Yes, you can diversify a market portfolio

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Wed Jun 06, 2018 3:49 pm

jalbert wrote:
Wed Jun 06, 2018 2:00 pm
The problem with long duration bonds for investment portfolios is that the institutional investors who drive the market for them are insurance companies and pension funds and other players who need to cover long duration liabilities. Because they are matching duration to liabilities, they are not taking term risk, so they do not require to be compensated for term exposure. That is why the yield curve slope flattens so much as you move out on the term structure (even when the yield curve has average or above average slope by historical standards).
I'm most interested understanding how assets perform in a portfolio context, and that's the focus of this post.

You are definitely not alone in preferring shorter duration bond funds in investor portfolios, but I think this stems from thinking about the bond fund in isolation rather than in the way it interacts with the portfolio as a whole. I think that's one advantage of the developments over the past ten years or so in looking at risk optimization and risk budgeting (developments which, I might add, seem to have been spurred by real investors finding out in 2008-09 that portfolio's they THOUGHT were diversified really weren't).

Let's focus on Treasuries for a minute, in part because we have a pretty good series of index-based ETFs that have a track record of more than 10 years.
  • iShares Short Treasury Bond ETF (SHV)
  • iShares 1-3 Year Treasury Bond ETF (SHY)
  • iShares 3-7 Year Treasury Bond ETF (IEI)
  • iShares 7-10 Year Treasury Bond ETF (IEF)
  • iShares 10-20 Year Treasury Bond ETF (TLH)
  • iShares 20+ Year Treasury Bond ETF (TLT)
I took a look at the performance of ETF individually from Feb 2007 to May 2018, and then also as the bond component in a portfolio with 60% VTSMX and 40% each ETF. First, here are the Sharpe ratios to evaluate your assertion that institutional investors "do not require to be compensated for term exposure". I don't total return data for Treasuries that predates this (Morningstar's databased doesn't have a single Long Government Bond fund with an inception date before 1985, for instance), so a caveat that this is only 11 years definitely applies

Image

The blue bars represent the ETFs alone, and you can see your claim has some support: the ETF with the highest risk-adjusted return is iShares 1-3 Year Treasury Bond ETF (SHY) at 0.90.

But when you look at the ETF as part of a portfolio instead of in isolation (red bars) you get a different story: the best risk-adjusted PORTFOLIO in this example is 60% VTSMX & 40% TLT. That combination DRAMATICALLY outperformed

Still, this post is really about risk management NOT about chasing better returns (risk-adjusted or otherwise). So look at how dramatically differently the duration impacts the ETF volatility versus a portfolio containing the ETF.

Image

The portfolio containing iShares 10-20 Year Treasury Bond ETF (TLH) had LESS volatility than the portfolio containing iShares 1-3 Year Treasury Bond ETF (SHY). It was a much better diversifier in that it balanced the risk of VTSMX better. That the TLH portfolio had a CAGR that was 180 bps higher than the SHY portfolio was a happy bonus.

Which leads to the next point:
jalbert wrote:
Wed Jun 06, 2018 2:00 pm
If you use diversification ratio rather than some measure of risk-adjusted return as a measuring tool for a diversification benefit it would seem to be imperative to ensure risk was adequately rewarded for the individual asset classes.
I think this is an important observation. Because most risk-optimization procedures do NOT, by design, include any information about expected returns it is necessary for the person constructing the portfolio to have some mechanism for ensuring that the final portfolio meets their needs in terms of both expected reward and risk. This can be done through asset selection (so that you are optimizing risk starting from a basket of assets that all have acceptable returns), weight constraints (limiting low return assets, like short duration bonds, to XX%% of the portfolio no matter what the optimization regime might prefer, etc.), factor exposure targets, etc. Diversification ratio algorithms will at least pare down the assets in the portfolio as part of the process: risk parity and inverse volatility approaches will include every asset you feed them. Attilio Meucci has done some interesting work on using PCA and other techniques to interrogate the effective number of bets a portfolio contains, though the math is little daunting.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
nisiprius
Advisory Board
Posts: 36457
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: Yes, you can diversify a market portfolio

Post by nisiprius » Wed Jun 06, 2018 4:14 pm

Added: Note: see SimpleGift's remarks below. I added complexity, and SimpleGift was intentionally trying to keep things simple. Apologies if this posting confuses anyone.



I dunno about SimpleGift's analysis. It seems to me that it leaves out the role of the riskless asset and the tangent line.

Consider the low risk-tolerant investor.

Image

Without the riskless asset, the investor's personal tolerance curve touches the efficient frontier at the orange dot. But, if a riskless asset is available--and if it has a high enough return--the tangent line touches the efficient frontier at the orange dot. Is the orange dot superior? Not for this investor, because although it has a higher risk-adjusted return, it has higher risk than this investor prefers.

But he's not required to accept it. He can mix it with the riskless asset to obtain any portfolio anywhere on the green dotted line. The green dotted line intersects his risk preference line at the purple dot. This dot represents a portfolio that is acceptable for his risk tolerance, but has higher return than the orange dot.

The investor is better off using the one-size-fits-all tangent portfolio, and adjusting to his risk preference by diluting the tangent portfolio with some of the riskless asset, then by going lower on the efficient frontier curve.

Now, there are two notes. The first is that I'm not sure whether this always works in the case where the riskless asset has very low return. I'm not sure whether or not there's a credible risk tolerance curve that can come in below the tangent line. The second is that it may not work for the highly risk-tolerant investor who wants more risk than the tangent portfolio, because in order to go up on the green tangent line past the tangent point, it is necessary to be able to borrow at the risk-free rate.
Last edited by nisiprius on Wed Jun 06, 2018 7:52 pm, edited 2 times 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.

golfCaddy
Posts: 696
Joined: Wed Jan 10, 2018 10:02 pm

Re: Yes, you can diversify a market portfolio

Post by golfCaddy » Wed Jun 06, 2018 4:36 pm

The issue isn't finding uncorrelated strategies. There are over 600 factors in the published literature, any of which should diversify a portfolio, using your definition. Dogecoin should diversify a portfolio. Art, wine, and collectible cars add diversification. Betting on sports and horses adds diversification. The issue is finding an uncorrelated strategy, which you expect to provide returns near that of the total market.

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Yes, you can diversify a market portfolio

Post by SimpleGift » Wed Jun 06, 2018 4:43 pm

nisiprius wrote:
Wed Jun 06, 2018 4:14 pm
I dunno about SimpleGift's analysis. It seems to me that it leaves out the role of the riskless asset and the tangent line.
You're right, I just presented the most basic aspects of portfolio selection theory, in response to poster beehave professing "complete unfamiliarity with portfolio theory." Plus, I was hoping to avoid derailing this thread into a detailed theory discussion.

The riskless asset and the tangent line are certainly further aspects of the theory, but the main points I was hoping to get across are 1) there are many combinations of risky assets that can be considered efficient portfolios, 2) the choice of the optimal portfolio depends on the individual investor's risk tolerance, and 3) the market portfolio is the optimal portfolio reflecting the choices of all investors in aggregate.

Opting for simplicity and brevity, my hope is it will lead to clarity for someone newly thinking about portfolio theory.
Last edited by SimpleGift on Wed Jun 06, 2018 4:44 pm, edited 2 times in total.
Cordially, Todd

jalbert
Posts: 3584
Joined: Fri Apr 10, 2015 12:29 am

Re: Yes, you can diversify a market portfolio

Post by jalbert » Wed Jun 06, 2018 4:43 pm

I took a look at the performance of ETF individually from Feb 2007 to May 2018, and then also as the bond component in a portfolio with 60% VTSMX and 40% each ETF.
Here is another biased historical sample that produces the opposite result:

https://www.portfoliovisualizer.com/bac ... easury3=40

I think backtests are problematic as a portfolio construction tool because they try their best to lead you down a path of confusing outcome with strategy.
Index fund investor since 1987.

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Wed Jun 06, 2018 5:11 pm

jalbert wrote:
Wed Jun 06, 2018 4:43 pm
Here is another biased historical sample that produces the opposite result:
I can see the attempt here was to create a cherrypicked illustration, unlike my example of using the fully available data, but I suspect you failed to notice that your PV backtest ended up being only two years long. Portfolio Visualizer's Long Term Treasury data begins in 1978.

Use the full range of data in that model (1978 to 2018) and you'll find the same thing I found using the iShares data: taken alone the shorter duration bonds seem like the best choice, but the longer bonds actually do the job we need them to do in the portfolio.

https://www.portfoliovisualizer.com/bac ... easury3=40

Diversification is more than just choosing a bunch of random assets and throwing them in portfolio so that you can call it "diversified". And the cold hard truth is that picking the right credit and term risk for your bond fund (say, BLV versus BND) matters way more than fretting over whether your large-cap growth fund holds 3,609 stocks (VTSMX) or just 508 (VFINX) ever will.

We all know the drawbacks to backtesting but we also know that it is possible to address those drawbacks. And I I know how hard it is to actually dig into the research (both theoretical and empirical) that has been published over the past decade, but I'd much rather assume that there are new things I can learn than to cling with zealotry to old beliefs.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

jalbert
Posts: 3584
Joined: Fri Apr 10, 2015 12:29 am

Re: Yes, you can diversify a market portfolio

Post by jalbert » Wed Jun 06, 2018 7:06 pm

They are all cherry-picked data, that’s the point. In case you didn’t notice, treasury rates went from their all-time historical high point in 1981 to their all-time historical low in 2012. Consideration of appreciation from falling rates as reward for taking term risk is confusing outcome with strategy.
Index fund investor since 1987.

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Wed Jun 06, 2018 7:42 pm

jalbert wrote:
Wed Jun 06, 2018 7:06 pm
They are all cherry-picked data, that’s the point. In case you didn’t notice, treasury rates went from their all-time historical high point in 1981 to their all-time historical low in 2012. Consideration of appreciation from falling rates as reward for taking term risk is confusing outcome with strategy.
if you’re using all of the data available, it isn’t cherry picking.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

golfCaddy
Posts: 696
Joined: Wed Jan 10, 2018 10:02 pm

Re: Yes, you can diversify a market portfolio

Post by golfCaddy » Wed Jun 06, 2018 7:47 pm

vineviz wrote:
Wed Jun 06, 2018 7:42 pm
jalbert wrote:
Wed Jun 06, 2018 7:06 pm
They are all cherry-picked data, that’s the point. In case you didn’t notice, treasury rates went from their all-time historical high point in 1981 to their all-time historical low in 2012. Consideration of appreciation from falling rates as reward for taking term risk is confusing outcome with strategy.
if you’re using all of the data available, it isn’t cherry picking.
Starting at 1978 doesn't count as all available data.

jalbert
Posts: 3584
Joined: Fri Apr 10, 2015 12:29 am

Re: Yes, you can diversify a market portfolio

Post by jalbert » Wed Jun 06, 2018 7:48 pm

Maybe not intentional cherry-picking, but a biased sample which is the fundamental principle.
Index fund investor since 1987.

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Wed Jun 06, 2018 8:12 pm

golfCaddy wrote:
Wed Jun 06, 2018 7:47 pm
Starting at 1978 doesn't count as all available data.
It counts when there is no data for the asset before 1978, as is the case here.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Yes, you can diversify a market portfolio

Post by SimpleGift » Wed Jun 06, 2018 9:21 pm

jalbert wrote:
Wed Jun 06, 2018 4:43 pm
I think backtests are problematic as a portfolio construction tool because they try their best to lead you down a path of confusing outcome with strategy.
An excellent and subtle point, one that deserves more attention than it gets on the Forum — but I don't think it negates the general thesis being advanced by the OP in this thread that one CAN diversify the market portfolio by adding asset classes or factor exposures with less than perfect correlation to beta, plus higher risks (standard deviation) and expected returns. Not only does this make sense according to portfolio selection theory, but it's increasingly easy for individual investors to accomplish with low cost ETFs and index funds.

But SHOULD one diversify the simple Boglehead 2- or 3-fund portfolio? Backtesting can't really answer this question with any satisfaction, I don't believe, as the results are so period dependent (chart below) — and will never be able to tell us anything about how different portfolios will perform in the decades ahead.
Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
Cordially, Todd

Random Walker
Posts: 2979
Joined: Fri Feb 23, 2007 8:21 pm

Re: Yes, you can diversify a market portfolio

Post by Random Walker » Wed Jun 06, 2018 9:42 pm

SimpleGift wrote:
Wed Jun 06, 2018 9:21 pm
Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
Yes, we invest looking forward. We don’t know what will happen in the future. Past history has displayed independent and unique sources of return. Out of sample tests (different time periods, different geographic markets, different asset classes in some cases) further support the case that these sources of return are real. Looking forward, each has an intuitive risk based or behavior based reason to expect it to persist in the future. Diversifying across factors should lessen the potential for a long period of negative returns. Modern portfolio theory is very real. We can’t predict the future, so makes sense to diversify broadly across the sources of return that have worked in the past and we can rationally expect to occur in the future.

Dave

jalbert
Posts: 3584
Joined: Fri Apr 10, 2015 12:29 am

Re: Yes, you can diversify a market portfolio

Post by jalbert » Wed Jun 06, 2018 10:05 pm

Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
No argument there. Everyone wants to know what the holy grail of diversification would be. Nobody does. In general, it takes a very long time for return data for risky assets to converge on their mean (expected value). And even if you did a carefully designed experiment to get a reliable estimate for expected return, once you invest in the position, the same problem comes up in that it can take a long time for the investment outcome to converge on the mean. And this is if it is a stationary process which may not even be true.
Index fund investor since 1987.

golfCaddy
Posts: 696
Joined: Wed Jan 10, 2018 10:02 pm

Re: Yes, you can diversify a market portfolio

Post by golfCaddy » Wed Jun 06, 2018 11:35 pm

Random Walker wrote:
Wed Jun 06, 2018 9:42 pm
Looking forward, each has an intuitive risk based or behavior based reason to expect it to persist in the future.
Dave
Except, there's not an intuitive explanation for all these factors.

small - if small is either risk or behavioral based, it shouldn't disappear if you exclude the month of January. Are small stocks only more risky during January? Or do behavioral biases only occur in one month?

quality - There's no risk or behavioral explanation for why quality should beat junk. Quality should be less risky than junk by any reasonable definition of the term. On the behavioral side, no one says I want to buy stocks that are absolute, complete junk.

momentum - Even Fama acknowledges, it's difficult to explain momentum in terms of systemic risk. That doesn't mean the behavioral explanations are any better. What's the intuitive reason momentum excludes the past month? If theory of momentum is people follow trends, you would think the past month would be the most relevant. Why is momentum based on 12-month performance instead of 2-month performance or 24-month performance?

rgs92
Posts: 2139
Joined: Mon Mar 02, 2009 8:00 pm

Re: Yes, you can diversify a market portfolio

Post by rgs92 » Wed Jun 06, 2018 11:40 pm

As I recall reading very often, Mr. Bogle has shown that the S&P 500 by itself gives you all the diversification you need. Even using the total stock market index instead provides only the slightest improvement in performance and risk, and it is such a small difference that it is statistically unreliable.

So there seems to be little need for anything but the S&P 500 from what I have read despite all the theoretical arguments for foreign/small cap/mid cap funds.

(Full disclosure, I use total stock market and about 10% international, but I really can't honestly defend it. And VBINX/VBIAX, Vanguard Balanced Index, my major holdings, uses total stock market.)

And there is even less reason to go beyond the total bond market, since that's for just two purposes, risk mitigation and some modest steady income. Squeezing extra income out of this section of your portfolio leads to interest rate or credit risk, defeating the purpose of the fixed income allocation.

I hope Taylor agrees with this.

User avatar
Taylor Larimore
Advisory Board
Posts: 27307
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: Yes, you can diversify a market portfolio

Post by Taylor Larimore » Thu Jun 07, 2018 6:09 am

I hope Taylor agrees with this.
rgs92:

I do.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

rkhusky
Posts: 5399
Joined: Thu Aug 18, 2011 8:09 pm

Re: Yes, you can diversify a market portfolio

Post by rkhusky » Thu Jun 07, 2018 6:27 am

vineviz wrote:
Wed Jun 06, 2018 7:42 pm
if you’re using all of the data available, it isn’t cherry picking.
True, but there are still coincidental time-dependent effects when one uses just one set of data. One should look at a variety of time periods to see if the hypothesis is robust.

rkhusky
Posts: 5399
Joined: Thu Aug 18, 2011 8:09 pm

Re: Yes, you can diversify a market portfolio

Post by rkhusky » Thu Jun 07, 2018 6:54 am

Random Walker wrote:
Wed Jun 06, 2018 9:42 pm
SimpleGift wrote:
Wed Jun 06, 2018 9:21 pm
Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
Yes, we invest looking forward. We don’t know what will happen in the future. Past history has displayed independent and unique sources of return. Out of sample tests (different time periods, different geographic markets, different asset classes in some cases) further support the case that these sources of return are real. Looking forward, each has an intuitive risk based or behavior based reason to expect it to persist in the future. Diversifying across factors should lessen the potential for a long period of negative returns. Modern portfolio theory is very real. We can’t predict the future, so makes sense to diversify broadly across the sources of return that have worked in the past and we can rationally expect to occur in the future.

Dave
A group of stocks is a source of return, i.e. the goods and services that those companies provide are sources of return. Factors are not sources of return.

A group of small value stocks is a source of return, positive or negative. A group of small growth stocks is a source of return, positive or negative.

Factor analysis is a filtering method, whose goal is to come up with a group of stocks that will outperform in the future (although the original goal was simply to predict the return of a group of stocks using a few free parameters and the returns of the corresponding sub-groups of stocks) . Each stock in the group has the same factor characteristics. The factor characteristics are not different sources of return. The group of stocks is a source of return different from the stocks outside the group.

Filtering a large group of stocks down to a much smaller group is not diversification, it is the opposite of diversification. One is making a bet that a small segment of the market will outperform the market itself.

User avatar
nisiprius
Advisory Board
Posts: 36457
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: Yes, you can diversify a market portfolio

Post by nisiprius » Thu Jun 07, 2018 7:29 am

rkhusky wrote:
Thu Jun 07, 2018 6:54 am
...A group of stocks is a source of return, i.e. the goods and services that those companies provide are sources of return. Factors are not sources of return.

A group of small value stocks is a source of return, positive or negative. A group of small growth stocks is a source of return, positive or negative.

Factor analysis is a filtering method...

Filtering a large group of stocks down to a much smaller group is not diversification, it is the opposite of diversification. One is making a bet that a small segment of the market will outperform the market itself.
Yes to all of the above.

I think that claims made by factor advocates have shifted. When I first started following this forum in 2007, when factor investing was often referred to as "slice and dice," advocates were talking about diversifying across asset classes. Small, value, and of course small value, were asset classes. Roger Gibson wrote about "Asset Allocation and the Rewards of Multiple-Asset-Class Investing." The portfolios being explored were long-only portfolios. DFA was admired as a source of superior building blocks for these strategies because their funds succeeded in having heavier loads on the desired factors, while people discussed whether competitive products (like "pure value" ETFs) might be better because of being "more valuey." Nobody used short positions. Of late, the language has shifted from diversifying across asset classes to casual talk of diversifying across risk factors. But risk factors do not exist in the wild as investible assets. They exist only as mathematical constructs that can be approximated only by long-short portfolios, which must then be leveraged up because the market exposure of the short positions subtracts from overall return.

I will make one mild comment in favor of filtering being diversification. There is a homely analogy which might apply to factor-based investing (if it works). In engineering, it is sometimes found that much of the strain is being taken by part of the structure. For example, if you try to bend a solid bar (or the bone of an animal), the stress and strain are higher near the surface, and low near the center. This means that the parts near the center are contributing weight without contributing strength, and one can achieve a higher strength-for-weight ratio by hollowing out the structure element. In the case of human engineering, this might mean using tubes rather than solid bars. In the case of some birds, the bones are hollow and actually contain a kind of triangular truss-like lattice that looks remarkably like human engineering. I think this is what factor investors are trying to do and claim to be doing: they consider market beta to be dead weight and are trying to hollow out the portfolio by removing the parts that provide only market beta and nothing else.
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: 2979
Joined: Fri Feb 23, 2007 8:21 pm

Re: Yes, you can diversify a market portfolio

Post by Random Walker » Thu Jun 07, 2018 7:54 am

golfCaddy wrote:
Wed Jun 06, 2018 11:35 pm
Random Walker wrote:
Wed Jun 06, 2018 9:42 pm
Looking forward, each has an intuitive risk based or behavior based reason to expect it to persist in the future.
Dave
Except, there's not an intuitive explanation for all these factors.

small - if small is either risk or behavioral based, it shouldn't disappear if you exclude the month of January. Are small stocks only more risky during January? Or do behavioral biases only occur in one month?

quality - There's no risk or behavioral explanation for why quality should beat junk. Quality should be less risky than junk by any reasonable definition of the term. On the behavioral side, no one says I want to buy stocks that are absolute, complete junk.

momentum - Even Fama acknowledges, it's difficult to explain momentum in terms of systemic risk. That doesn't mean the behavioral explanations are any better. What's the intuitive reason momentum excludes the past month? If theory of momentum is people follow trends, you would think the past month would be the most relevant. Why is momentum based on 12-month performance instead of 2-month performance or 24-month performance?
To me, small stocks are intuitively more risky than big. Which company would be safer to loan money to, small or big?
I agree with you on profitability/quality. I’ve read some sort of risk story about profitable more subject to competitors coming in the business, but that does seem tenuous to me. Momentum is entirely behavioral. Can’t speak for the definition excluding most recent month, but human nature is so darn persistent and predictably subject to fear and greed.

Dave

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 8:15 am

rkhusky wrote:
Thu Jun 07, 2018 6:54 am
A group of stocks is a source of return, i.e. the goods and services that those companies provide are sources of return. Factors are not sources of return.
I feel like the terms here aren't quite consistent.

It's definitely true that the theoretical value of a stock is based on the present value of the expected future cash flows, and we shouldn't lose site of that.

But it is still true that style and risk factors are a source of return to the investor at the portfolio construction stage of investing. The portfolio (if well constructed) diversifies away the unsystematic risks associated with the discounted cash flows of each company, leaving the investor exposed to the systematic risks they chose to take on. Factors are just one way to evaluate which risks the portfolio is exposed to and, consequently, estimate its expected return.

So factor allocation is a potential source of return to the investor just like asset allocation is.
rkhusky wrote:
Thu Jun 07, 2018 6:54 am
Filtering a large group of stocks down to a much smaller group is not diversification, it is the opposite of diversification. One is making a bet that a small segment of the market will outperform the market itself.
This misrepresents the function of diversification, which is to manage risk (usually defined as portfolio volatility and/or concentrated exposure to uncertainty. Having a bunch of highly correlated stocks is not diversification, it's just an expensive mess.

Imagine you started with an equal-weight portfolio of the following ten stocks:
  • BAC (Bank of America Corporation)
  • C (Citigroup Inc.)
  • GILD (Gilead Sciences, Inc.)
  • GS (Goldman Sachs Group, Inc.)
  • JPM (J P Morgan Chase & Co)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • USB (U.S. Bancorp)
  • V (Visa Inc.)
  • WFC (Wells Fargo & Company)
"Filtering down", as you say, that list to an equal-weight portfolio of the following four stocks would indeed improve the portfolio's diversification:
  • GILD (Gilead Sciences, Inc.)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • V (Visa Inc.)

It does this without any view about the expected returns of the various stocks: you aren't making a bet on their outperformance, merely using the covariances to minimize the risks.

Even intuitively I think you can see that you've reduced your concentrated exposure to the financial services sector by spreading your risk more evenly to other sectors of the economy.

Quantitatively, you've increased the diversification ratio (one of the easiest-to-understand measures of diversification), cut the volatility nearly in half, and reduced the portfolio's conditional value-at-risk by 35%.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
nisiprius
Advisory Board
Posts: 36457
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: Yes, you can diversify a market portfolio

Post by nisiprius » Thu Jun 07, 2018 8:59 am

vineviz wrote:
Wed Jun 06, 2018 7:42 pm
if you’re using all of the data available, it isn’t cherry picking.
golfCaddy wrote:
Wed Jun 06, 2018 7:47 pm
Starting at 1978 doesn't count as all available data.
vineviz wrote:It counts when there is no data for the asset before 1978, as is the case here.
We can, however, look at the data back to 1926, from the Ibbotson SBBI Classic Yearbook, for "large-company stocks," "long-term government bonds," and "short-term government bonds."

As I understand it, Vineviz is saying that
  • long-term bonds are a better diversifier for stocks than intermediate-term bonds;
  • using all available data for certain specific ETFs, it was so effective that, in a 60/40 portfolio, that replacing lower-risk intermediate-term bonds with high-risk long-term bonds actually reduced the risk of the portfolio as a whole (as measured by standard deviation);
  • most important, that this paradoxical phenomenon (higher term risk in bonds results in lower risk in the portfolio) is so robust and reliable that one ought to adopt it in real-world portfolios and expect to see benefit from it over most reasonable time periods.
However, from 1926 through 2017, using the SBBI data sets, we did not see this. Substituting long-term for intermediate-term bonds in a 60/40 portfolio
  • increased the standard deviation of the portfolio as a whole,
  • also increased the return,
  • and overall resulted in a virtually identical Sharpe ratio. No harm, but, also, no improvement.
There is an interesting theoretical possibility that, under restricted and unusual conditions, increasing the risk of components in a portfolio can reduce the risk of the portfolio as a whole. But it is rare in practice.

Image
Image
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.

User avatar
aburntoutcase
Posts: 132
Joined: Tue Mar 13, 2018 12:06 pm

Re: Yes, you can diversify a market portfolio

Post by aburntoutcase » Thu Jun 07, 2018 9:29 am

vineviz wrote:
Tue Jun 05, 2018 6:00 pm
Likewise, the Vanguard Total Bond Market Index fund represents an impressive portfolio, but it does not have the level of exposure to credit risk and term risk it needs in order to balance the stocks held in VTSMX.
You should have stopped right there. It makes no sense whatsoever to say that adding credit risk to a bond fund balances the risk of stocks held in the portfolio. The credit risk in corporate bonds is correlated with the underlying equity risk.

DaufuskieNate
Posts: 248
Joined: Wed May 28, 2014 11:53 am

Re: Yes, you can diversify a market portfolio

Post by DaufuskieNate » Thu Jun 07, 2018 9:30 am

golfCaddy wrote:
Wed Jun 06, 2018 11:35 pm

quality - There's no risk or behavioral explanation for why quality should beat junk. Quality should be less risky than junk by any reasonable definition of the term. On the behavioral side, no one says I want to buy stocks that are absolute, complete junk.
Individual investors have shown a preference for small, high investment/low profitability stocks. These stocks typically have a great growth story associated with them and people are looking for that one chance to make a lot of money, much like a lottery. We all have heard the stories about normal people investing early on in some of the great business successes and becoming millionaires. This demand for positively skewed, lottery-like investments drives up prices and reduces future returns. Institutions are limited in their ability to take the other side and short these small stocks due to availability and costs of shorting. Quality is not really about taking on more risk. It's about weeding out lottery-like stocks that have excess demand beyond their fundamentals as individuals look for one really big payoff.

Random Walker
Posts: 2979
Joined: Fri Feb 23, 2007 8:21 pm

Re: Yes, you can diversify a market portfolio

Post by Random Walker » Thu Jun 07, 2018 9:40 am

Nisiprius wrote “There is an interesting theoretical possibility that, under restricted and unusual conditions, increasing the risk of components in a portfolio can reduce the risk of the portfolio as a whole. But it is rare in practice.”

I don’t think it’s that rare at all in practice. One needs to decide on a target expected return. If one keeps the target expected return constant, then in addition to adding asset classes with increased expected return, the investor will be decreasing overall equity allocation and increasing overall safe bond allocation. A more efficient portfolio can have the same mean expected return and a smaller expected SD. This more efficient portfolio will have less volatility, smaller max drawdowns, and greater compound return for same simple mean return than less efficient portfolio. Check out Larry Swedroe’s Black Swan book.

Dave

User avatar
Epsilon Delta
Posts: 7430
Joined: Thu Apr 28, 2011 7:00 pm

Re: Yes, you can diversify a market portfolio

Post by Epsilon Delta » Thu Jun 07, 2018 10:45 am

vineviz wrote:
Thu Jun 07, 2018 8:15 am

Imagine you started with an equal-weight portfolio of the following ten stocks:
  • BAC (Bank of America Corporation)
  • C (Citigroup Inc.)
  • GILD (Gilead Sciences, Inc.)
  • GS (Goldman Sachs Group, Inc.)
  • JPM (J P Morgan Chase & Co)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • USB (U.S. Bancorp)
  • V (Visa Inc.)
  • WFC (Wells Fargo & Company)
"Filtering down", as you say, that list to an equal-weight portfolio of the following four stocks would indeed improve the portfolio's diversification:
  • GILD (Gilead Sciences, Inc.)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • V (Visa Inc.)

It does this without any view about the expected returns of the various stocks: you aren't making a bet on their outperformance, merely using the covariances to minimize the risks.

Even intuitively I think you can see that you've reduced your concentrated exposure to the financial services sector by spreading your risk more evenly to other sectors of the economy.

Quantitatively, you've increased the diversification ratio (one of the easiest-to-understand measures of diversification), cut the volatility nearly in half, and reduced the portfolio's conditional value-at-risk by 35%.
Why not filter down to an un-equally weighted portfolio?
  • 25% SO (Southern Company)
  • 25% GILD (Gilead Sciences, Inc.)
  • 25% V (Visa Inc.)
  • 4% BAC (Bank of America Corporation)
  • 4% C (Citigroup Inc.)
  • 4% GS (Goldman Sachs Group, Inc.)
  • 4% JPM (J P Morgan Chase & Co)
  • 3% MS (Morgan Stanley)
  • 3% USB (U.S. Bancorp)
  • 3% WFC (Wells Fargo & Company)
Yes that adds complexity, but it's a toy example. In real life VTSMX exists.

User avatar
nisiprius
Advisory Board
Posts: 36457
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: Yes, you can diversify a market portfolio

Post by nisiprius » Thu Jun 07, 2018 11:06 am

Random Walker wrote:
Thu Jun 07, 2018 9:40 am
Nisiprius wrote “There is an interesting theoretical possibility that, under restricted and unusual conditions, increasing the risk of components in a portfolio can reduce the risk of the portfolio as a whole. But it is rare in practice.”

I don’t think it’s that rare at all in practice. One needs to decide on a target expected return. If one keeps the target expected return constant, then in addition to adding asset classes with increased expected return, the investor will be decreasing overall equity allocation and increasing overall safe bond allocation. A more efficient portfolio can have the same mean expected return and a smaller expected SD. This more efficient portfolio will have less volatility, smaller max drawdowns, and greater compound return for same simple mean return than less efficient portfolio. Check out Larry Swedroe’s Black Swan book.

Dave
Vineviz's example shows a fixed 60/40 allocation. He doesn't readjust the allocation for constant return. He keeps it at 60/40. He observes that over one time period, replacing the bond allocation with riskier bonds lowers the risk of the portfolio as a whole. That is the situation which I am saying is rare.
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.

rgs92
Posts: 2139
Joined: Mon Mar 02, 2009 8:00 pm

Re: Yes, you can diversify a market portfolio

Post by rgs92 » Thu Jun 07, 2018 11:08 am

Thank you Taylor. Best to you.

Random Walker
Posts: 2979
Joined: Fri Feb 23, 2007 8:21 pm

Re: Yes, you can diversify a market portfolio

Post by Random Walker » Thu Jun 07, 2018 11:15 am

Nisiprius,
That certainly makes sense to me. Sorry I misinterpreted. I remember reading once that for equity heavy portfolios (probably way beyond 60/40), may as well extend bond duration and take term premium because the portfolio risk is so overwhelmingly dominated by equity. Increasing duration in that case simply doesn’t add much to overall portfolio risk.

Dave

User avatar
David Jay
Posts: 5440
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Yes, you can diversify a market portfolio

Post by David Jay » Thu Jun 07, 2018 11:34 am

nisiprius wrote:
Thu Jun 07, 2018 8:59 am
[*]most important, that this paradoxical phenomenon (higher term risk in bonds results in lower risk in the portfolio) is so robust and reliable that one ought to adopt it in real-world portfolios and expect to see benefit from it over most reasonable time periods.[/list]
And herein lies the rub for real-world portfolios. Most us us are less than 20 years from retirement (few of us get serious about retirement until our 30s - I am always impressed with the 20-somethings who "get it") and the start of decumulation.
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

User avatar
nisiprius
Advisory Board
Posts: 36457
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: Yes, you can diversify a market portfolio

Post by nisiprius » Thu Jun 07, 2018 11:45 am

Random Walker wrote:
Thu Jun 07, 2018 11:15 am
Nisiprius,
That certainly makes sense to me. Sorry I misinterpreted.
I wasn't clear.
I remember reading once that for equity heavy portfolios (probably way beyond 60/40), may as well extend bond duration and take term premium because the portfolio risk is so overwhelmingly dominated by equity. Increasing duration in that case simply doesn’t add much to overall portfolio risk.

Dave
Well, my personal view is that you might as well increase duration up to about intermediate term, for the reasons you mention--portfolio risk is so overwhelmingly dominated by equity.

But when you are in the long-term space, you are starting to get into serious interest-rate risk and very serious inflation risk (unless you use TIPS), and your whole rationale for bonds had better change. It's no longer clear that the risk of bonds doesn't matter. You had better have a conviction in their ability to actively work the MPT miracle, fluctuating both with and against stocks, during whatever time period you hope to be invested.
Last edited by nisiprius on Thu Jun 07, 2018 11:56 am, 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.

Ben Mathew
Posts: 96
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle
Contact:

Re: Yes, you can diversify a market portfolio

Post by Ben Mathew » Thu Jun 07, 2018 11:55 am

golfCaddy wrote:
Wed Jun 06, 2018 11:35 pm
Except, there's not an intuitive explanation for all these factors.

small - if small is either risk or behavioral based, it shouldn't disappear if you exclude the month of January. Are small stocks only more risky during January? Or do behavioral biases only occur in one month?

quality - There's no risk or behavioral explanation for why quality should beat junk. Quality should be less risky than junk by any reasonable definition of the term. On the behavioral side, no one says I want to buy stocks that are absolute, complete junk.

momentum - Even Fama acknowledges, it's difficult to explain momentum in terms of systemic risk. That doesn't mean the behavioral explanations are any better. What's the intuitive reason momentum excludes the past month? If theory of momentum is people follow trends, you would think the past month would be the most relevant. Why is momentum based on 12-month performance instead of 2-month performance or 24-month performance?
I agree with this wholeheartedly. Factor models often try to recharacterize historical excess returns as rational risk premiums. There are some people in academia who are very invested in the idea that markets are highly efficient. So any systematic excess returns have to be recast as risk premiums for them to be comfortable with intellectually. There really is only one source of risk that falls naturally out of basic economic theory -- correlation to the aggregate economy. There has to be a risk premium for holding assets that do badly when the economy also does badly and everyone needs the money. This source of risk is captured by the beta of CAPM. This risk premium explains why stocks yield more than bonds, and why call options on stock ETFs yield much more than put options on stock ETFs. All other notions of risk are questionable after-the-fact concoctions. The historical excess returns from small, value, momentum, and quality cannot be explained by beta. These are therefore unexplained excess returns. Calling them reasonable premiums for taking on some sort of risk we don't fully understand is BS. This is not just a matter of semantics. It has real world consequences. Risk premia don't go away when everyone knows about it. Everyone knows that stocks yield more than bonds. But this premium is not going away because stocks lose value in a downturn when you need it the most. People know that stocks pay much more than bonds, but still hold bonds because they want the safety. But unjustified excess returns will go away if everyone catches on. They are much more fragile. I tilt towards small and value stocks. But I view this as betting on potentially underpriced stocks offering an unjustified return boost rather than as just compensation for taking on some sort of additional risk. If everyone felt this way, these premiums would go away (and, who knows, maybe they already have). I just don't have the same confidence in the persistence of the excess returns offered by my factor tilts as I do in my stock premium.

garlandwhizzer
Posts: 1933
Joined: Fri Aug 06, 2010 3:42 pm

Re: Yes, you can diversify a market portfolio

Post by garlandwhizzer » Thu Jun 07, 2018 1:15 pm

SimpleGift wrote:
Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
As usual, Todd hits the nail on the head IMO.

Garland Whizzer

rkhusky
Posts: 5399
Joined: Thu Aug 18, 2011 8:09 pm

Re: Yes, you can diversify a market portfolio

Post by rkhusky » Thu Jun 07, 2018 1:36 pm

vineviz wrote:
Thu Jun 07, 2018 8:15 am
Imagine you started with an equal-weight portfolio of the following ten stocks:
  • BAC (Bank of America Corporation)
  • C (Citigroup Inc.)
  • GILD (Gilead Sciences, Inc.)
  • GS (Goldman Sachs Group, Inc.)
  • JPM (J P Morgan Chase & Co)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • USB (U.S. Bancorp)
  • V (Visa Inc.)
  • WFC (Wells Fargo & Company)
"Filtering down", as you say, that list to an equal-weight portfolio of the following four stocks would indeed improve the portfolio's diversification:
  • GILD (Gilead Sciences, Inc.)
  • MS (Morgan Stanley)
  • SO (Southern Company)
  • V (Visa Inc.)

It does this without any view about the expected returns of the various stocks: you aren't making a bet on their outperformance, merely using the covariances to minimize the risks.

Even intuitively I think you can see that you've reduced your concentrated exposure to the financial services sector by spreading your risk more evenly to other sectors of the economy.

Quantitatively, you've increased the diversification ratio (one of the easiest-to-understand measures of diversification), cut the volatility nearly in half, and reduced the portfolio's conditional value-at-risk by 35%.
I don't agree that the smaller list is more diversified. Suppose that instead of picking 4 companies out of the original 10, Morgan Stanley bought the other 6 companies that were not chosen. You end up with the exact same 4 companies in your portfolio. Is there any difference in the diversification of the two 4-stock portfolios? Or between the 10-stock portfolio and the 2nd 4-stock portfolio?

By picking only 2 of the available financial companies, you are taking the risk that one or both of those companies will do poorly and the 6 companies not chosen will do better than average. You diversify your risk by investing in all 8 financial companies according to their share of the market.

Broken Man 1999
Posts: 1375
Joined: Wed Apr 08, 2015 11:31 am

Re: Yes, you can diversify a market portfolio

Post by Broken Man 1999 » Thu Jun 07, 2018 2:03 pm

Ah, this is a good thread!

I would suggest anyone who wants to learn more about factor investing should pick up copies of Larry Swedroe's latest two books, and read them both. Then read them both again. And even once again.
  • Reducing the Risk of Black Swans: Using the Science of Investing to Capture Returns with Less Volatility, 2018 EditionFeb 14, 2018
    by Larry Swedroe and Kevin Grogan
  • Your Complete Guide to Factor-Based Investing: The Way Smart Money Invests TodayOct7, 2016 by Andrew L. Berkin and Larry E. Swedroe
Broken Man 1999
Last edited by Broken Man 1999 on Thu Jun 07, 2018 2:06 pm, edited 1 time in total.
“If I cannot drink Bourbon and smoke cigars in Heaven than I shall not go. " -Mark Twain

jalbert
Posts: 3584
Joined: Fri Apr 10, 2015 12:29 am

Re: Yes, you can diversify a market portfolio

Post by jalbert » Thu Jun 07, 2018 2:04 pm

garlandwhizzer wrote:
Thu Jun 07, 2018 1:15 pm
SimpleGift wrote:
Thus it really comes down to how much one believes in the theory of multi-asset or factor investing and is willing to take their chances with investment portfolios that deviate from a standard neutral "market portfolio." Just my perspective.
As usual, Todd hits the nail on the head IMO.

Garland Whizzer
My quibble is with using long-term duration and long-duration credit as factors to add to the diversification mix of risk assets. I am not interested in a bond portfolio that depends on factor diversification working as expected to provide protection to the portfolio in an economic crisis or downturn. I don’t hold total bond but the long duration works there because it is counterbalanced by short-term instruments to provide income and liquidity when needed.
I guess one could argue that holding cash for liquidity and investing the remainder in the most efficient portfolio possible would be the MPT solution to the problem. But holding cash and long treasuries instead of intermediate treasuries is not really changing the term exposure.

I also don’t take much interest in asset classes that only add value to a portfolio if an uncertain diversification works as expected, vs assets classes where one expects risk to be rewarded for the asset class standalone as well as in a portfolio.

For instance, it is fine that security-specific risk of a stock is not rewarded because diversifying it away can be done reliably. But the fact that commodities have a zero expected real return means they can only add value to a diversified portfolio if the diversification works, and it is not a reliable diversification.
Last edited by jalbert on Fri Jun 08, 2018 10:15 am, edited 1 time in total.
Index fund investor since 1987.

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 2:39 pm

nisiprius wrote:
Thu Jun 07, 2018 11:06 am
Vineviz's example shows a fixed 60/40 allocation. He doesn't readjust the allocation for constant return. He keeps it at 60/40. He observes that over one time period, replacing the bond allocation with riskier bonds lowers the risk of the portfolio as a whole. That is the situation which I am saying is rare.
My post was meant to emphasize two important general statements, both of which are undoubtedly true and neither of which seems to be widely understood by individual investors.
  1. The market portfolio does NOT necessarily have the highest possible lever of diversification. It is possible to further diversify it.
  2. The behavior of a portfolio depends on both the characteristics of the component assets AND their interaction with the portfolio in aggregate. In other words, a well-constructed portfolio is more than the sum of its parts.
I've tried to illustrate those two points as well as I can, sometimes employing simplifying assumptions (e.g. holding the stock/bond allocation fixed) to make these concepts easier to understand. Obviously it's not clear that I succeeded.

It's definitely true that one benefit of taking more credit and term risk in the bond portion of a risky portfolio is that you could reallocate your risk budget to other factors (market beta, SmB, HmL, BaB, etc) if that seemed optimal. You could also allocate less overall capital to your risky portfolio and hold more riskless assets (e.g cash or money markets) if THAT seemed optimal.

So although the question of whether or not long term bonds outperform intermediate bonds over such-and-such time time period is interesting, it is also somewhat beside the point. That said, the backtesting that nisiprius misses some nuance (his approach added term risk but reduced credit risk relative to the example I started with, for instance, and his equity portfolio is simplified (only large cap domestic stocks). I point that out only because they are somewhat germane to my point #2 above.
Last edited by vineviz on Thu Jun 07, 2018 2:55 pm, edited 1 time in total.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 2:53 pm

Epsilon Delta wrote:
Thu Jun 07, 2018 10:45 am

Why not filter down to an un-equally weighted portfolio?

Yes that adds complexity, but it's a toy example. In real life VTSMX exists.
Sure, that works too. I was trying to illustrate a specific point (that you CAN increase diversification by reducing the NUMBER of holdings), but what really matter is the weight of each independent source of risk on the aggregate portfolio.

For sure you could dramatically improve the diversification of my original toy portfolio just by adjusting the weights.

In fact, the maximally diversified portfolio (based on 60 months of data) is this:

Code: Select all

Ticker	Name	Allocation
JPM	J P Morgan Chase & Co	8.17%
BAC	Bank of America Corporation	8.33%
V	Visa Inc.	11.41%
WFC	Wells Fargo & Company	10.03%
GS	Goldman Sachs Group, Inc.	1.20%
MS	Morgan Stanley	0.20%
SO	Southern Company	46.99%
GILD	Gilead Sciences, Inc.	13.68%
This portfolio has a diversification ratio of 2.01 and a monthly volatility of 2.52% versus my 4-stock EW example with a DR of 1.74 and a monthly volatility of 3.34%.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
David Jay
Posts: 5440
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Yes, you can diversify a market portfolio

Post by David Jay » Thu Jun 07, 2018 2:57 pm

vineviz wrote:
Thu Jun 07, 2018 2:39 pm
So although the question of whether or not long term bonds outperform intermediate bonds over such-and-such time time period is interesting, it is also somewhat beside the point.
That was in response to nisiprius' observation about practical portfolio construction considerations. In theory, where there is no need to consider portfolio sequence of returns issues, performance over "such-and-such time period" may be besides the point. For individuals facing very real retirement dates, loss of asset value is very much the point in practice.
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 3:08 pm

rkhusky wrote:
Thu Jun 07, 2018 1:36 pm
I don't agree that the smaller list is more diversified. Suppose that instead of picking 4 companies out of the original 10, Morgan Stanley bought the other 6 companies that were not chosen. You end up with the exact same 4 companies in your portfolio. Is there any difference in the diversification of the two 4-stock portfolios? Or between the 10-stock portfolio and the 2nd 4-stock portfolio?

By picking only 2 of the available financial companies, you are taking the risk that one or both of those companies will do poorly and the 6 companies not chosen will do better than average. You diversify your risk by investing in all 8 financial companies according to their share of the market.
Well, there is no doubt that the 4-stock portfolio I illustrated is MORE diversified than the 10-stock portfolio.

Your hypothetical acquisition spree is interesting, though, because it would effectively turn MS into basket all all seven formerly-independent stocks. Assuming that V, SO, and GILD are each at 25% then the new hypothetical MS conglomerate at 25% would in fact impact the portfolio differently than the historical MS would.

Even though most of risks of the 8 financial stocks overlap, they don't overlap perfectly. So if Morgan Stanley bought the other six (leaving out Visa for arguments sake) then the resulting 4-stock portfolio would certainly be MORE diversified than the one I presented. Not dramatically so, but (depending on the relative impact of each acquisition on Morgan Stanley's overall business) the impact is one you could definitely measure.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
David Jay
Posts: 5440
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Yes, you can diversify a market portfolio

Post by David Jay » Thu Jun 07, 2018 3:35 pm

vineviz wrote:
Thu Jun 07, 2018 3:08 pm
Well, there is no doubt that the 4-stock portfolio I illustrated is MORE diversified than the 10-stock portfolio.
Behold, the spherical cow. SD is lower over 60 months, so by definition 4 companies is more diversified than 10 companies.

[edit] Now all we have to do is find a company whose stock has ZERO SD and we will be perfectly diversified with just one stock.
Last edited by David Jay on Thu Jun 07, 2018 3:39 pm, edited 1 time in total.
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

Random Walker
Posts: 2979
Joined: Fri Feb 23, 2007 8:21 pm

Re: Yes, you can diversify a market portfolio

Post by Random Walker » Thu Jun 07, 2018 3:38 pm

Broken Man 1999 wrote:
Thu Jun 07, 2018 2:03 pm
Ah, this is a good thread!

I would suggest anyone who wants to learn more about factor investing should pick up copies of Larry Swedroe's latest two books, and read them both. Then read them both again. And even once again.
  • Reducing the Risk of Black Swans: Using the Science of Investing to Capture Returns with Less Volatility, 2018 EditionFeb 14, 2018
    by Larry Swedroe and Kevin Grogan
  • Your Complete Guide to Factor-Based Investing: The Way Smart Money Invests TodayOct7, 2016 by Andrew L. Berkin and Larry E. Swedroe
Broken Man 1999
Totally agree. For those who want a good short introduction to those reads or not interested in entire books, I’d recommend a short essay by Larry entitled Effective Diversification In A 3 Factor World. The essay only focuses on tilting to size and value, but it certainly shows how a more efficient portfolio can be built by diversifying across uncorrelated sources of return. Once an investor appreciates the ability to improve portfolio efficiency and the benefits of it, he can take that to the next level with the 5 or 6 factors in the factor book or even the alternatives mentioned in the newest edition of Black Swans. The link to the short 3 page essay is below. I think it is 3 of the most significant pages of investment reading I have ever found.

http://thebamalliance.com/pdfs/Effectiv ... wedroe.pdf




Dave

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 3:48 pm

David Jay wrote:
Thu Jun 07, 2018 2:57 pm
vineviz wrote:
Thu Jun 07, 2018 2:39 pm
So although the question of whether or not long term bonds outperform intermediate bonds over such-and-such time time period is interesting, it is also somewhat beside the point.
That was in response to nisiprius' observation about practical portfolio construction considerations. In theory, where there is no need to consider portfolio sequence of returns issues, performance over "such-and-such time period" may be besides the point. For individuals facing very real retirement dates, loss of asset value is very much the point in practice.
Oh yeah, definitely. I hope I'm not implying that I think diversification optimization is something anyone should pursue blindly: obviously you still need some mechanism for making sure that the portfolio has an expected return that suits your needs.

Honestly, I'm really only advocating that investors be more aware of the risks to which their portfolios are exposed. What they do with that awareness is ultimately a personal decision.

Mathematically though, if you take on more term and credit risk in your bond funds then you don't need nearly the market risk in your equity holdings. What follows is an INSANELY egregious example of cherrypicking, but hypothetically starting at the beginning of 1930 (as far back as Portfolio Visualizer lets me go) imagine you had two options.

Portfolio A is 60% large stocks and 40% intermediate government bonds.

Portfolio B is 30% small cap stocks, 45% long term corporate bonds, and 25% 30 day Treasury bills.

Rebalancing annually, your probably wouldn't be shocked to learn 88 years later that Portfolio B had substantially lower volatility and less severe drawdowns. And maybe you wouldn't be surprised that they achieved almost exactly the same CAGR (8.38% and 8.34% respectively), but I was.

https://www.portfoliovisualizer.com/bac ... tion6_2=25

I'm not suggesting that anyone pick a portfolio like this cherrypicked example. I AM suggesting that if investors think about their risk as something they allocate like their capital, it can lead to fresh thinking about how to best prepare for their "very real retirement dates".
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
vineviz
Posts: 1231
Joined: Tue May 15, 2018 1:55 pm

Re: Yes, you can diversify a market portfolio

Post by vineviz » Thu Jun 07, 2018 3:51 pm

David Jay wrote:
Thu Jun 07, 2018 3:35 pm
Behold, the spherical cow. SD is lower over 60 months, so by definition 4 companies is more diversified than 10 companies.
I feel like you're not working very hard to grasp my actual point.

Let me turn it around: what would lead you to assume that a ten stock portfolio is naturally more diversified than a four stock portfolio?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
sapphire96
Posts: 44
Joined: Fri Jun 16, 2017 8:08 pm

Re: Yes, you can diversify a market portfolio

Post by sapphire96 » Thu Jun 07, 2018 4:25 pm

vineviz wrote:
Thu Jun 07, 2018 3:08 pm
rkhusky wrote:
Thu Jun 07, 2018 1:36 pm
I don't agree that the smaller list is more diversified. Suppose that instead of picking 4 companies out of the original 10, Morgan Stanley bought the other 6 companies that were not chosen. You end up with the exact same 4 companies in your portfolio. Is there any difference in the diversification of the two 4-stock portfolios? Or between the 10-stock portfolio and the 2nd 4-stock portfolio?

By picking only 2 of the available financial companies, you are taking the risk that one or both of those companies will do poorly and the 6 companies not chosen will do better than average. You diversify your risk by investing in all 8 financial companies according to their share of the market.
Well, there is no doubt that the 4-stock portfolio I illustrated is MORE diversified than the 10-stock portfolio.

Your hypothetical acquisition spree is interesting, though, because it would effectively turn MS into basket all all seven formerly-independent stocks. Assuming that V, SO, and GILD are each at 25% then the new hypothetical MS conglomerate at 25% would in fact impact the portfolio differently than the historical MS would.

Even though most of risks of the 8 financial stocks overlap, they don't overlap perfectly. So if Morgan Stanley bought the other six (leaving out Visa for arguments sake) then the resulting 4-stock portfolio would certainly be MORE diversified than the one I presented. Not dramatically so, but (depending on the relative impact of each acquisition on Morgan Stanley's overall business) the impact is one you could definitely measure.
While I understand where you are coming from, I don’t necessarily agree with the logic that the four-stock portfolio is more diversified than the ten-stock portfolio...

All of the statistics for these companies are looking at past data, and past performance is not a guarantee of future results. All it takes is one financial scandal, or a major black swan event, to cause one of those holdings to significantly suffer, which is why I hesitate to fully agree with your argument. Yes, the backtest shows the four stocks as less volatile than the ten stocks, but that is not indicative of how the future performance will be. In other words, I don’t doubt the past results, but I doubt the future results.

But you do raise an interesting question... which is more diversified: a sector ETF (containing tens, if not hundreds, of stocks) or a portfolio with one stock in each sector?

Kind regards,

Ben

MIretired
Posts: 733
Joined: Fri Sep 06, 2013 12:35 pm

Re: Yes, you can diversify a market portfolio

Post by MIretired » Thu Jun 07, 2018 4:30 pm

vineviz,
What might your answer/opinions to this be:

OK, sure, according to the Cramer episodes ,'am I diversified', the 4 stocks in 4 relatively(2 types of financials?) different sectors IS MORE diversified than 6 finanacials, 1 energy?, and 1 healthcare(biotech.)

How about a universe of all US, and then selecting 1 or 2 stocks from each of S&P's 11 sectors.
You will probably get a more diversified portfolio after figuring out the appropriate weights. AND it's less # of stocks.

OK, I posted up thread that the TSM seems on the eff. frontier of all US stocks.

I read the 2nd paper from 2013 you originally made a link to in op.

OK. I understand that 1 CAPM short coming is that people can't 'borrow' and leverage up the Sharpe optimal portfolio ('the market portfolio) at the risk free rate. And so what probably happens is that if you want more beta than the optimal portfolio, you will instead just over weight higher beta stocks. Thereby skewing the real world's 'market portfolio' capitalization.

Now, in the DR theory paper, it would call the TSM an upto 3600+ risk factor portfolio. Or after determining a most diversified portfolio(MDP), the # of stocks selected for unique risk would be the # of risks.
The same for my 11 or 22 stocks of 11 different sectors. Upto 11 or 22 different risks. And they can probably, realistically be more diversified than the TSM or my 22 stocks,say, cap weighted.
For 1 reason, the slide up towards growth or higher beta stock weighting taking place in the TSM.

I know you are primarily showing the difference in bonds selection. Maybe adding small caps as you did. So you could dismiss this as too smallish a concern.

But, the stocks the MDP selects is going to vary per time period of measure. And this becomes a problem. This reduces diversification over time; or the reliability of the diversification over time. Maybe this will be the reason this is dismissed as being too narrow of pickin's universe.

User avatar
David Jay
Posts: 5440
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Yes, you can diversify a market portfolio

Post by David Jay » Thu Jun 07, 2018 5:40 pm

vineviz wrote:
Thu Jun 07, 2018 3:51 pm
David Jay wrote:
Thu Jun 07, 2018 3:35 pm
Behold, the spherical cow. SD is lower over 60 months, so by definition 4 companies is more diversified than 10 companies.
I feel like you're not working very hard to grasp my actual point.

Let me turn it around: what would lead you to assume that a ten stock portfolio is naturally more diversified than a four stock portfolio?
There are dramatic reductions in non-systemic risk as a portfolio increases to 30 stocks with continuing benefits to over 100 stocks. That is my entire point in this thread, measuring “risk” as the SD of the particular assets is only looking at a very narrow subset of risks.

In a portfolio containing a single-digit number of individual stocks, I would suggest that more is almost categorically better than less. Hence my tongue-in-cheek remark about finding the perfect single stock with no volatility whatsoever (perfect diversification in one stock as measured by standard deviation).

What if one of your 4 stocks is the next Kodak (a company that is about to be inundated by sudden technological change)?
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

golfCaddy
Posts: 696
Joined: Wed Jan 10, 2018 10:02 pm

Re: Yes, you can diversify a market portfolio

Post by golfCaddy » Thu Jun 07, 2018 5:55 pm

DaufuskieNate wrote:
Thu Jun 07, 2018 9:30 am
golfCaddy wrote:
Wed Jun 06, 2018 11:35 pm

quality - There's no risk or behavioral explanation for why quality should beat junk. Quality should be less risky than junk by any reasonable definition of the term. On the behavioral side, no one says I want to buy stocks that are absolute, complete junk.
Individual investors have shown a preference for small, high investment/low profitability stocks. These stocks typically have a great growth story associated with them and people are looking for that one chance to make a lot of money, much like a lottery. We all have heard the stories about normal people investing early on in some of the great business successes and becoming millionaires. This demand for positively skewed, lottery-like investments drives up prices and reduces future returns. Institutions are limited in their ability to take the other side and short these small stocks due to availability and costs of shorting. Quality is not really about taking on more risk. It's about weeding out lottery-like stocks that have excess demand beyond their fundamentals as individuals look for one really big payoff.
You seem to be confusing value with quality. They are independent factors.

golfCaddy
Posts: 696
Joined: Wed Jan 10, 2018 10:02 pm

Re: Yes, you can diversify a market portfolio

Post by golfCaddy » Thu Jun 07, 2018 6:00 pm

Random Walker wrote:
Thu Jun 07, 2018 7:54 am
To me, small stocks are intuitively more risky than big. Which company would be safer to loan money to, small or big?
Why would small stocks only be more risky in the month of January? The small factor disappears entirely if you exclude January.
Momentum is entirely behavioral. Can’t speak for the definition excluding most recent month, but human nature is so darn persistent and predictably subject to fear and greed.
What is it about human nature that makes 12-months the magic number for momentum and not 24 months or 2 months? Are we supposed to assume greed and fear operate on 12-month cycles like the earth revolving around the sun?

Dave

Post Reply