Practice vs theory in portfolio strategies

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Practice vs theory in portfolio strategies

Post by Astones »

I'd like to have a conversation about the applicability of some of the modern theories concerning portfolio strategies.
I do intend to work out a solution that is in line with what the theory would prescribe, but at the same time I don't want to be dogmatic and it's important to acknowledge where the flaws are when the theory is used in the real world.

For those who don't want to read the rest, this thread is about how -if at all- you deviate from the theoretically "ideal" portfolio (buying the whole financial market -including both stocks and bonds- with market-cap weighing) and what is your reasoning behind it.

To explain what I mean in more detail, the best thing to do is to just dive into the subject:

The theory
Theoretically, assuming that (a) the market is efficient (b) every security is equally available to all market participants (c) we can do leverage for free (d) the companies in the market remain the same over time, then the best strategy would be to basically make a Bill Sharpe portfolio, buy every possible security weighing them according to their market cap, and then tune the expected return according to our risk tolerance using leverage.

How it works out in practice
From the top of my head, there are few issues that come to my mind when we actually implement the strategy in practice:

1) the most important: leverage isn't free at all. So, if someone isn't satisfied with the expected return of the alleged efficient portfolio, you can't really move smoothly along the tangent line by just borrowing money, so a case can be made that increasing the risk by rather tuning the weights differently would end up being more cost efficient. This is what people usually do: you don't really weigh according to market cap in the case of different asset classes, but usually you allocate to stocks a weight that depends on your age and your risk tolerance. Then some people do a step further and they also change the allocation among stocks through various forms of factor investing.

2) Some securities are easier to buy than others. This kind of friction, especially when we try to diversify geographically, might lead to market capitalizations that are different with respect to the ones we'd have if all securities belonged to the same market. This could also be a valid argument in favor of the Bogle portfolio, of just buying the total US market as opposed to the world market.

3) Then there is the intrinsic problem that you can't actually buy every possible asset. You'd buy a sample through ETFs, but then the larger the turnover ratio due to companies being added/removed from the index, the lower your return would be with respect of having, yes, fewer companies, but with the certainty that those companies will remain the same over time.

Don't hesitate to add more practical issues that you believe would be present if we try to apply the theory under the efficient market hypothesis.
User avatar
nedsaid
Posts: 14561
Joined: Fri Nov 23, 2012 12:33 pm

Re: Practice vs theory in portfolio strategies

Post by nedsaid »

A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well. For example, Larry Swedroe says that this effect has taken away about 1/3 of the factor premiums.
A fool and his money are good for business.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well. For example, Larry Swedroe says that this effect has taken away about 1/3 of the factor premiums.
I love Larry Swedroe. That said, if small cap companies tend to have a larger growth rate, and a larger standard deviation on that growth, shouldn't the premium be somehow eternal ? Larger returns for larger risks is in line with the efficient hypothesis.
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

I like the idea of moving towards risk parity: diversify away from the factor that dominates most all our portfolios, the market factor. I believe in diversifying across geography, factors, styles, even alternatives. The best and cheapest source of diversification is high quality bonds. Everything else down the diversification/portfolio efficiency path involves decreased marginal benefit and increased marginal cost. Each investor needs to decide where to draw the line. For me that’s where theory meets reality. Personally, I’ve sort of gone all in on the theory side of things. Over the last decade, I would have done substantially better following Bogle’s Cost Matters Hypothesis and the remainder of his advice. We’ll see what the next decade brings.

Dave
KlangFool
Posts: 20488
Joined: Sat Oct 11, 2008 12:35 pm

Re: Practice vs theory in portfolio strategies

Post by KlangFool »

OP,

'I know that I know nothing. Hence, I do not put 100% of my portfolio into any single strategies.

A) 40% of my portfolio is in the 3 funds. -> Passive Index

B) 40% of my portfolio is actively managed -> Wellington Fund. Some times, the market can go crazy and no longer efficient

C) 20% of my portfolio is Larry portfolio -> 10% SCV and 10% Intermediate-Term Treasury.

I am not looking to make the most money. I am hedging to protect myself from any single strategy from going seriously wrong.

KlangFool
40% VWENX | 12.5% VFWAX/VTIAX | 11.5% VTSAX | 16% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 40% Wellington 40% 3-funds 20% Mini-Larry
User avatar
nisiprius
Advisory Board
Posts: 43360
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Practice vs theory in portfolio strategies

Post by nisiprius »

In real life, the efficient frontier squirms, writhes, and wriggles with surprisingly small changes in the endpoints of the data used to calculate the returns, standard deviations, and correlation coefficient. The grazing point of the tangent line slides back and forth, often kissing the hyperbola at points represent 100% or more devoted to one asset. Basically most of the time if you use MPT based on actual past data, it is just going to tell you to put most of your money into whatever did best recently... often leveraging up whatever did best and shorting whatever did poorly. MPT is not very useful unless you are a deity who knows the "true" underlying statistics, of which the real assets are merely a sample.

The Black-Litterman method allows you to be that deity, and to incorporate your "views" on how you think the assets out to behave in future instead of how they did behave in the past. It tends to tell you to put most of your money into whatever you think is going to do best in the future.
Last edited by nisiprius on Sat May 01, 2021 12:34 pm, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

Astones wrote: Sat May 01, 2021 12:11 pm
nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well. For example, Larry Swedroe says that this effect has taken away about 1/3 of the factor premiums.
I love Larry Swedroe. That said, if small cap companies tend to have a larger growth rate, and a larger standard deviation on that growth, shouldn't the premium be somehow eternal ? Larger returns for larger risks is in line with the efficient hypothesis.
Yes, if it’s a risk story, it should persist. That being said, the size of the risk premium can certainly shrink when the factor becomes well known. Moreover, markets are more liquid, easier, and cheaper to invest in now than in the past. One can much more readily access the size factor. So it’s rational for the size of the expected premium to shrink, but not disappear. Rational in an efficient market to see all investable assets have similar Sharpe ratios. If this were not the case, investors would buy the assets with higher Sharpes, bid their prices up, cause expected returns to decrease, until Sharpe’s equilibrate. Markets price risk.

Dave
Lock
Posts: 43
Joined: Sat Oct 05, 2019 8:52 pm

Re: Practice vs theory in portfolio strategies

Post by Lock »

If we’re going to be looking at this theoretically, crypto is a valuable asset class due to diversification and range of outcomes that needs to be added here.

Futures are the link in a lot of ways between the theory and practical. Without futures, your proposal becomes incredibly difficult to implement (depending on your risk goal of course).
User avatar
nedsaid
Posts: 14561
Joined: Fri Nov 23, 2012 12:33 pm

Re: Practice vs theory in portfolio strategies

Post by nedsaid »

Astones wrote: Sat May 01, 2021 12:11 pm
nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well. For example, Larry Swedroe says that this effect has taken away about 1/3 of the factor premiums.
I love Larry Swedroe. That said, if small cap companies tend to have a larger growth rate, and a larger standard deviation on that growth, shouldn't the premium be somehow eternal ? Larger returns for larger risks is in line with the efficient hypothesis.
From what I have read, the Size premium has disappeared. However when you screen a bit for Quality, the Size premium returns with a vengeance, that shows there is some junk in the Small-Cap sector. A Morningstar report specifically mentioned the S&P 600 Small-Cap Index as a good example of an index that screens for Quality, companies have to have actual earnings to be included in the S&P Indexes. Screening a bit for Quality helps Value as well.

Smaller stocks and Value stocks have what I call fundamental risk, they are more vulnerable to economic downturns and just plain old bad luck. Value companies tend to have more volatile earnings, more leverage on their balance sheets. Smaller companies don't have the resources that the larger companies have to weather recessions and frankly bad management decisions.

The Large Growth Stocks have their own risk, what I call pricing risk. Historically, investors can get too optimistic about Growth when times are good and too pessimistic about Value. There is a point where investor enthusiasm can turn great companies into poor investments. I think a lot of the story regarding factors is behavioral and not so much a risk story. Certainly, you are taking on certain risks when investing in Small-Cap stocks and when investing in Value but you take a behavioral risk with Large Growth. Wall Street is an expectations game, low expectations are easier to beat than high expectations. Sometimes the expectations for Growth just gets to be too high.

An implementation problem with factor investing is that the various US vs. International, Value vs Growth, Small vs. Large trends can last a long time. Value oriented investors like myself have waited a time time for a reversion to the mean in favor of Value, I have been waiting since 2009. Many investors trying factor strategies just don't have the patience required. There are signs that we might be in a Value trend, but I thought that a few years ago. We will have to wait and see if this Value trend will hold or just be temporary.
A fool and his money are good for business.
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

nisiprius wrote: Sat May 01, 2021 12:33 pm MPT is not very useful unless you are a deity who knows the "true" underlying statistics, of which the real assets are merely a sample.
I disagree. We can only guesstimate expected returns, use historical correlations and knowledge about how they can change, and use historical volatilities. But without looking for the ultimate portfolio on the future efficient frontier, we can use the general concepts of MPT to create portfolios more likely to be close to the elusive northwest corner. This is one of my favorite pieces of investment writing. This essay effectively summarizes the entirety of Gibson’s book Asset Allocation: Balancing Financial Risk. The specific asset classes are not important. The concepts are.

https://ivinvestor.com/wp-content/uploa ... sting1.pdf

Dave
User avatar
nedsaid
Posts: 14561
Joined: Fri Nov 23, 2012 12:33 pm

Re: Practice vs theory in portfolio strategies

Post by nedsaid »

nisiprius wrote: Sat May 01, 2021 12:33 pm In real life, the efficient frontier squirms, writhes, and wriggles with surprisingly small changes in the endpoints of the data used to calculate the returns, standard deviations, and correlation coefficient. The grazing point of the tangent line slides back and forth, often kissing the hyperbola at points represent 100% or more devoted to one asset. Basically most of the time if you use MPT based on actual past data, it is just going to tell you to put most of your money into whatever did best recently... often leveraging up whatever did best and shorting whatever did poorly. MPT is not very useful unless you are a deity who knows the "true" underlying statistics, of which the real assets are merely a sample.

The Black-Litterman method allows you to be that deity, and to incorporate your "views" on how you think the assets out to behave in future instead of how they did behave in the past. It tends to tell you to put most of your money into whatever you think is going to do best in the future.
Here at Bogleheads, most all of us assume that the Equity Risk Premium will persist in the future. Indeed our portfolios are mostly built on this assumption. This isn't playing God but simply using market history as a guide as to what is likely to happen in the future.

But you are right in that making assumptions about the future has a lot of uncertainty to it. But educated guesses are a whole lot better than nothing.

It is possible that computer algorithms, Artificial Intelligence, legions of smart people scouring the markets for anomalies have collapsed the Factor Premiums to zero. Based upon my knowledge of human nature, human behavior, and market history, I don't think so. Despite reams of data showing that individual stock picking by small investors is not a good idea, the little guy is rushing to trading platforms like Robinhood. I thought that retaining my individual stock portfolio made me a dinosaur, it turns out that I was both behind the times and ahead of the times. In other words, people are reverting to the very behavior that made the factor premiums work in the first place.

So maybe this small investor speculation in individual stocks is ultimately creating a reverse Robin Hood effect, stealing from the poor and giving to the rich. Perhaps a new generation of suckers with little investment experience or perspective on market history will fuel the factors again. What is old is new again, sort of back to the future.
A fool and his money are good for business.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

Lock wrote: Sat May 01, 2021 12:34 pm If we’re going to be looking at this theoretically, crypto is a valuable asset class due to diversification and range of outcomes that needs to be added here.

Futures are the link in a lot of ways between the theory and practical. Without futures, your proposal becomes incredibly difficult to implement (depending on your risk goal of course).
I have always intended portfolio theories as specifically devoted to assets with an intrinsic value -which would exclude commodities and bitcoins- but your point is well taken and yes, why not, a case can be made that in a more general interpretation bitcoins and commodities should be included in the abstract theoretical portfolio, weighed according to their market cap.
KlangFool
Posts: 20488
Joined: Sat Oct 11, 2008 12:35 pm

Re: Practice vs theory in portfolio strategies

Post by KlangFool »

OP,

I do not see anything to dispute the Larry portfolio principle. Aka, pairing ultra-risky SCV and ultra-safe Treasury in order to take advantage of the inherent human behavior. Overly optimistic and overly pessimistic in market cycle. It works out very well in March 2020.

So, there are at least a different way to look at SCV. Instead of using SCV as standalone, use it as part of the Larry portfolio.

KlangFool
40% VWENX | 12.5% VFWAX/VTIAX | 11.5% VTSAX | 16% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 40% Wellington 40% 3-funds 20% Mini-Larry
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

KlangFool wrote: Sat May 01, 2021 1:32 pm So, there are at least a different way to look at SCV. Instead of using SCV as standalone, use it as part of the Larry portfolio.
KlangFool
I know some young advisors at Larry’s firm are basically 100% SV, but in general I think Larry recommends the heavy tilt to SV in conjunction with decreasing overall equity allocation. For example 40% SV / 60% bonds aa an alternative to 60% TSM / 40% bonds. From a factor point of view, it’s a move in the direction of risk parity.

Dave
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sat May 01, 2021 11:31 am The theory
Theoretically, assuming that (a) the market is efficient (b) every security is equally available to all market participants (c) we can do leverage for free (d) the companies in the market remain the same over time, then the best strategy would be to basically make a Bill Sharpe portfolio, buy every possible security weighing them according to their market cap, and then tune the expected return according to our risk tolerance using leverage.
This is not the full picture of the theory. Bill Sharpe portfolio = optimal risky portfolio for everyone is the practical implication of CAPM, which requires more assumptions than the 3 you stated. One critical assumption under CAPM (that is not used in later asset pricing theory) is all investors are mean-variance investors.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 8:52 am This is not the full picture of the theory. Bill Sharpe portfolio = optimal risky portfolio for everyone is the practical implication of CAPM, which requires more assumptions than the 3 you stated. One critical assumption under CAPM (that is not used in later asset pricing theory) is all investors are mean-variance investors.
Your point is well taken but I believe that practically real systems are close enough to respecting this condition for it not to be an issue. It seems to me that people make imperfect decisions but the vast majority of market participants are trying to maximize their returns given their risk tolerance. Not all of them, but those who don't seem a tiny minority to me. Is it a fair assessment ?
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 9:13 am
acegolfer wrote: Sun May 02, 2021 8:52 am This is not the full picture of the theory. Bill Sharpe portfolio = optimal risky portfolio for everyone is the practical implication of CAPM, which requires more assumptions than the 3 you stated. One critical assumption under CAPM (that is not used in later asset pricing theory) is all investors are mean-variance investors.
Your point is well taken but I believe that practically real systems are close enough to respecting this condition for it not to be an issue. It seems to me that people make imperfect decisions but the vast majority of market participants are trying to maximize their returns given their risk tolerance. Not all of them, but those who don't seem a tiny minority to me. Is it a fair assessment ?
If what you are saying true, then CAPM (a single factor model) should explain cross section of stock returns. Unfortunately, it's a fact that CAPM can't. Instead, a multi-factor model (that doesn't rely on mean-variance investors) better explains the reality. Investors are risk averse, but risk is not equal to variance.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 9:16 am If what you are saying true, then CAPM (a single factor model) should explain cross section of stock returns. Unfortunately, it's a fact that CAPM can't. Instead, a multi-factor model (that doesn't rely on mean-variance investors) better explains the reality. Investors are risk averse, but risk is not equal to variance.
Can you elaborate a little bit on this concept of cross section of stock returns, as well as in the implications of risk not being equal to variance ?

I have always -maybe naively- thought about risk as the standard deviation of returns (sqrt of variance), assuming the return distribution to be symmetric in first approximation.

Also, what changes should someone make in the portfolio to capture these aspects ?
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 9:29 am Can you elaborate a little bit on this concept of cross section of stock returns, as well as in the implications of risk not being equal to variance ?

I have always -maybe naively- thought about risk as the standard deviation of returns (sqrt of variance), assuming the return distribution to be symmetric in first approximation.

Also, what changes should someone make in the portfolio to capture these aspects ?
Good questions.
1. CAPM states the market factor alone can explain the individual stock returns. A stock with higher beta (=loading of market factor) will have higher expected return. This proved to be wrong in 70s/80s. In 90s, multi-factor models (eg. Fama-French 3-factor model) were introduced to fill the gap in reality.

2. Variance is just one measurement of risk and arguably the most popular one. It has worked better than any other measures because the distribution of stock returns mimics a normal distribution. So all the theory in 50s/60s (aka modern portfolio theory) were developed under this assumption. In addition to variance, ppl are also concerned with other risk such as recession risk. If company A does much worse in a bad economy than company B, stock A may be considered more risky, even if A and B have the same variance.

3. How should an individual create portfolio? Fama suggests everyone starts from the market portfolio because it's the center of universe and efficient. Then based on one's preference, he should decide whether/which way to tilt.
Last edited by acegolfer on Sun May 02, 2021 9:44 am, edited 2 times in total.
User avatar
willthrill81
Posts: 24507
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Practice vs theory in portfolio strategies

Post by willthrill81 »

nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well.
To the extent that is accurate, the market are not as efficient as we've been told that it is. If they were, all such strategies would already have been known and accounted for by the market a priori.
“Good and ill have not changed since yesteryear; nor are they one thing among Elves and Dwarves and another among Men.” J.R.R. Tolkien, The Lord of the Rings
User avatar
nedsaid
Posts: 14561
Joined: Fri Nov 23, 2012 12:33 pm

Re: Practice vs theory in portfolio strategies

Post by nedsaid »

willthrill81 wrote: Sun May 02, 2021 9:41 am
nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well.
To the extent that is accurate, the market are not as efficient as we've been told that it is. If they were, all such strategies would already have been known and accounted for by the market a priori.
You aren't supposed to say that! :wink:
A fool and his money are good for business.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 9:39 am Good questions.
1. CAPM states the market factor alone can explain the individual stock returns. A stock with higher beta (=loading of market factor) will have higher expected return. This proved to be wrong in 70s/80s. In 90s, multi-factor models (eg. Fama-French 3-factor model) were introduced to fill the gap in reality.

2. Variance is just one measurement of risk and arguably the most popular one. It has worked better than any other measures because the distribution of stock returns mimics a normal distribution. So all the theory in 50s/60s (aka modern portfolio theory) were developed under this assumption. In addition to variance, ppl are also concerned with other risk such as recession risk. If company A does much worse in a bad economy than company B, stock A may be considered more risky, even if A and B have the same variance.

3. How should an individual create portfolio? Fama suggests everyone starts from the market portfolio because it's the center of universe and efficient. Then based on one's preference, he should decide whether/which way to tilt.
Thanks for taking the time to write the explanation.

1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 9:54 am Thanks for taking the time to write the explanation.

1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
Again that depends on individual preference. Some prefer to leverage to increase risk and return. Others prefer to tilt to exposed to non-market factor risks and increase return. Some (including myself) are perfectly content with sticking to market portfolio.
User avatar
willthrill81
Posts: 24507
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Practice vs theory in portfolio strategies

Post by willthrill81 »

nedsaid wrote: Sun May 02, 2021 9:49 am
willthrill81 wrote: Sun May 02, 2021 9:41 am
nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well.
To the extent that is accurate, the market are not as efficient as we've been told that it is. If they were, all such strategies would already have been known and accounted for by the market a priori.
You aren't supposed to say that! :wink:
Just stirrin' the pot! :twisted:
“Good and ill have not changed since yesteryear; nor are they one thing among Elves and Dwarves and another among Men.” J.R.R. Tolkien, The Lord of the Rings
User avatar
nedsaid
Posts: 14561
Joined: Fri Nov 23, 2012 12:33 pm

Re: Practice vs theory in portfolio strategies

Post by nedsaid »

willthrill81 wrote: Sun May 02, 2021 10:06 am
nedsaid wrote: Sun May 02, 2021 9:49 am
willthrill81 wrote: Sun May 02, 2021 9:41 am
nedsaid wrote: Sat May 01, 2021 11:38 am A big issue is that when strategies became well known, money rushes in, and those strategies either cease to work or work not as well.
To the extent that is accurate, the market are not as efficient as we've been told that it is. If they were, all such strategies would already have been known and accounted for by the market a priori.
You aren't supposed to say that! :wink:
Just stirrin' the pot! :twisted:
Yes, I have been known to stir the pot too.
A fool and his money are good for business.
User avatar
willthrill81
Posts: 24507
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Practice vs theory in portfolio strategies

Post by willthrill81 »

Random Walker wrote: Sat May 01, 2021 12:50 pm
nisiprius wrote: Sat May 01, 2021 12:33 pm MPT is not very useful unless you are a deity who knows the "true" underlying statistics, of which the real assets are merely a sample.
I disagree. We can only guesstimate expected returns, use historical correlations and knowledge about how they can change, and use historical volatilities. But without looking for the ultimate portfolio on the future efficient frontier, we can use the general concepts of MPT to create portfolios more likely to be close to the elusive northwest corner. This is one of my favorite pieces of investment writing. This essay effectively summarizes the entirety of Gibson’s book Asset Allocation: Balancing Financial Risk. The specific asset classes are not important. The concepts are.

https://ivinvestor.com/wp-content/uploa ... sting1.pdf

Dave
I see both your point and nisi's. The efficient frontier has substantially changed its boundary over time and in seemingly not very predictable ways. But even so, estimating the efficient frontier and allocating accordingly may lead to a more efficient portfolio than otherwise.

Frankly, a good bit of this probably comes down to whether one is a satisficer or an optimizer with regard to one's portfolio.
“Good and ill have not changed since yesteryear; nor are they one thing among Elves and Dwarves and another among Men.” J.R.R. Tolkien, The Lord of the Rings
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 10:05 am Again that depends on individual preference. Some prefer to leverage to increase risk and return. Others prefer to tilt to exposed to non-market factor risks and increase return. Some (including myself) are perfectly content with sticking to market portfolio.
I thought we were talking about the theory here. If it's true that your larger expected returns are due to larger risks, then the implication is that a leveraged market portfolio would have a better risk/return profile because you are moving along the tangent line, whereas by over-weighing the small cap or value stocks you are not.
User avatar
cos
Posts: 401
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

Re: Practice vs theory in portfolio strategies

Post by cos »

Astones wrote: Sun May 02, 2021 9:54 am Thanks for taking the time to write the explanation.

1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
Not quite. In terms of straight risk-adjusted returns, maybe, but exposure to multiple different risk factors also offers diversification since they are fairly uncorrelated with each other.
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 10:12 am
acegolfer wrote: Sun May 02, 2021 10:05 am Again that depends on individual preference. Some prefer to leverage to increase risk and return. Others prefer to tilt to exposed to non-market factor risks and increase return. Some (including myself) are perfectly content with sticking to market portfolio.
I thought we were talking about the theory here. If it's true that your larger expected returns are due to larger risks, then the implication is that a leveraged market portfolio would have a better risk/return profile because you are moving along the tangent line, whereas by over-weighing the small cap or value stocks you are not.
If you want to talk about theory, then I suggest you read https://eml.berkeley.edu/~craine/EconH1 ... _world.pdf (written by Fama's colleague and SIL). It explains efficient surface and tangency portfolio in a multi dimension risk framework (instead of mean-variance framework behind CAPM). In this multi-dimension framework, everyone has a different tangency portfolio because they have different indifference curve.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

cos wrote: Sun May 02, 2021 10:18 am Not quite. In terms of straight risk-adjusted returns, maybe, but exposure to multiple different risk factors also offers diversification since they are fairly uncorrelated with each other.
Yeah, perhaps you have in mind Larry Swedroe ? He made a similar point. His position was that you can increase the risk by over-weighing small cap stocs, and then you can compensate by increasing the weights of long term bonds, so that you are lowering the overall risk by decreasing the correlations among your assets.

But this is valid if the efficient frontier is calculated among stocks only. If you calculate it including both stocks and bonds (Sharpe portfolio), then there is no escape, the correlation is already factored in and you can't do better than the tangent line (assuming free leverage, of course).
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

Astones wrote: Sun May 02, 2021 9:54 am 1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
I think there is a substantial difference between increasing risk by leveraging TSM versus tilting a TSM dominated portfolio towards SV. FF showed that size and value are sources of risk and return that are unique and independent from the market factor. Size and value are also uncorrelated with each other. The leveraged TSM portfolio is just taking on more of the same type of risk, market risk. Increasing expected return by adding size and value implies taking on additional different risks than market risk. This is more efficient. Last I looked, the correlations were about
Market-size 0.4
Market-value 0.14
Size-value 0.14

Dave
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

Random Walker wrote: Sun May 02, 2021 1:25 pm
Astones wrote: Sun May 02, 2021 9:54 am 1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
I think there is a substantial difference between increasing risk by leveraging TSM versus tilting a TSM dominated portfolio towards SV. FF showed that size and value are sources of risk and return that are unique and independent from the market factor. Size and value are also uncorrelated with each other. The leveraged TSM portfolio is just taking on more of the same type of risk, market risk. Increasing expected return by adding size and value implies taking on additional different risks than market risk. This is more efficient. Last I looked, the correlations were about
Market-size 0.4
Market-value 0.14
Size-value 0.14

Dave
Ok, but I hope you'll agree with me that if this is the case -you can improve your risk/expected return profile by changing weights- then it means that Fama is wrong, and market is not efficient at all.
To me efficient market means precisely that a total market index must lie on the efficient tangent line, and of course there is no way to beat that line, by construction.
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

Astones wrote: Sun May 02, 2021 2:06 pm Ok, but I hope you'll agree with me that if this is the case -you can improve your risk/expected return profile by changing weights- then it means that Fama is wrong, and market is not efficient at all.
To me efficient market means precisely that a total market index must lie on the efficient tangent line, and of course there is no way to beat that line, by construction.
That’s not what the efficient market hypothesis says. It makes no prediction about the optimal portfolio to hold.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

vineviz wrote: Sun May 02, 2021 2:09 pm That’s not what the efficient market hypothesis says. It makes no prediction about the optimal portfolio to hold.
I believe it does, actually.
There's the need to dig a little to reach that conclusion, since it's not explicitly worded in the definition of EMH, but if the prices faithfully represent the risk/return profile according to the information available, it follows that a total market, cap weighted index must lie on the efficient frontier. Not only that, but it must be precisely where the tangent and the frontier overlap.
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 2:18 pm
vineviz wrote: Sun May 02, 2021 2:09 pm That’s not what the efficient market hypothesis says. It makes no prediction about the optimal portfolio to hold.
I believe it does, actually.
There's the need to dig a little to reach that conclusion, since it's not explicitly worded in the definition of EMH, but if the prices faithfully represent the risk/return profile according to the information available, it follows that a total market, cap weighted index must lie on the efficient frontier. Not only that, but it must be precisely where the tangent and the frontier overlap.
They are 2 different "efficient". EMH doesn't imply the market portfolio must be on efficient frontier. It doesn't mean market is the tangency portfolio. Those 2 are implications of CAPM not EMH.
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

Astones wrote: Sun May 02, 2021 2:18 pm
vineviz wrote: Sun May 02, 2021 2:09 pm That’s not what the efficient market hypothesis says. It makes no prediction about the optimal portfolio to hold.
I believe it does, actually.
There's the need to dig a little to reach that conclusion, since it's not explicitly worded in the definition of EMH, but if the prices faithfully represent the risk/return profile according to the information available, it follows that a total market, cap weighted index must lie on the efficient frontier. Not only that, but it must be precisely where the tangent and the frontier overlap.
It doesn't follow from the efficient market hypothesis precisely because the EMH explicitly does not contain an asset pricing model. There's no way to construct an efficient frontier using just EMH: you also need to specify a pricing model

Sounds like you're using the Sharpe-Lintner CAPM - at least implicitly - as your pricing model, so your conclusion requires both EMH and CAPM to be true. We know from decades of testing that the EMH is mostly true and the CAPM is mostly false, and CAPM is a particularly poor choice when the problem already specified that there are multiple sources of systematic risk (e.g. market, size, value).
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

Astones wrote: Sun May 02, 2021 2:06 pm
Random Walker wrote: Sun May 02, 2021 1:25 pm
Astones wrote: Sun May 02, 2021 9:54 am 1. Here maybe you can solve a doubt that I have been having for a while. It seems to me that what Fama says is that value and small-cap stocks tend to outperform because they are more risky, and this is why they don't break the EMH. But if this is the case, then over-weighing them would not be better than just leveraging the market portfolio (assuming free leverage) ?
I think there is a substantial difference between increasing risk by leveraging TSM versus tilting a TSM dominated portfolio towards SV. FF showed that size and value are sources of risk and return that are unique and independent from the market factor. Size and value are also uncorrelated with each other. The leveraged TSM portfolio is just taking on more of the same type of risk, market risk. Increasing expected return by adding size and value implies taking on additional different risks than market risk. This is more efficient. Last I looked, the correlations were about
Market-size 0.4
Market-value 0.14
Size-value 0.14

Dave
Ok, but I hope you'll agree with me that if this is the case -you can improve your risk/expected return profile by changing weights- then it means that Fama is wrong, and market is not efficient at all.
To me efficient market means precisely that a total market index must lie on the efficient tangent line, and of course there is no way to beat that line, by construction.
This is very dicey territory for me. I’m pretty sure that Fama says TSM is always on the efficient frontier. And I’ve asked Larry Swedroe this question as well, and he agrees TSM is on the efficient frontier. But that being said, I also believe, like Larry has taught, that a portfolio can be made more efficient by diversifying across independent sources of risk. That means there are portfolios above and to the left of TSM on plot of return v. SD. I have difficulty making sense of this. The only way I have rationalized it in my head is perceiving a distinction between an efficient market and an efficient portfolio.
Take a look at the Gibson paper I linked above. If TSM is just one asset in a portfolio, with one source of return (market beta), then a more efficient portfolio than TSM alone should be attainable by adding other uncorrelated assets. Interested in what you think. I struggle with this as well.

Dave
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Random Walker wrote: Sun May 02, 2021 2:51 pm
This is very dicey territory for me. I’m pretty sure that Fama says TSM is always on the efficient frontier. And I’ve asked Larry Swedroe this question as well, and he agrees TSM is on the efficient frontier. But that being said, I also believe, like Larry has taught, that a portfolio can be made more efficient by diversifying across independent sources of risk. That means there are portfolios above and to the left of TSM on plot of return v. SD. I have difficulty making sense of this. The only way I have rationalized it in my head is perceiving a distinction between an efficient market and an efficient portfolio.

This paper explains why MKT is on the efficient surface but not mean-variance efficient. https://eml.berkeley.edu/~craine/EconH1 ... _world.pdf
Random Walker
Posts: 4847
Joined: Fri Feb 23, 2007 8:21 pm

Re: Practice vs theory in portfolio strategies

Post by Random Walker »

acegolfer wrote: Sun May 02, 2021 2:55 pm
Random Walker wrote: Sun May 02, 2021 2:51 pm
This is very dicey territory for me. I’m pretty sure that Fama says TSM is always on the efficient frontier. And I’ve asked Larry Swedroe this question as well, and he agrees TSM is on the efficient frontier. But that being said, I also believe, like Larry has taught, that a portfolio can be made more efficient by diversifying across independent sources of risk. That means there are portfolios above and to the left of TSM on plot of return v. SD. I have difficulty making sense of this. The only way I have rationalized it in my head is perceiving a distinction between an efficient market and an efficient portfolio.

This paper explains why MKT is on the efficient surface but not mean-variance efficient. https://eml.berkeley.edu/~craine/EconH1 ... _world.pdf
I’ve read that paper before and liked it a lot. As I think you noted, it expands the concept to multiple dimensions of risk beyond the single measure of volatility. That would certainly make sense that MKT accounts for all the dimensions of risk and weights them appropriately, and therefore would land on the efficient surface. I’ll have to read the essay.
Dave
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

vineviz wrote: Sun May 02, 2021 2:36 pm
Sounds like you're using the Sharpe-Lintner CAPM - at least implicitly - as your pricing model, so your conclusion requires both EMH and CAPM to be true. We know from decades of testing that the EMH is mostly true and the CAPM is mostly false, and CAPM is a particularly poor choice when the problem already specified that there are multiple sources of systematic risk (e.g. market, size, value).
I might be getting confused so let's try to work this out.

Let's forget about CAPM for a moment and let's focus on EMH.

EMH states that you can't consistently generate alpha ( ="beat the market" ), at least according to the definition provided by investopedia.

If it were possible to build a portfolio with higher expected returns for the same risk compared to the market portfolio, wouldn't you be generating alpha by definition?
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

Astones wrote: Sun May 02, 2021 3:11 pm EMH states that you can't consistently generate alpha ( ="beat the market" ), at least according to the definition provided by investopedia.
If that's what investopedia actually says then Investopedia is wrong. Or at least incomplete.

Alpha is a measure of risk-adjusted return. Which begs the question, alpha relative to what? This is where you have to go back and pick the asset pricing model and all the baggage that comes with it.

If you want to fully understand the EMH you'll have to dig into the literature around it, or at least some of the better text books. In any case you could do a lot worse than Burton Malkiel's review.
....equity prices adjust to new information without delay and, as a result, no arbitrage opportunities exist that would allow investors to achieve above-average returns without accepting above-average risk.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 3:11 pm
Let's forget about CAPM for a moment and let's focus on EMH.

EMH states that you can't consistently generate alpha ( ="beat the market" ), at least according to the definition provided by investopedia.

If it were possible to build a portfolio with higher expected returns for the same risk compared to the market portfolio, wouldn't you be generating alpha by definition?
1. One can't consistently generate alpha is one of the implications, not the definition of EMH.
2. You can't really discuss alpha without assuming an asset pricing model.
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 3:22 pm
1. One can't consistently generate alpha is one of the implications, not the definition of EMH.
2. You can't really discuss alpha without assuming an asset pricing model.
Whether it is explicitly contained in the definition or if we just infer it from it, the point is the same.

If the market is efficient, then you can't build a portfolio providing better returns for the same risks as the market. Do we agree over this?

If we disagree, then I don't get how at the same time we can say that the market is efficient, yet there are strategies that are theoretically better than the market on a risk/return framework.
vineviz wrote: Sun May 02, 2021 3:18 pm If you want to fully understand the EMH you'll have to dig into the literature around it, or at least some of the better text books. In any case you could do a lot worse than Burton Malkiel's review.
I'll do my research, but from what I see, to me the quote you just posted seems to imply that the market portfolio lies, indeed, on the frontier ?
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

Astones wrote: Sun May 02, 2021 3:39 pm If the market is efficient, then you can't build a portfolio providing better returns for the same risks as the market. Do we agree over this?
That's a reasonable statement, as far as it goes.

But with multiple dimensions of risk, it's unlikely that two different portfolios will present an investor with the same risks. Or that two different investors will perceived the same portfolio as having the same risks.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
acegolfer
Posts: 2503
Joined: Tue Aug 25, 2009 9:40 am

Re: Practice vs theory in portfolio strategies

Post by acegolfer »

Astones wrote: Sun May 02, 2021 3:39 pm you can't build a portfolio providing better returns for the same risks as the market. Do we agree over this?
That's what Fama said and I agree. Under any asset pricing model that we can think of, market portfolio is efficient.
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

acegolfer wrote: Sun May 02, 2021 5:09 pm
Astones wrote: Sun May 02, 2021 3:39 pm you can't build a portfolio providing better returns for the same risks as the market. Do we agree over this?
That's what Fama said and I agree. Under any asset pricing model that we can think of, market portfolio is efficient.
But only in aggregate: it need not be efficient for any particular investor.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

acegolfer wrote: Sun May 02, 2021 5:09 pm That's what Fama said and I agree. Under any asset pricing model that we can think of, market portfolio is efficient.
Doesn't it imply that the market portfolio will be on the intersection between frontier and tangent ?
User avatar
willthrill81
Posts: 24507
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Practice vs theory in portfolio strategies

Post by willthrill81 »

vineviz wrote: Sun May 02, 2021 5:13 pm
acegolfer wrote: Sun May 02, 2021 5:09 pm
Astones wrote: Sun May 02, 2021 3:39 pm you can't build a portfolio providing better returns for the same risks as the market. Do we agree over this?
That's what Fama said and I agree. Under any asset pricing model that we can think of, market portfolio is efficient.
But only in aggregate: it need not be efficient for any particular investor.
On a related note, the market's aggregate holdings are not necessarily optimal in any way for any particular investor. For instance, TBM is not mathematically optimal for anyone whose investment horizon does not match the average duration of TBM. Further, the relative sizes of the stock and bond markets (roughly 43/57, IIRC) is not a compelling reason in itself for any particular investor to hold these assets in the same proportions.
“Good and ill have not changed since yesteryear; nor are they one thing among Elves and Dwarves and another among Men.” J.R.R. Tolkien, The Lord of the Rings
Topic Author
Astones
Posts: 236
Joined: Mon Apr 12, 2021 12:48 pm

Re: Practice vs theory in portfolio strategies

Post by Astones »

willthrill81 wrote: Sun May 02, 2021 5:25 pm
On a related note, the market's aggregate holdings are not necessarily optimal in any way for any particular investor. For instance, TBM is not mathematically optimal for anyone whose investment horizon does not match the average duration of TBM. Further, the relative sizes of the stock and bond markets (roughly 43/57, IIRC) is not a compelling reason in itself for any particular investor to hold these assets in the same proportions.
That's not a problem, at least theoretically, because once you find the intersection between frontier and tangent, then you can move along the tangent by tuning the allocation you give to the risk-free asset.
User avatar
vineviz
Posts: 9846
Joined: Tue May 15, 2018 1:55 pm

Re: Practice vs theory in portfolio strategies

Post by vineviz »

Astones wrote: Sun May 02, 2021 5:32 pm
willthrill81 wrote: Sun May 02, 2021 5:25 pm
On a related note, the market's aggregate holdings are not necessarily optimal in any way for any particular investor. For instance, TBM is not mathematically optimal for anyone whose investment horizon does not match the average duration of TBM. Further, the relative sizes of the stock and bond markets (roughly 43/57, IIRC) is not a compelling reason in itself for any particular investor to hold these assets in the same proportions.
That's not a problem, at least theoretically, because once you find the intersection between frontier and tangent, then you can move along the tangent by tuning the allocation you give to the risk-free asset.
This is where theory uncouples from practice: who can both borrow and lend at the risk free rate?

And even here, it’s almost certainly not the global market weight portfolio that is the tangency portfolio for you.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Post Reply