.
I think that's right. I've never seen growth characterized as having negative expected return. It's just that value is expected to earn a premium relative to growth. I suspect both would be expected to earn a positive premium over t bills because both are substantially exposed to market risk
That's exactly what I meant, of course.
But the key here is that the factor "experts" try to convince us that there is a relatively uncorrelated component of return in "value", and/or that part of the excess return is risk-adjusted excess return. The key for whether or not tilting to "value" is worth it, is whether at least one of these 2 things is true.
ivgrivchuck
Posts: 1672
Joined: Sun Sep 27, 2020 6:20 pm

Re: Avantis Funds: what's so special about them?

Post by ivgrivchuck »

comeinvest wrote: Thu Mar 30, 2023 10:17 pm
km91 wrote: Thu Mar 30, 2023 9:56 pm
JSPECO9 wrote: Thu Mar 30, 2023 9:12 pm
I don't think that's true, maybe someone with more expertise can correct me. Growth stocks are still risky and warrant a risk premium over bonds. It's just value stocks are riskier and warrant a premium over growth stocks.

Someone feel free to correct me.
I think that's right. I've never seen growth characterized as having negative expected return. It's just that value is expected to earn a premium relative to growth. I suspect both would be expected to earn a positive premium over t bills because both are substantially exposed to market risk
That's exactly what I meant, of course.
But the key here is that the factor "experts" try to convince us that there is a relatively uncorrelated component of return in "value", and/or that part of the excess return is risk-adjusted excess return. The key for whether or not tilting to "value" is worth it, is whether at least one of these 2 things is true.
Time will tell. Value stocks are riskier than growth stocks, so their expected return is higher. Nobody seriously disputes that. The real question is if value stocks provide better risk-adjusted returns, and that can be true only if indeed there are at least two uncorrelated components of return: "value risk" and "market risk".
25% VTI | 25% VXUS | 12.5% AVUV | 10% AVDV | 2.5% VWO | 25% BND/SCHR/SCHP
rushrocker
Posts: 248
Joined: Wed Dec 28, 2022 11:54 am

Re: Avantis Funds: what's so special about them?

Post by rushrocker »

There is some benefit to holding growth in taxable and value in tax deferred, if you are up for that challenge. You can probably gain about 0.10-0.15 bp per year, if you do it perfectly in equal weights, since the taxable dividends are often much lower for growth funds.

Otherwise, it seems to me you should pick one or the other, tilted from a total market fund.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

comeinvest wrote: Thu Mar 30, 2023 10:17 pm That's exactly what I meant, of course.
But the key here is that the factor "experts" try to convince us that there is a relatively uncorrelated component of return in "value", and/or that part of the excess return is risk-adjusted excess return. The key for whether or not tilting to "value" is worth it, is whether at least one of these 2 things is true.
Risk-adjusted return is a bit of a red herring. If the factors represent true risks separate from the market factor we should expect them to have excess volatility adjusted returns. The factors supposedly capture dimensions of risk that aren't beta/market correlation. The risk is there, the Sharpe ratio just doesn't see it. I wouldn't say the equity factors are uncorrelated from each other or the market factor, but certainly the correlations are less than one. The factors do seem to have a contrarian bend to them so I suspect the "uncorrelated source of return" story is mostly true
Apathizer
Posts: 2507
Joined: Sun Sep 26, 2021 2:56 pm

Re: Avantis Funds: what's so special about them?

Post by Apathizer »

km91 wrote: Thu Mar 30, 2023 10:46 pm
comeinvest wrote: Thu Mar 30, 2023 10:17 pm That's exactly what I meant, of course.
But the key here is that the factor "experts" try to convince us that there is a relatively uncorrelated component of return in "value", and/or that part of the excess return is risk-adjusted excess return. The key for whether or not tilting to "value" is worth it, is whether at least one of these 2 things is true.
Risk-adjusted return is a bit of a red herring. If the factors represent true risks separate from the market factor we should expect them to have excess volatility adjusted returns. The factors supposedly capture dimensions of risk that aren't beta/market correlation. The risk is there, the Sharpe ratio just doesn't see it. I wouldn't say the equity factors are uncorrelated from each other or the market factor, but certainly the correlations are less than one. The factors do seem to have a contrarian bend to them so I suspect the "uncorrelated source of return" story is mostly true
Both correlation and dispersion. Assets can be closely correlated, but with a wide dispersion. This seems to be the case with factors and total market. For instance, while small value is closely correlated with the total market there's a wide dispersion. Some days the total market might rise 0.5% while small value rises 1.5%. And of course the inverse can be true as well.

This dispersion is an important aspect of diversification. For instance in 2022 both the market and small value had negative returns, but small values losses were much lower. This is an example of how an asset can be closely correlated with the market but still provide important diversification benefits.
ROTH: 50% AVGE, 10% DFAX, 40% BNDW. Taxable: 50% BNDW, 40% AVGE, 10% DFAX.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Avantis Funds: what's so special about them?

Post by comeinvest »

Apathizer wrote: Fri Mar 31, 2023 12:29 am
km91 wrote: Thu Mar 30, 2023 10:46 pm
comeinvest wrote: Thu Mar 30, 2023 10:17 pm That's exactly what I meant, of course.
But the key here is that the factor "experts" try to convince us that there is a relatively uncorrelated component of return in "value", and/or that part of the excess return is risk-adjusted excess return. The key for whether or not tilting to "value" is worth it, is whether at least one of these 2 things is true.
Risk-adjusted return is a bit of a red herring. If the factors represent true risks separate from the market factor we should expect them to have excess volatility adjusted returns. The factors supposedly capture dimensions of risk that aren't beta/market correlation. The risk is there, the Sharpe ratio just doesn't see it. I wouldn't say the equity factors are uncorrelated from each other or the market factor, but certainly the correlations are less than one. The factors do seem to have a contrarian bend to them so I suspect the "uncorrelated source of return" story is mostly true
Both correlation and dispersion. Assets can be closely correlated, but with a wide dispersion. This seems to be the case with factors and total market. For instance, while small value is closely correlated with the total market there's a wide dispersion. Some days the total market might rise 0.5% while small value rises 1.5%. And of course the inverse can be true as well.

This dispersion is an important aspect of diversification. For instance in 2022 both the market and small value had negative returns, but small values losses were much lower. This is an example of how an asset can be closely correlated with the market but still provide important diversification benefits.
I think positive expected return of the factor (on top of the market risk) is a requirement, though. Correlation less than 1 is not enough. Otherwise you could just split the 500 constituents of the index into two parts by some arbitrary criterion like industry, which would not be sensical.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

Apathizer wrote: Thu Mar 30, 2023 8:07 pm
rkhusky wrote: Thu Mar 30, 2023 7:32 pm
Apathizer wrote: Thu Mar 30, 2023 5:55 pm
rkhusky wrote: Thu Mar 30, 2023 9:23 am There is a less than perfect correlation between LV, LG, SV, and SG. So, for maximum diversity one should own all of them. TSM is the easy way to do that. Although one could also use four funds at various proportions.
What exactly do you mean own all of them? If you believe nothing is correlated and owning everything is beneficial, that would seem to me in equal weight index rather than market cap weights.

You don't really get the benefit of imperfect correlation by only owning a modicum of stocks that are not large caps, and large growth especially. Getting diversification benefits from small value stocks means owning them in weights higher than they exist in the market.
Equal weights for the four sectors is certainly reasonable.
You realize owning everything in equal weight with regular re-balancing would essentially mean 0% long term returns?
No, can you show the math for that?
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

km91 wrote: Thu Mar 30, 2023 9:56 pm
JSPECO9 wrote: Thu Mar 30, 2023 9:12 pm
I don't think that's true, maybe someone with more expertise can correct me. Growth stocks are still risky and warrant a risk premium over bonds. It's just value stocks are riskier and warrant a premium over growth stocks.

Someone feel free to correct me.
I think that's right. I've never seen growth characterized as having negative expected return. It's just that value is expected to earn a premium relative to growth. I suspect both would be expected to earn a positive premium over t bills because both are substantially exposed to market risk
Growth absolutely has a negative expected return, because value has a positive expected return. It is the same variable in the model.

That doesn’t mean all growth stocks are expected to have negative return, because the market factor has a positive expected return, and a portfolio of growth stocks has loading on market beta
Crom laughs at your Four Winds
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

muffins14 wrote: Fri Mar 31, 2023 8:28 am Growth absolutely has a negative expected return, because value has a positive expected return. It is the same variable in the model.
The stock market is not a zero sum game. Growth and value stocks have both had positive returns over the years.

How are you calculating the expected return?

Note that expected returns of growth, value, etc are inputs to the factor model, not outputs.
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

rkhusky wrote: Fri Mar 31, 2023 9:05 am
muffins14 wrote: Fri Mar 31, 2023 8:28 am Growth absolutely has a negative expected return, because value has a positive expected return. It is the same variable in the model.
The stock market is not a zero sum game. Growth and value stocks have both had positive returns over the years.

How are you calculating the expected return?

Note that expected returns of growth, value, etc are inputs to the factor model, not outputs.
Please see the second paragraph of my post:

“That doesn’t mean all growth stocks are expected to have negative return, because the market factor has a positive expected return, and a portfolio of growth stocks has loading on market beta”

Market beta = positive expected return
Value factor coefficient HmL = positive expected return.

value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.

A portfolio of growth stocks likely has market beta ~ 1 and HmL < 0. Therefore is has positive expected return, but lower expected return than a value portfolio, which has market beta ~1 but HmL > 0
Crom laughs at your Four Winds
Topic Author
JSPECO9
Posts: 458
Joined: Thu Nov 24, 2016 11:34 pm

Re: Avantis Funds: what's so special about them?

Post by JSPECO9 »

muffins14 wrote: Fri Mar 31, 2023 9:12 am
rkhusky wrote: Fri Mar 31, 2023 9:05 am
muffins14 wrote: Fri Mar 31, 2023 8:28 am Growth absolutely has a negative expected return, because value has a positive expected return. It is the same variable in the model.
The stock market is not a zero sum game. Growth and value stocks have both had positive returns over the years.

How are you calculating the expected return?

Note that expected returns of growth, value, etc are inputs to the factor model, not outputs.
Please see the second paragraph of my post:

“That doesn’t mean all growth stocks are expected to have negative return, because the market factor has a positive expected return, and a portfolio of growth stocks has loading on market beta”

Market beta = positive expected return
Value factor coefficient HmL = positive expected return.

value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.

A portfolio of growth stocks likely has market beta ~ 1 and HmL < 0. Therefore is has positive expected return, but lower expected return than a value portfolio, which has market beta ~1 but HmL > 0
Then what's the point of holding the market portfolio if an all-value portfolio still likely also has market beta ~ 1? What would be the point of including growth stocks?
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

Because sometimes the value portfolio has negative returns relative to a growth portfolio.

The degree to which one’s portfolio returns different than the total market can create anxiety and second-guessing, at best, and at worst, value could have negative returns for 10-20 years and a growth portfolio would have been better.

Maybe then not going for market beta = 1 and HmL = 1 makes sense, and something of a more moderate tilt is better.

As Taylor says, would you want to ignore the biggest/best companies completely? I’m sure google, Microsoft, Tesla, Facebook, Netflix, Amazon etc have helped my returns since 2015. I feel good owning a bunch of VTI instead of just leaving it up to reliance on correlations
Crom laughs at your Four Winds
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
I don't think this part is right. HmL is high minus low. The value premium is the performance spread of value over growth. Both could have positive expected returns and still produce a positive value premium. Value could have an expected return of 11% and growth could have an expected return of 9% for example. The value premium would be 2% and both value and growth would be expected to have positive returns
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
That’s not true.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 10:37 am
muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
I don't think this part is right. HmL is high minus low. The value premium is the performance spread of value over growth. Both could have positive expected returns and still produce a positive value premium. Value could have an expected return of 11% and growth could have an expected return of 9% for example. The value premium would be 2% and both value and growth would be expected to have positive returns
Or expected return of growth stocks could be 10% and the expected return of value stocks could be 5%, resulting in a value premium of -5%.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 10:42 am Or expected return of growth stocks could be 10% and the expected return of value stocks could be 5%, resulting in a value premium of -5%.
If the data and factor models are to be believed the expected value premium should be positive over the long-term. The realized value premium over any holding period could be positive or negative though
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 10:57 am
rkhusky wrote: Fri Mar 31, 2023 10:42 am Or expected return of growth stocks could be 10% and the expected return of value stocks could be 5%, resulting in a value premium of -5%.
If the data and factor models are to be believed the expected value premium should be positive over the long-term. The realized value premium over any holding period could be positive or negative though
The factor model is agnostic as to whether premiums are positive or negative. Computing expected returns is performed outside the model and then input to the model.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 11:01 am Computing expected returns is performed outside the model and then input to the model.
The model uses actual returns as an input and the output tells us about a historical relationship that existed in the past. From this we make an inference about the future
Apathizer
Posts: 2507
Joined: Sun Sep 26, 2021 2:56 pm

Re: Avantis Funds: what's so special about them?

Post by Apathizer »

rkhusky wrote: Fri Mar 31, 2023 8:25 am
Apathizer wrote: Thu Mar 30, 2023 8:07 pm
rkhusky wrote: Thu Mar 30, 2023 7:32 pm
Apathizer wrote: Thu Mar 30, 2023 5:55 pm
rkhusky wrote: Thu Mar 30, 2023 9:23 am There is a less than perfect correlation between LV, LG, SV, and SG. So, for maximum diversity one should own all of them. TSM is the easy way to do that. Although one could also use four funds at various proportions.
What exactly do you mean own all of them? If you believe nothing is correlated and owning everything is beneficial, that would seem to me in equal weight index rather than market cap weights.

You don't really get the benefit of imperfect correlation by only owning a modicum of stocks that are not large caps, and large growth especially. Getting diversification benefits from small value stocks means owning them in weights higher than they exist in the market.
Equal weights for the four sectors is certainly reasonable.
You realize owning everything in equal weight with regular re-balancing would essentially mean 0% long term returns?
No, can you show the math for that?
Sure. There's an article below. Basically, a small portion of stocks have high returns while the majority have low or negative returns. Only about 42% of all common stocks have returns higher the one-month treasury bills. Nearly 60% of all stocks have flat or negative returns.

So if most stocks have flat or negative returns and you hold all stocks in equal weight with regular re-balancing, most of your holdings will have flat or negative returns. This will be countered by the few with positive returns, but this will be a small portion of your entire portfolio, so returns will be very low if not zero.

Incidentally, an equal weight index with regular re-balancing would be extraordinary expensive, complicated, and inefficient. Since stock prices constantly fluctuate, an equal weight index would need to constantly buy and sell all stocks to maintain equal weight, so it wouldn't be a practical approach to portfolio construction.
https://www.sciencedirect.com/science/a ... 5X18301521
ROTH: 50% AVGE, 10% DFAX, 40% BNDW. Taxable: 50% BNDW, 40% AVGE, 10% DFAX.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 11:10 am
rkhusky wrote: Fri Mar 31, 2023 11:01 am Computing expected returns is performed outside the model and then input to the model.
The model uses actual returns as an input and the output tells us about a historical relationship that existed in the past. From this we make an inference about the future
In other words.
The modeling process uses actual returns of the factors and of a portfolio (or single investment) to train a model for that portfolio, i.e.determine the free parameters associated with the portfolio. Once the model is determined, it accepts values for the factors as input and outputs the return of the portfolio associated with those factor values.

One can input the training data to the model to determine the quality of the model. Or you can input computed values of future factor returns to predict future returns for the portfolio.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

Apathizer wrote: Fri Mar 31, 2023 11:44 am
rkhusky wrote: Fri Mar 31, 2023 8:25 am
Apathizer wrote: Thu Mar 30, 2023 8:07 pm
rkhusky wrote: Thu Mar 30, 2023 7:32 pm
Apathizer wrote: Thu Mar 30, 2023 5:55 pm
What exactly do you mean own all of them? If you believe nothing is correlated and owning everything is beneficial, that would seem to me in equal weight index rather than market cap weights.

You don't really get the benefit of imperfect correlation by only owning a modicum of stocks that are not large caps, and large growth especially. Getting diversification benefits from small value stocks means owning them in weights higher than they exist in the market.
Equal weights for the four sectors is certainly reasonable.
You realize owning everything in equal weight with regular re-balancing would essentially mean 0% long term returns?
No, can you show the math for that?
Sure. There's an article below. Basically, a small portion of stocks have high returns while the majority have low or negative returns. Only about 42% of all common stocks have returns higher the one-month treasury bills. Nearly 60% of all stocks have flat or negative returns.

So if most stocks have flat or negative returns and you hold all stocks in equal weight with regular re-balancing, most of your holdings will have flat or negative returns. This will be countered by the few with positive returns, but this will be a small portion of your entire portfolio, so returns will be very low if not zero.

Incidentally, an equal weight index with regular re-balancing would be extraordinary expensive, complicated, and inefficient. Since stock prices constantly fluctuate, an equal weight index would need to constantly buy and sell all stocks to maintain equal weight, so it wouldn't be a practical approach to portfolio construction.
https://www.sciencedirect.com/science/a ... 5X18301521
Sorry, but I was speaking of equal weighting four funds covering the LG, LV, SG, SV sectors. I think TSM is easier, but could understand why someone might try the other to get more bang out of the different correlations.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 12:03 pm In other words.
The modeling process uses actual returns of the factors and of a portfolio (or single investment) to train a model for that portfolio, i.e.determine the free parameters associated with the portfolio. Once the model is determined, it accepts values for the factors as input and outputs the return of the portfolio associated with those factor values.

One can input the training data to the model to determine the quality of the model. Or you can input computed values of future factor returns to predict future returns for the portfolio.
I think that's right. I guess if you're inputting your own computed values for future factor returns into the model the question is where did those values come from, if not from the model itself and the historical data? If I compute expected returns where value is less than growth is that really a relationship that the historical data describes?
GP813
Posts: 1243
Joined: Wed Dec 11, 2019 9:11 am

Re: Avantis Funds: what's so special about them?

Post by GP813 »

Apathizer wrote: Fri Mar 31, 2023 11:44 am
rkhusky wrote: Fri Mar 31, 2023 8:25 am
Apathizer wrote: Thu Mar 30, 2023 8:07 pm
rkhusky wrote: Thu Mar 30, 2023 7:32 pm
Apathizer wrote: Thu Mar 30, 2023 5:55 pm
What exactly do you mean own all of them? If you believe nothing is correlated and owning everything is beneficial, that would seem to me in equal weight index rather than market cap weights.

You don't really get the benefit of imperfect correlation by only owning a modicum of stocks that are not large caps, and large growth especially. Getting diversification benefits from small value stocks means owning them in weights higher than they exist in the market.
Equal weights for the four sectors is certainly reasonable.
You realize owning everything in equal weight with regular re-balancing would essentially mean 0% long term returns?
No, can you show the math for that?
Sure. There's an article below. Basically, a small portion of stocks have high returns while the majority have low or negative returns. Only about 42% of all common stocks have returns higher the one-month treasury bills. Nearly 60% of all stocks have flat or negative returns.

So if most stocks have flat or negative returns and you hold all stocks in equal weight with regular re-balancing, most of your holdings will have flat or negative returns. This will be countered by the few with positive returns, but this will be a small portion of your entire portfolio, so returns will be very low if not zero.

Incidentally, an equal weight index with regular re-balancing would be extraordinary expensive, complicated, and inefficient. Since stock prices constantly fluctuate, an equal weight index would need to constantly buy and sell all stocks to maintain equal weight, so it wouldn't be a practical approach to portfolio construction.
https://www.sciencedirect.com/science/a ... 5X18301521

Look at the S&P 500 YTD for an example. Almost all the gains have come from the largest "tech" heavy companies.

https://finviz.com/map.ashx?t=sec&st=ytd
Apathizer
Posts: 2507
Joined: Sun Sep 26, 2021 2:56 pm

Re: Avantis Funds: what's so special about them?

Post by Apathizer »

rkhusky wrote: Fri Mar 31, 2023 12:09 pm
Apathizer wrote: Fri Mar 31, 2023 11:44 am
rkhusky wrote: Fri Mar 31, 2023 8:25 am
Apathizer wrote: Thu Mar 30, 2023 8:07 pm
rkhusky wrote: Thu Mar 30, 2023 7:32 pm
Equal weights for the four sectors is certainly reasonable.
You realize owning everything in equal weight with regular re-balancing would essentially mean 0% long term returns?
No, can you show the math for that?
Sure. There's an article below. Basically, a small portion of stocks have high returns while the majority have low or negative returns. Only about 42% of all common stocks have returns higher the one-month treasury bills. Nearly 60% of all stocks have flat or negative returns.

So if most stocks have flat or negative returns and you hold all stocks in equal weight with regular re-balancing, most of your holdings will have flat or negative returns. This will be countered by the few with positive returns, but this will be a small portion of your entire portfolio, so returns will be very low if not zero.

Incidentally, an equal weight index with regular re-balancing would be extraordinary expensive, complicated, and inefficient. Since stock prices constantly fluctuate, an equal weight index would need to constantly buy and sell all stocks to maintain equal weight, so it wouldn't be a practical approach to portfolio construction.
https://www.sciencedirect.com/science/a ... 5X18301521
Sorry, but I was speaking of equal weighting four funds covering the LG, LV, SG, SV sectors. I think TSM is easier, but could understand why someone might try the other to get more bang out of the different correlations.
Equally weighting all 9 general sectors would have increased total return over MCW, but tilting to factors would have increased returns even more. Small growth as a whole has had poor returns, so if anything reducing allocation makes sense.
ROTH: 50% AVGE, 10% DFAX, 40% BNDW. Taxable: 50% BNDW, 40% AVGE, 10% DFAX.
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

rkhusky wrote: Fri Mar 31, 2023 10:38 am
muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
That’s not true.
Yes, it is. That is how the model is constructed

It is like saying lower is the opposite of higher. North is the opposite of south, down is the opposite of up
Last edited by muffins14 on Fri Mar 31, 2023 1:18 pm, edited 1 time in total.
Crom laughs at your Four Winds
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

km91 wrote: Fri Mar 31, 2023 10:37 am
muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
I don't think this part is right. HmL is high minus low. The value premium is the performance spread of value over growth. Both could have positive expected returns and still produce a positive value premium. Value could have an expected return of 11% and growth could have an expected return of 9% for example. The value premium would be 2% and both value and growth would be expected to have positive returns
please read my full post. I'm saying that the FACTOR HmL is implying that value has positive return and thus growth has negative return

HmL is indeed high minus low.

a growth portfolio has HML < 0
a value portfolio has HML > 0

More HmL => more expected return, because the HML coefficient in the model has a positive coefficient.

That does now mean growth STOCKS generally have negative expected return. The return of a portfolio is like return = beta_mkt * mkt + beta_hml * hml + beta_smb * smb + ...

Even though your growth portfolio might have hml < 0, the fact that beta_mkt > 0 means you still get return, obviously
Crom laughs at your Four Winds
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
Crom laughs at your Four Winds
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

muffins14 wrote: Fri Mar 31, 2023 1:12 pm
rkhusky wrote: Fri Mar 31, 2023 10:38 am
muffins14 wrote: Fri Mar 31, 2023 9:12 am
value and growth are two ends of the same variable, so if value has positive expectation value, growth has negative expectation value.
That’s not true.
Yes, it is. That is how the model is constructed
The model factor is constructed as H minus L, but knowing H-L doesn’t tell you anything about H and L separately. 102-100 = 36-34 = 2347-2345 = 2.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

muffins14 wrote: Fri Mar 31, 2023 1:17 pm please read my full post. I'm saying that the FACTOR HmL is implying that value has positive return and thus growth has negative return

HmL is indeed high minus low.

a growth portfolio has HML < 0
a value portfolio has HML > 0

More HmL => more expected return, because the HML coefficient in the model has a positive coefficient.

That does now mean growth STOCKS generally have negative expected return. The return of a portfolio is like return = beta_mkt * mkt + beta_hml * hml + beta_smb * smb + ...

Even though your growth portfolio might have hml < 0, the fact that beta_mkt > 0 means you still get return, obviously
Got it, I see what you're saying. Yes, exposure to growth ie HML <0, should have negative expected return. The growth buyer is paying the premium, the value buyer is earning the premium
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: Avantis Funds: what's so special about them?

Post by abc132 »

I like that they looked at the data and are not trying to time the market in and out of certain factors. They provide a solid way to get exposure to various factors. I think their implementation of a strategy is good. Time will tell if the strategy pans out.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

muffins14 wrote: Fri Mar 31, 2023 1:21 pm These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
You mean basic terminology like expected return is the integral of return over the return probability distribution?
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

rkhusky wrote: Fri Mar 31, 2023 1:34 pm
muffins14 wrote: Fri Mar 31, 2023 1:21 pm These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
You mean basic terminology like expected return is the integral of return over the return probability distribution?
right - mathematical expectation value.

Obviously the distribution is unknown, but the factor models force a ~normal distribution for the average factor contribution to total return (since it is a linear regression)

I'm not saying that's the god-given return distribution, that's just what the model is.

All models are wrong, some are useful. In this case, if discussing factors, we should at least discuss the existing factor models and
Crom laughs at your Four Winds
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 12:09 pm
rkhusky wrote: Fri Mar 31, 2023 12:03 pm In other words.
The modeling process uses actual returns of the factors and of a portfolio (or single investment) to train a model for that portfolio, i.e.determine the free parameters associated with the portfolio. Once the model is determined, it accepts values for the factors as input and outputs the return of the portfolio associated with those factor values.

One can input the training data to the model to determine the quality of the model. Or you can input computed values of future factor returns to predict future returns for the portfolio.
I think that's right. I guess if you're inputting your own computed values for future factor returns into the model the question is where did those values come from, if not from the model itself and the historical data? If I compute expected returns where value is less than growth is that really a relationship that the historical data describes?
One can make the assumption that the future will look the same as the past, and use an average over some past time period. One can try to forecast future market or business conditions.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 1:34 pm
muffins14 wrote: Fri Mar 31, 2023 1:21 pm These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
You mean basic terminology like expected return is the integral of return over the return probability distribution?
But this is why a lot of investors find factor investing to be a bunch of academic hocus pocus. Leaning too heavily into the statistical conceptualization of the factors just confuses the issue more than anything. You don't need to have a Phd in stats to understand the intuition underpinning the factors. Yes, the statistics tell us a relationship exists in the data, we need to rely on other evidence to prove out that the relationship exists in the real world
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 1:58 pm
rkhusky wrote: Fri Mar 31, 2023 1:34 pm
muffins14 wrote: Fri Mar 31, 2023 1:21 pm These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
You mean basic terminology like expected return is the integral of return over the return probability distribution?
But this is why a lot of investors find factor investing to be a bunch of academic hocus pocus. Leaning too heavily into the statistical conceptualization of the factors just confuses the issue more than anything. You don't need to have a Phd in stats to understand the intuition underpinning the factors. Yes, the statistics tell us a relationship exists in the data, we need to rely on other evidence to prove out that the relationship exists in the real world
When practitioners take common words and use them for their jargon, it can cause issues. Take “premium” in factor language. In common parlance, premium means above average, but in the factor model it can be positive or negative. A negative premium seems like an oxymoron.

The same with “expected”. In common parlance, it can connate various levels of certainty, from “I expect to be rich someday.” to “I expected you at 7:00 sharp.”. When someone uses “expected return” on Bogleheads, should we assume they are using the term in the mathematical sense or the colloquial sense?
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

muffins14 wrote: Fri Mar 31, 2023 1:49 pm
rkhusky wrote: Fri Mar 31, 2023 1:34 pm
muffins14 wrote: Fri Mar 31, 2023 1:21 pm These threads are so frustrating because people just don't even agree on basic terminology or statistical concepts for what factors mean, so I don't know how any discussion is even possible
You mean basic terminology like expected return is the integral of return over the return probability distribution?
right - mathematical expectation value.

Obviously the distribution is unknown, but the factor models force a ~normal distribution for the average factor contribution to total return (since it is a linear regression)

I'm not saying that's the god-given return distribution, that's just what the model is.

All models are wrong, some are useful. In this case, if discussing factors, we should at least discuss the existing factor models and
Since stock returns aren’t normal, perhaps that’s a source of error in the factor models.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 3:23 pm
When practitioners take common words and use them for their jargon, it can cause issues. Take “premium” in factor language. In common parlance, premium means above average, but in the factor model it can be positive or negative. A negative premium seems like an oxymoron.

The same with “expected”. In common parlance, it can connate various levels of certainty, from “I expect to be rich someday.” to “I expected you at 7:00 sharp.”. When someone uses “expected return” on Bogleheads, should we assume they are using the term in the mathematical sense or the colloquial sense?
I think it's mostly semantics. "Insurance premium." "The electrician can get it done tomorrow but we'll have to pay a premium." The colloquially meaning applies. If you are the premium seller you receive positive premium, if you are the buyer you receive negative premium, we obviously don't think in these terms in everyday life but it's mostly straightforward

Expected means expected. You don't need a mathematical definition of expected to understand the intuition
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

In terms of using a factor model, suppose I apply the 3-factor FF model to a Growth index fund. I might find that beta_mkt is 1, beta_hml is -0.3 and beta_smb is -0.2. That means that if growth stocks do better than value stocks and big stocks do better than small stocks, the Growth index fund should do better than the Market. And vice-versa.
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 3:42 pm Expected means expected. You don't need a mathematical definition of expected to understand the intuition
Except where the intuition doesn’t match the mathematical, ie intuition says high degree of certainty and mathematics says low degree of certainty. The expected return of value stocks versus expected return of the market can be in that category.
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 3:54 pm Except where the intuition doesn’t match the mathematical, ie intuition says high degree of certainty and mathematics says low degree of certainty. The expected return of value stocks versus expected return of the market can be in that category.
"I expect the market to return 10%", "I expect value stocks to return 12%", "I expect value stocks to have higher returns than growth stocks". I think the intuition holds. Probabilistic expectations are another thing. But I doubt thinking in probabilities makes anyone's understanding clearer, it certainly doesn't for me
muffins14
Posts: 5529
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Avantis Funds: what's so special about them?

Post by muffins14 »

km91 wrote: Fri Mar 31, 2023 4:02 pm
rkhusky wrote: Fri Mar 31, 2023 3:54 pm Except where the intuition doesn’t match the mathematical, ie intuition says high degree of certainty and mathematics says low degree of certainty. The expected return of value stocks versus expected return of the market can be in that category.
"I expect the market to return 10%", "I expect value stocks to return 12%", "I expect value stocks to have higher returns than growth stocks". I think the intuition holds. Probabilistic expectations are another thing. But I doubt thinking in probabilities makes anyone's understanding clearer, it certainly doesn't for me
I think if you are talking about factors, you should *only* use the word expect to mean mathematical expectation. otherwise you're letting your biases speak, instead of deferring to the model of returns
Crom laughs at your Four Winds
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

muffins14 wrote: Fri Mar 31, 2023 4:15 pm I think if you are talking about factors, you should *only* use the word expect to mean mathematical expectation. otherwise you're letting your biases speak, instead of deferring to the model of returns
Probably but it's not a natural way of thinking for most people, myself included. At least for bogleheads, I think most of us realize that in the context of asset returns "expected" return doesn't mean our return is rolling in to the station any day now, and that there is a greater degree or uncertainty/unknowability than the colloquial sense of the word implies
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 4:02 pm
rkhusky wrote: Fri Mar 31, 2023 3:54 pm Except where the intuition doesn’t match the mathematical, ie intuition says high degree of certainty and mathematics says low degree of certainty. The expected return of value stocks versus expected return of the market can be in that category.
"I expect the market to return 10%", "I expect value stocks to return 12%", "I expect value stocks to have higher returns than growth stocks". I think the intuition holds. Probabilistic expectations are another thing. But I doubt thinking in probabilities makes anyone's understanding clearer, it certainly doesn't for me
And would your perception or investing strategy change if the width of the return distributions were +- 20% and +- 25% respectively (ie 10% +- 20% and 12% +- 25%).
km91
Posts: 1389
Joined: Wed Oct 13, 2021 12:32 pm

Re: Avantis Funds: what's so special about them?

Post by km91 »

rkhusky wrote: Fri Mar 31, 2023 4:40 pm And would your perception or investing strategy change if the width of the return distributions were +- 20% and +- 25% respectively (ie 10% +- 20% and 12% +- 25%).
Maybe maybe not. At some point you have to draw a line between what is signal and what is noise. Personally, I think the deeper we dive into the statistics, the more noise we're getting. I wouldn't outright dismiss the distribution data but the variance would have to be fairly extreme to overcome the other, non statistical evidence that I believe supports factor investing
rkhusky
Posts: 17764
Joined: Thu Aug 18, 2011 8:09 pm

Re: Avantis Funds: what's so special about them?

Post by rkhusky »

km91 wrote: Fri Mar 31, 2023 4:51 pm
rkhusky wrote: Fri Mar 31, 2023 4:40 pm And would your perception or investing strategy change if the width of the return distributions were +- 20% and +- 25% respectively (ie 10% +- 20% and 12% +- 25%).
Maybe maybe not. At some point you have to draw a line between what is signal and what is noise. Personally, I think the deeper we dive into the statistics, the more noise we're getting. I wouldn't outright dismiss the distribution data but the variance would have to be fairly extreme to overcome the other, non statistical evidence that I believe supports factor investing
Fair enough.
Gaston
Posts: 1220
Joined: Wed Aug 21, 2013 7:12 pm

Re: Avantis Funds: what's so special about them?

Post by Gaston »

JSPECO9 wrote: Mon Dec 12, 2022 12:16 pm Just finished reading "Your Complete Guide to Factor-Based Investing" by Larry Swedroe and Andrew Berkin.
I saw your reference to this book, bought it and just finished it. Really liked it, particularly the level of detail it contains.

Thanks for mentioning it.
“My opinions are just that - opinions.”
Post Reply