Factor ETFs For Diversification Or “Diworsification”

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
pdavi21
Posts: 1253
Joined: Sat Jan 30, 2016 4:04 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by pdavi21 » Mon May 20, 2019 4:38 pm

vineviz wrote:
Mon May 20, 2019 4:26 pm
pdavi21 wrote:
Mon May 20, 2019 4:05 pm
Because we all invest in long-short funds and individual securities on Bogleheads.org
Okay, but that doesn't magically make a flawed analysis useful.

It's possible to either make a well-reasoned analysis of the relative risk of factors or to refrain from characterizing that relative risk until you have to tools to do it correctly.
It magically makes the academic analysis useless.

Please tell me which value fund(s) you invest in and why you expect them to be measurably and meaningfully riskier than their growth counterparts going foreward.
"We spend a great deal of time studying history, which, let's face it, is mostly the history of stupidity." -Stephen Hawking

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Mon May 20, 2019 4:47 pm

pdavi21 wrote:
Mon May 20, 2019 4:38 pm
It magically makes the academic analysis useless.
What?

The statistical tools required to undertake a cross-sectional analysis of stock returns aren't rudimentary but they aren't SO complicated that an amateur investor can't get the gist of it if they try.

If you're going to claim that a well-established piece of financial economics is somehow flawed, I don't think it's not too much to ask that you at least attempt to understand how financial economics works.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

pdavi21
Posts: 1253
Joined: Sat Jan 30, 2016 4:04 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by pdavi21 » Mon May 20, 2019 6:08 pm

vineviz wrote:
Mon May 20, 2019 4:47 pm
pdavi21 wrote:
Mon May 20, 2019 4:38 pm
It magically makes the academic analysis useless.
What?

The statistical tools required to undertake a cross-sectional analysis of stock returns aren't rudimentary but they aren't SO complicated that an amateur investor can't get the gist of it if they try.

If you're going to claim that a well-established piece of financial economics is somehow flawed, I don't think it's not too much to ask that you at least attempt to understand how financial economics works.
The academic analysis is useless because funds aren't factors. So unless you invest in factors, the analysis on which factor is riskier is useless.

Which value fund(s) do you invest in that are expected to have higher risk than their growth counterparts? Still not getting an answer. Maybe just a fund you recommend?

EDIT: The flaw is more likely in the funds than the academic paper, but there is a flaw. That's why the majority of Value MF have had similar volatility (usually lower) over the last 20 years than their growth counterparts.

EDIT2: Let's talk about Beta risk too:
"diversification benefit" values from you:
"Vanguard S&P Small-Cap 600 Value ETF (VIOV): 1.04
Vanguard S&P Small-Cap 600 ETF (VIOO): 1.03
Vanguard S&P Small-Cap 600 Growth ETF (VIOG): 1.03
Vanguard Small-Cap Value ETF (VBR): 1.02
Vanguard Small-Cap Growth ETF (VBK): 1.02
Vanguard Small-Cap ETF (VB): 1.02"

Could that mean there is no discernible difference in Beta risk between value and growth too? Sure looks like it...or that Value has lower Beta risk for VIOV.
"We spend a great deal of time studying history, which, let's face it, is mostly the history of stupidity." -Stephen Hawking

Park
Posts: 682
Joined: Sat Nov 06, 2010 4:56 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Park » Mon May 20, 2019 7:15 pm

vineviz wrote:
Mon May 20, 2019 12:49 pm
Park wrote:
Mon May 20, 2019 12:29 pm
OTOH, how much does one want to invest in 3% of the total stock market? Although you're diversifying among factors in that 3%, there is a loss of diversification, when you ignore the other 97%.
This is a common misconception: diversification benefits accrue from correlation and variance, neither of which are dependent on market capitalization. A portfolio that is 50% large cap and 50% small cap is, empirically, much more diversified (not less) than a portfolio that is 97% large cap and 3% small cap.
vineviz wrote:
Mon May 20, 2019 4:22 pm
Basically, diversification is the process of spreading the total amount risk in a portfolio more evenly across the sources of risk. A well-diversified portfolio allocates its risk across many independent sources as feasible.
What about unknown risks? The only way I know of dealing with unknown risks is with diversification. And when you eliminate 97% of stocks, I think a good argument can be made that you're decreasing your ability to mitigate unknown risks.

Park
Posts: 682
Joined: Sat Nov 06, 2010 4:56 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Park » Mon May 20, 2019 7:33 pm

Lastrun wrote:
Mon May 20, 2019 7:39 am
typical.investor wrote:
Mon May 20, 2019 1:21 am


In any case, if size and value explain 90% of returns for diversified portfolios, then why is a portfolio (as suggested in this thread) of 25% size/value, 25% momentum, 25% low volatility and 25% quality called "diversification"?

Why have 75% of your portfolio in the factors that explain 10% of returns. I understand at times that the correlation is low, but there is nothing to say these factors aren't derived from common elements such as growth and inflation.
I think it may even be stronger than 90%--the wiki claims market, size and value "explain" 95% citing this Dimensional white paper by Fama https://www.ifa.com/academic-papers/mul ... a_2006.pdf

But I struggle with the question. No so much of a "how much to tilt" question, but how to achieve a sufficient diversification benefit with the least amount of tilt.

I also realize that the research is moving forward--heard Fama say once or twice that you could throw value out of their five-factor model and make it a four-factor model
Thanks for the posts. By multifactor investing, are you diversifying into factors that are of small relative importance? Does this emphasis on factors of small relative importance decrease one's ability to diversify using factors of greater relative importance?

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Mon May 20, 2019 8:11 pm

pdavi21 wrote:
Mon May 20, 2019 6:08 pm
The flaw is more likely in the funds than the academic paper, but there is a flaw.
I suspect the problem is more likely that you just don't understand the research well enough to implement it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

stlutz
Posts: 5465
Joined: Fri Jan 02, 2009 1:08 am

Re: Factor ETFs For Diversification Or “Diworsification”

Post by stlutz » Mon May 20, 2019 8:15 pm

Park wrote:
Mon May 20, 2019 7:33 pm
Lastrun wrote:
Mon May 20, 2019 7:39 am
typical.investor wrote:
Mon May 20, 2019 1:21 am


In any case, if size and value explain 90% of returns for diversified portfolios, then why is a portfolio (as suggested in this thread) of 25% size/value, 25% momentum, 25% low volatility and 25% quality called "diversification"?

Why have 75% of your portfolio in the factors that explain 10% of returns. I understand at times that the correlation is low, but there is nothing to say these factors aren't derived from common elements such as growth and inflation.
I think it may even be stronger than 90%--the wiki claims market, size and value "explain" 95% citing this Dimensional white paper by Fama https://www.ifa.com/academic-papers/mul ... a_2006.pdf

But I struggle with the question. No so much of a "how much to tilt" question, but how to achieve a sufficient diversification benefit with the least amount of tilt.

I also realize that the research is moving forward--heard Fama say once or twice that you could throw value out of their five-factor model and make it a four-factor model
Thanks for the posts. By multifactor investing, are you diversifying into factors that are of small relative importance? Does this emphasis on factors of small relative importance decrease one's ability to diversify using factors of greater relative importance?
To clarify a few things on what risk factor models say and don't say.

1) "Explains" is always somewhat of an unfortunate word to use in this regard, as what is being measured is simply the correction of one thing to multiple other things. "Correlation" and "explanation" aren't really the same thing. Which leads to the point that:

2) There is no one true risk model. For example, Fama and French introduced a 5 factor model that added profitability and investment as factors, and they found that doing this made value redundant as a factor. Here is a paper that proposes a two factor model (risk and uncertainty) to which value and size don't add anything. https://www.federalreserve.gov/pubs/fed ... 145pap.pdf. Going in the other direction, there are commercially available risk modeling tools that use dozens of factors. The F/F 3 factor model is a super-useful tool; it's not the only one.

3) Factor models appear to explain a lot more when like is being compared against like. For example, the F/F factors are calculated based cap-weighted portfolios. When comparing a a cap-weighted investment against them (such as VBR or IJS), they leave very little residual return that is unexplained. If I instead use a non-cap weighted ETF like RZV, there is quite a bit of residual that is unexplained. In this case, adding more factors (and more different types of factors) helps quite a bit. If I used the F/F model to analyze an equally-weighted portfolio of 15 stocks I selected, the model would perform even worse. Note that unexplained returns are not an indication of a bad investment; they are an indication of an incomplete risk factor model.

4) Factor models measure concentration of risk. High [absolute] numbers in your output indicate that your risks are concentrated, not that they they diversified. VTI has a market factor of 1.0. This indicates that my risk is totally correlated to the market (duh). If I replace that with 100% IJS, I still have the same market risk exposure, but I also have very high size and value risk as well. This indicates a portfolio where risks are highly concentrated in how one group of stocks perform. That's not necessarily bad if I'm convinced small/value will outperform the market, but the factor model is telling you that your portfolio is quite concentrated. In contrast, the naive equal-weighted portfolio I proposed above (IJS+MTUM+QUAL+USMV) still has a very high market factor (~.9), but other exposures are rather small, which indicates that my risk is much more diversified. If I want diversification of my market risk, I'll need to start adding things like bonds, gold, or residential real estate.

5) For folks like us, qualitative understanding provides a better guide to assembling a diversified portfolio. It was only when "factor" became a buzzword that people start arguing that buying only one type of stock is more diversified than buying many types of stocks. Quantitative factor modeling is actually most useful for searching out unknown or unaccounted for risks. That is, when unexpected correlations arise between assets--that indicates something worth exploring further in terms of asset pricing and portfolio management.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Mon May 20, 2019 8:40 pm

stlutz wrote:
Mon May 20, 2019 8:15 pm
This indicates a portfolio where risks are highly concentrated in how one group of stocks perform.
You’re using words as if they mean the complete opposite of what they actually mean.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

pdavi21
Posts: 1253
Joined: Sat Jan 30, 2016 4:04 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by pdavi21 » Mon May 20, 2019 9:37 pm

vineviz wrote:
Mon May 20, 2019 8:11 pm
pdavi21 wrote:
Mon May 20, 2019 6:08 pm
The flaw is more likely in the funds than the academic paper, but there is a flaw.
I suspect the problem is more likely that you just don't understand the research well enough to implement it.
How do you implement it? I'm all ears.
"We spend a great deal of time studying history, which, let's face it, is mostly the history of stupidity." -Stephen Hawking

stlutz
Posts: 5465
Joined: Fri Jan 02, 2009 1:08 am

Re: Factor ETFs For Diversification Or “Diworsification”

Post by stlutz » Mon May 20, 2019 9:43 pm

vineviz wrote:
Mon May 20, 2019 8:40 pm
stlutz wrote:
Mon May 20, 2019 8:15 pm
This indicates a portfolio where risks are highly concentrated in how one group of stocks perform.
You’re using words as if they mean the complete opposite of what they actually mean.
Suppose I have a portfolio of smallcap value banks. A risk model would show high values on market, size, value, and bank risk. Suppose I then add a bunch of smallcap value tech and energy stocks to the portfolio. My bank score will go way down. Is my portfolio now more or less diversified?

Diversification is a good thing. That does not mean that the best portfolio must therefore be the most diversified one.

The risks of growth stocks are very different from the risks of value stocks. Owning both is more diversified. Focusing on value might well be better for your long-term returns.

Heck, if I went out and bought $10,000 worth of fill dirt, my portfolio would be more diversified; it certainly wouldn't be better.

And, just to be clear, I don't believe that that VTI represents the most diversified domestic stock portfolio. It's usually good enough, but the most optimally-diversified portfolio would likely have different industry weights, for example. And this is where relying on a single risk model can let one down. "Market risk" is really a weighted sum of a whole host of other risks--putting a lot of reliance on it can be problematic.

The F/F model run in 1999 showed all risks outside of market risk to be diversified away in the total market. That's by definition. It's also misleading. A deeper look showed a market that was highly concentrated in one type of stock. Risk modeling back then showed that many stocks outside of big cap tech were acting exactly like big cap tech stocks. Your concentration of risk in big tech was actually higher than even what the technology portfolio weight showed! Using different models to find high concentrations of risk perhaps might have led one to take some action to adjust their portfolio at the time. I would argue that a portfolio of 1/4 VTI,, 1/4 REITs, 1/4 Utilities, and 1/4 small/value was more diversified in 1999 than was 100% VTI. I think using multiple different types of risk models would validate this conclusion.

But to get back the original issue, the way one would validate that contention is to examine whether the risk exposures to big cap tech were lowered without simply being replaced by other similarly high concentrations of other risks. Again, low absolute risk scores indicate diversification of risk; high scores indicate concentration of it.

Park
Posts: 682
Joined: Sat Nov 06, 2010 4:56 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Park » Mon May 20, 2019 9:52 pm

https://seekingalpha.com/instablog/7181 ... ndex-funds

Warren Buffett:

"If you are not a professional investor, if your goal is not to manage money in such a way that you get a significantly better return than world, then I believe in extreme diversification. I believe that 98 or 99 percent - maybe more than 99 percent - of people who invest should extensively diversify and not trade. That leads them to an index fund with very low costs. All they're going to do is own a part of America. They've made a decision that owning a part of America is worthwhile. I don't quarrel with that at all - that is the way they should approach it."

DecumulatorDoc
Posts: 191
Joined: Sun Jan 28, 2018 7:48 am

Re: Factor ETFs For Diversification Or “Diworsification”

Post by DecumulatorDoc » Mon May 20, 2019 10:25 pm

vineviz wrote:
Sun May 19, 2019 8:25 pm
Taylor Larimore wrote:
Sun May 19, 2019 7:07 pm
Currently, Total US Market's 5-year annualized return is 10.76%. Small Value's 5-year annualized return is 5.25%. This is serious under-performance -- especially for retirees.
Bogleheads,

These numbers aren’t quite what actual investors have realized: iShares S&P Small-Cap 600 Value ETF has produced a CAGR if 8.38% over the past five years: not poor by any measure. https://www.portfoliovisualizer.com/fun ... F17%2F2019

And because we are long-term investors, we can’t forget the 20.66% CAGR produced by iShares S&P Small-Cap 600 Value ETF over the previous 5 years.
Taylor continues to use Morningstar category performance for SCV, which includes underperforming active funds. :oops:
The rest of us know to use appropriate benchmarks for comparison. He also stays away from longer term 10 year returns like 14.4% for iShares S&P Small-Cap 600 Value ETF.

I'm fine with it all...happy that most of my equities are in US Total Market and my modest tilt to SCV of 10% is there for the long term. Everyone can argue but nobody can predict the most optimal allocation or strategy moving forward.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Tue May 21, 2019 7:37 am

pdavi21 wrote:
Mon May 20, 2019 6:08 pm
The academic analysis is useless because funds aren't factors. So unless you invest in factors, the analysis on which factor is riskier is useless.
No. My point was that the factor analysis is absolutely NOT useless in evaluating funds if you know how to do it correctly. You just have to read the instruction manual.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
nedsaid
Posts: 12768
Joined: Fri Nov 23, 2012 12:33 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by nedsaid » Tue May 21, 2019 9:46 am

acegolfer wrote:
Mon May 20, 2019 12:27 pm
nedsaid wrote:
Mon May 20, 2019 12:02 pm
acegolfer wrote:
Mon May 20, 2019 11:38 am
nedsaid wrote:
Sun May 19, 2019 11:09 am
I have advocated for Small/Value tilted portfolios for those who believe in Academic Research and the simple 3 fund portfolio for those that don't.
I bet Fama would disagree with this.
It is a way of making peace on the forum. I believe in the Academic Research and have a mild Small/Value tilt with aggressive growth earnings and price momentum funds mixed in my portfolio.
I also believe in the academic research (my profession) and understand your investment decision. But I was debating the logic behind the advice you gave to others. Whether to tilt or not should not be based solely on volatility risk. The fact that CAPM failed means there are non-volatility risks that matter to the investors.
My advice on whether to tilt or not to tilt is not at all on based upon volatility risk. It has to do with people's fundamental beliefs about the markets. If someone thinks that the academic research is bunk, it just seems a waste of time to persuade them to tilt.

Indeed, I have made behavioral and not risk arguments about the factors. My belief is that factors persist because human nature and human behavior don't change over time, we as a whole cannot help ourselves, we make the same behavioral errors over and over. Factors don't persist because they are riskier, that would just take us back to CAPM. As I recall, one thing that got the factors arguments started was that the so-called low volatility stocks had higher returns than the model predicted. Also certain higher volatility stocks had lower returns than predicted. Something else was at work in the markets than beta.

I have argued that Value suffers from fundamental risk, higher debt levels and more volatile earnings. I have further argued that Growth suffers from pricing risk, the risk that the markets have built too much expectations into these stocks. Also discussed that risk can be expressed in different ways, Bill Bernstein calls short term volatility shallow risk and the risk of loss of purchasing power over time deep risk.
A fool and his money are good for business.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Tue May 21, 2019 11:34 am

stlutz wrote:
Mon May 20, 2019 9:43 pm
Suppose I have a portfolio of smallcap value banks. A risk model would show high values on market, size, value, and bank risk. Suppose I then add a bunch of smallcap value tech and energy stocks to the portfolio. My bank score will go way down. Is my portfolio now more or less diversified?
Leaving aside the fact that there is no such thing as a "bank score", I think it should be obvious that a portfolio which includes energy, tech, and bank stocks is more diversified than a portfolio which only includes bank stocks. I'm not sure what could be controversial about that.
stlutz wrote:
Mon May 20, 2019 9:43 pm
Diversification is a good thing. That does not mean that the best portfolio must therefore be the most diversified one.
I agree with this, and I don't think I've seen anyone claim otherwise.
stlutz wrote:
Mon May 20, 2019 9:43 pm
The risks of growth stocks are very different from the risks of value stocks.
I also agree with this. In fact, combining a large cap growth index with a small cap value index is one of the simplest ways to create a stock portfolio that is reasonably well diversified.
stlutz wrote:
Mon May 20, 2019 9:43 pm
And, just to be clear, I don't believe that that VTI represents the most diversified domestic stock portfolio. It's usually good enough, but the most optimally-diversified portfolio would likely have different industry weights, for example.
Yep. You're still doing great.
stlutz wrote:
Mon May 20, 2019 9:43 pm
But to get back the original issue, the way one would validate that contention is to examine whether the risk exposures to big cap tech were lowered without simply being replaced by other similarly high concentrations of other risks. Again, low absolute risk scores indicate diversification of risk; high scores indicate concentration of it.
Here's where your argument starts to get into trouble.

For one thing, the argument is confused by the use non-specific or meaningless terms ("concentrations of other risk", " absolute risk scores", "diversification of risk").

But if by "low absolute risk scores" you mean portfolio volatility, your last sentence isn't correctly stated. The level of risk in a portfolio and the diversification of risk in a portfolio are two different things. It is entirely possible to have an undiversified portfolio with a low overall level of risk, a well-diversified portfolio with a high overall level of risk, and anything in between.

Portfolio diversification works by combining multiple independent sources of risk in a balanced manner. The way to diversify a portfolio of tech stocks, for instance, is to replace some portion of those stocks with an asset hat is less than perfectly correlated with tech stocks. The effectiveness of this diversification would depend on the variance of the second asset and the correlation of that asset with tech stocks.

When the topic is diversification, talking about the concentration of assets is much less useful than talking about the concentration of risk. Even in that context, the use of words like "concentration" is problematic because concentration itself depends on other dimensions both how correlated the risks are and how powerful they are (i.e. how high their variance is) they are.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
hdas
Posts: 1270
Joined: Thu Jun 11, 2015 8:24 am

Re: Factor ETFs For Diversification Or “Diworsification”

Post by hdas » Thu May 23, 2019 9:58 am

vineviz wrote:
Mon May 20, 2019 12:45 pm
hdas wrote:
Mon May 20, 2019 12:06 pm

Totally agree with you that growth isn't the opposite of value. But isn't earnings growth a factor in the sense that is a independent variable that partially explains stock performance?
When Morningstar developed the equity style box (which weren't based on factors as much as they were on classic ad hoc fund categories), it became a powerfully simple way to explain mutual fund performance. I suppose it was not completely an accident that the style box captures the first two factors that Fama-French added to the CAPM, but it definitely built an implicit expectation among investors that value and growth were two ends of the same dimension.

Anyway, when I say that growth isn't a factor I really mean that it's not a statistically valid factor. You could, of course, sort the universe of stocks by some growth metric (EPS growth rate or whatever) and construct long-short portfolios in the same manner that the HML and SMB portfolios were constructed. Billions of CPU cycles have been wasted by researchers doing just that.

When you do it, though, what you find is that the growth metrics don't improve the pricing model. In other words, they don't add enough explanatory power to justify their inclusion. That's the reason that none of the factor researches (Fama-French, Carhart, Asness, MSCI, etc.) include it in their models or smart beta products..

Whatever performance typical growth funds exhibit is generally explainable by accepted factors (beta, size, value, momentum, and quality): adding explicit growth terms to the regression doesn't help explain the cross-section of returns.
I take your point about the significance of earnings growth,...and that's all we have for data at the moment. However, one feature I like about this pseudo-factor, is the "adaptability" as it is the less depending on a specific industry. It seems to be a combination of momentum + quality.

From this interesting site
THE GROWTH FACTOR

Calculated by a combination of sales and earnings-per-share growth over the last three years, the growth factor shows the greatest sector rotation of all factors and illustrates the ever-changing nature of the economy. The Technology sector currently features larger in the long portfolio, reflecting the strong growth of tech companies in recent years. The short portfolio is dominated by the Energy sector thanks to lower oil and gas prices.
Image

Perhaps I'm just arguing my book. :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
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by vineviz » Thu May 23, 2019 10:22 am

hdas wrote:
Thu May 23, 2019 9:58 am
I take your point about the significance of earnings growth,...and that's all we have for data at the moment. However, one feature I like about this pseudo-factor, is the "adaptability" as it is the less depending on a specific industry. It seems to be a combination of momentum + quality.

From this interesting site
I'm glad to see someone else reads FactorResearch.com. Nicolas Rabener always seem to have something intelligent and relevant to say.

In their smart beta ETF rundown, they showed that their "Growth Factor" had underperformed the market since 2000 (though not so much lately) and had the lowest returns of the major factors they compared.

Image

In another piece, it looked to me the excess returns of Vanguard Small-Cap Growth ETF (VBK) were the most correlated to their growth factor long-short portfolio. A traditional factor analysis of VBK would indicated that you are right: their growth "factor" probably represents some combination of momentum, quality, and size.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

acegolfer
Posts: 1486
Joined: Tue Aug 25, 2009 9:40 am

Re: Factor ETFs For Diversification Or “Diworsification”

Post by acegolfer » Sat May 25, 2019 10:04 am

vineviz wrote:
Mon May 20, 2019 4:49 am
I see now that French is referring to the low market-to-book portfolios as "growth", so your confusion is understandable. When I studied finance with Fama, he never let us call those portfolios "growth" portfolios, because nothing in their construction was actually measuring growth. Maybe they've adopted the colloquial usage now, which is unfortunate.
True. In 90's, Fama didn't call low ME/BE stocks as "growth" stocks. He didn't even call high ME/BE stocks as "value" stocks, either. His 6 portfolios which were used to construct SMB and HML were called "S/L, S/M, S/H, B/L, B/M, B/H". Later, they changed these 6 terms to "small growth, small neutral, small value, big growth, big neutral, big value".

What's so unfortunate about the new terminology that average investors are more comfortable with?

Park
Posts: 682
Joined: Sat Nov 06, 2010 4:56 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Park » Sun May 26, 2019 5:53 pm

There are those who question whether factor investing will continue to be as profitable in the future, as it has in backtesting. This is based on a conclusion, which I think is correct, that a large part (good majority?) of factor returns are behavioral in origin and not risk based.

My guess is that factor returns will decrease, but not disappear. Why?

I know of no one who thinks that there won't be bear and bull markets in the future. Although financial economics has progressed greatly, there were still two significant bear markets in the last 20 years. IMO, a significant cause of those bear markets was human behavior. Fear and greed have not been repealed.

What Is a stock market? It's a market of stocks. The same human behavior that drives stocks markets also drives stocks. If one thinks that the human behavior that results in factor returns will become insignificant, will the same change in human behavior also apply to stock markets? Will bear markets become less of an issue in the future? If you don't think that bear markets will become less of an issue in the future, why should human behavior change at the level of individual stocks, but not at the level of markets?

Typographical correction: the last use of the word level was written as less, and I corrected that.
Last edited by Park on Tue May 28, 2019 7:34 am, edited 1 time in total.

User avatar
packer16
Posts: 1278
Joined: Sat Jan 04, 2014 2:28 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by packer16 » Sun May 26, 2019 7:17 pm

IMO there is a difference between value investing and factor investing. Value investing is estimating an intrinsic value & buying securities that trade a significant discounts to intrinsic value. Factor investing is a relative pricing exercise. The pricing exercise changes over time based upon the popular characteristic of the day/week/month/year. Factors are good explaining historical pricing of securities based upon how the market weights various characteristics over a given period of time. However these weights change over time & thus when you invest this way you are depending on the characteristics being weighted in the same way over your investment period. When you do this with more than one factor you are IMO making the situation of the historical weights = future weights a much less likely occurrence. This is why IMO multi-factor strategies for the most part have a hard time doing better than plain vanilla indexing.

The difference between pricing and valuation is an important distinction in looking at value investing versus value factor investing. Value investing can also include growth stocks if the growth is not reflected in the stock price.

Packer
Buy cheap and something good might happen

User avatar
nedsaid
Posts: 12768
Joined: Fri Nov 23, 2012 12:33 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by nedsaid » Mon May 27, 2019 2:22 pm

packer16 wrote:
Sun May 26, 2019 7:17 pm
IMO there is a difference between value investing and factor investing. Value investing is estimating an intrinsic value & buying securities that trade a significant discounts to intrinsic value. Factor investing is a relative pricing exercise. The pricing exercise changes over time based upon the popular characteristic of the day/week/month/year. Factors are good explaining historical pricing of securities based upon how the market weights various characteristics over a given period of time. However these weights change over time & thus when you invest this way you are depending on the characteristics being weighted in the same way over your investment period. When you do this with more than one factor you are IMO making the situation of the historical weights = future weights a much less likely occurrence. This is why IMO multi-factor strategies for the most part have a hard time doing better than plain vanilla indexing.

The difference between pricing and valuation is an important distinction in looking at value investing versus value factor investing. Value investing can also include growth stocks if the growth is not reflected in the stock price.

Packer
A good post. I explain that there are at least two definitions of Value.

First definition is the Graham-Dodd definition which has to do with stocks priced significantly below intrinsic value. Intrinsic value involves some subjective judgments in my opinion, for one thing you are making a judgment on the management team, deciding whether or not they can turn things around. As the famous professor said, "Every good valuation has a story behind it." Just like Peter Lynch would say to develop a narrative for each stock you buy. Also, accounting has difficulty valuing intangible assets and the going concern value of the company itself. This method is a bottoms up analysis company by company. Of course, one pours over the financials of each company doing a lot of analysis but also there is effort to see what is behind the numbers.

The Academic definition of Value is picking the cheapest 30% of the stock market using such metrics as Price to Earnings, Price to Cash Flow, Price to Sales, and Price to Book. It is what I call the CostCo approach, pretty much you are picking stocks in bulk, running the financials through computer screens to find all stocks that meet your preset criteria. It is more of a top-down approach.

There is a third definition of Value and that would be Growth At a Reasonable Price (GARP) and it seems Buffett is more in this camp than in the Graham-Dodd camp. Packer16 alluded to that above.

Investment firms hire analysts to analyze the market sector by sector and within a sector, stock by stock. Hard to see where you will find much in the way of market inefficiencies with the largest, most followed, and most traded of the Large-Cap stocks. Obviously some companies attract more analyst coverage than others. But also consider how software and super fast computer hardware makes crunching the numbers so much faster. Instead of analysts cranking away 8-5 Monday through Friday, the computers can crank away 24-7. This has created problems even for the great Warren Buffett.

The research seems to show that the Value premium, if it even still exists, is pretty slim with Large-Cap stocks. You see most of the Value effect now in Mid-Caps and Small-Caps. Of course, those are areas of the market with less analyst coverage and less trading volume.
A fool and his money are good for business.

Topic Author
Random Walker
Posts: 4188
Joined: Fri Feb 23, 2007 8:21 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Random Walker » Mon May 27, 2019 6:30 pm

I’ve read that when one adds the profitability factor to a value screen, that effectively accounts for Buffett’s success. Is that basically GARP?

Dave

User avatar
nedsaid
Posts: 12768
Joined: Fri Nov 23, 2012 12:33 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by nedsaid » Mon May 27, 2019 6:45 pm

Random Walker wrote:
Mon May 27, 2019 6:30 pm
I’ve read that when one adds the profitability factor to a value screen, that effectively accounts for Buffett’s success. Is that basically GARP?

Dave
I think you are getting close. Growth at a Reasonable Price isn't really a Value strategy as such but it is a variant.

Growth at a reasonable price goes something like this. You wanted to own a great Blue Chip stock trading at let's say 20 times earnings. It has earnings growth of 8% to 9% a year but commands such a premium because of the amazing consistency of its growth and the very strong balance sheet. You didn't buy it because you don't want to pay the high price as it is priced for a Goldilocks scenario. As Morningstar would say, this company has a wide economic moat, so the barriers of entry for competitors is very high. The industry the company is in has very good prospects. Great company in a great industry with high barriers to entry for additional competitors and that is why the company trades at such a premium. Bear market hits, the stock is available at a 16 P/E ratio, the company NEVER trades at this multiple. Meanwhile, the business just keeps chugging along with a minor earnings disappointment but the earnings miss was due to short term factors that won't repeat. You swoop in and buy 100 shares.
A fool and his money are good for business.

User avatar
packer16
Posts: 1278
Joined: Sat Jan 04, 2014 2:28 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by packer16 » Mon May 27, 2019 7:52 pm

Random Walker wrote:
Mon May 27, 2019 6:30 pm
I’ve read that when one adds the profitability factor to a value screen, that effectively accounts for Buffett’s success. Is that basically GARP?

Dave
I think this may explain some of what Buffett did. What I find interesting is academics claim stuff like this but any portfolio put together with these factors doesn't look like Buffett's portfolio at all. It is like the financial porn you see in popular money magazine's that "These are the stocks that Warren Buffett would buy." IMO it is financial porn for academics.

Packer
Buy cheap and something good might happen

stan1
Posts: 7710
Joined: Mon Oct 08, 2007 4:35 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by stan1 » Mon May 27, 2019 7:58 pm

Random Walker wrote:
Mon May 27, 2019 6:30 pm
I’ve read that when one adds the profitability factor to a value screen, that effectively accounts for Buffett’s success. Is that basically GARP?

Dave
Long before I knew anything different I thought PRIMECAP was the master of growth at a reasonable price (and I think that's what sold me on it around 1994).

stan1
Posts: 7710
Joined: Mon Oct 08, 2007 4:35 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by stan1 » Mon May 27, 2019 8:32 pm

Random Walker wrote:
Mon May 27, 2019 6:30 pm
I’ve read that when one adds the profitability factor to a value screen, that effectively accounts for Buffett’s success. Is that basically GARP?

Dave
I just found this analysis of Berkshire Hathaway from a factor perspective.

https://www.factorresearch.com/research ... f-all-time

Summary:
Long Value, Size, Quality, and Low Volatility factors and short Growth and Dividend Yield (neutral Momentum)

Park
Posts: 682
Joined: Sat Nov 06, 2010 4:56 pm

Re: Factor ETFs For Diversification Or “Diworsification”

Post by Park » Tue May 28, 2019 7:55 am

Park wrote:
Sun May 26, 2019 5:53 pm
There are those who question whether factor investing will continue to be as profitable in the future, as it has in backtesting. This is based on a conclusion, which I think is correct, that a large part (good majority?) of factor returns are behavioral in origin and not risk based.

My guess is that factor returns will decrease, but not disappear. Why?

I know of no one who thinks that there won't be bear and bull markets in the future. Although financial economics has progressed greatly, there were still two significant bear markets in the last 20 years. IMO, a significant cause of those bear markets was human behavior. Fear and greed have not been repealed.

What Is a stock market? It's a market of stocks. The same human behavior that drives stocks markets also drives stocks. If one thinks that the human behavior that results in factor returns will become insignificant, will the same change in human behavior also apply to stock markets? Will bear markets become less of an issue in the future? If you don't think that bear markets will become less of an issue in the future, why should human behavior change at the level of individual stocks, but not at the level of markets?

Typographical correction: the last use of the word level was written as less, and I corrected that.
Once again, will factor returns be arbitraged away, as a large part of the historical returns are behavioral in origin?

Let's look at small value stocks. One could envision a scenario where it becomes a crowded trade. After all, they are only 3% of the market. And it's easy nowadays to invest in them, and it wasn't in the past.

But the outperformance of small value stocks is related to the underperformance of large and growth stocks. The top 10% of stocks (approximately top 500 stocks) have underperformed the rest of the market by about 1.5% historically. I doubt that will change. Liquidity issues cause larger investors to gravitate towards those stocks. As for growth stock underperformance, it may lessen, but I doubt it will go away. There has always been bubbles, and there always will be. Growth stocks will tend to be where those bubbles occur. The greater the bubble, the more that large cap stocks will be part of it. The use of shorting to profit from growth stock underperformance is either difficult ( small growth stocks) or hazardous (large growth stocks). It is hazardous, because it isn't easy to predict the extent and duration of a bubble. How many people made money by shorting the internet bubble of 1999?

Post Reply